Kulite Pressure Transducer Handbook

123

Pressure Transducer Handbook

TABLE OF CONTENTS SECTION 1 - Introduction 1.1 Introduction 1.2 Product Overview SECTION 2 - Kulite Sensing Technology 2.1 Pressure Transducers 2.2 Theory of Operation 2.3 Transducer Types 2.4 The Piezoresistive Effect 2.5 The Piezoresistive Effect in Silicon 2.6 Sensing Elements 2.6.1 Discrete Gauges 2.6.2 Integral Gauges 2.6.3 The Wheatstone Bridge 2.6.4 Key Characteristics of Kulite Piezoresistive Technology 2.7 Microphones 2.8 Integrated Sensor Design 2.8.1 Dielectrically Isolated Design (SOI) 2.8.2 Diaphragm Characteristics 2.8.3 Isolated Capsule Design 2.8.4 Dual Diaphragm Technology 2.8.5. Redundant & Combination Pressure/ Temperature Transducers 2.8.6 Kulite Leadless Design 2.8.6.1 Leadless/ Acceleration Compensated Design 2.8.7 Temperature compensation 2.8.7.1 Bridge Zero and Zero Shift Compensation 2.8.7.2 Bridge Sensitivity Compensation 2.8.8 Mechanical Design 2.8.9 Silicon Carbide SECTION 3 - Performance Characteristics 3.1 Dynamic Range 3.1.1 Definition 3.1.2 Range 3.1.3 Overrange

3.2 Sensitivity 3.2.1 Sensitivity Calibration 3.2.2 Polarity

3.3 Nonlinearity, Hysteresis and Non-repeatability

2

3.3.1 Definitions 3.3.3.1 Linearity 3.3.1.2 Hysteresis 3.3.1.3 Non-repeatability 3.3.2 Nonlinearity 3.3.3 Hysteresis 3.3.4 Non-repeatability 3.3.5 Combining Nonlinearity, Hysteresis and Non-repeatability

3.4 Zero Measurand Output (ZMO) 3.4.1 Mounting Effects 3.4.2 Warm-up 3.4.3 Thermal Stability 3.4.4 Effect of Overpressure

3.5 Phase Shift

3.6 Input and Output Resistance

3.7 Thermal Sensitivity Shift and Zero Shift 3.7.1 Thermal Sensitivity Shift 3.7.2 Thermal Zero Shift 3.7.3 Thermal Transient Response

3.8 Photo Flash Response

3.9 Transducer Resonant Frequency

3.10 Frequency Response 3.10.1 Rise and Response Time

3.11 Acceleration Sensitivity

3.12 Burst Pressure

3.13 Full Scale Output

3.14 Supply Voltage or Excitation

3.15 Input / Output Resistance

3.16 Insulation Resistance SECTION 4 - Environmental Limits 4.1 Diaphragm Loading

4.2 Temperature

4.3 Acceleration, Shock and Vibration

4.4 RF and Magnetic Fields

4.5 Sealing and Hermeticity

4.6 Media Compatibility 4.6.1 Pressure Sensitive End 4.6.2 Electrical Lead End

4.7 Nuclear Radiation SECTION 5 - Application Information 5.1 Connection Diagrams

123 Semiconductor Products, Inc. Page 0-3

5.2 Mounting Techniques 5.2.1 Strain Sensitivity 5.2.1.1 Threaded Mounting Configurations 5.2.1.2 Cylindrical configurations 5.2.1.3 Thin Line Transducers (flat-pack) 5.2.2 Strain Measurement 5.3 Insulation

5.4 Cabling 5.4.1 Standard Cables 5.4.2 Splicing and Extension Cables 5.4.3 Loading Effects 5.4.4 Effects of Cable on Transducer Sensitivity 5.4.4.1 Excitation voltage Drop 5.4.4.2 Signal Attenuation 5.4.4.3 RC Filtering 5.5 Measurement of Dynamic Pressures 5.5.1 Acoustic and Fluid flow Effects 5.5.1.1 Acoustic Fundamentals 5.5.1.2 Sound Speed in Gas 5.5.1.3 Organ Pipe Resonance 5.5.1.4 Cavity Resonances (Helmholtz) 5.5.1.5 Transmitting Tube Connected to Cavity 5.5.1.6 Pressures in a Flowing Fluid 5.5.1.7 Pressure Shock Wave Effects 5.5.2 Acoustic Limitations of a Pressure Probe 5.5.3 Dynamic Response of a Transducer in a Liquid System 5.5.4 Dynamic Pressure Measurements at High Temperatures

SECTION 6 - Electronics 6.1 Power for Excitation 6.1.1 DC Power Supplies 6.1.1.1 Constant Current Power Sources 6.1.1.2 External Sensing 6.1.2 AC Excitation 6.2 Signal Conditioning 6.2.1 Analogue Amplifiers 6.2.2 Digital Corrected Analogue Output 6.2.3 Digital Output 6.2.4 Pressure Switch Output 6.2.5 Solid State Replacements for Electro-Mechanical Pressure Transducers 6.2.6 Wireless Transmission 6.3 Readout and Recording Devices 6.3.1 Input Characteristics 6.3.2 Meter Characteristics 6.3.3 Errors in Digitising SECTION 7 - Measurement of Transient Pressure Pulses 7.1 Dynamic Range

4

7.2 Low Frequency Response

7.3 High Frequency Response

7.4 Phase Shift

7.5 Special considerations for Air Blast Measurements 7.5.1 Rise and Response Times 7.5.2 Spatial Averaging of Pressure Across Diaphragms 7.5.3 Mechanical Protection

SECTION 8 - Calibration 8.1 Temperature Calibrations

8.2 Electrical Calibrations

8.3 Static Calibrations 8.3.1 Dead Weight Testers

8.4 Dynamic Calibrations 8.4.1 Oscillating Pressure Calibrations 8.4.1.1 Hydraulic Pressure Generator 8.4.1.2 Vibrating Liquid Column 8.4.1.3 Inlet Modulated Pressure Generator (IMPG) 8.4.1.4 Gulton Whistle 8.4.1.5 Gas Pistonphone 8.4.2 Step Pressure Generators 8.4.2.1 Fast Acting Valves 8.4.2.2 Gas Shock Tubes SECTION 9 – Glossary, Unit Conversions & Kulite Reports 9.1 Glossary

9.2 Pressure Unit Conversions 9.2.1 Units of Measurement 9.2.2 Decibel Formulae

9.3 Kulite Reports

9.4 Kulite Patents


Section 1 - Introduction 1.1 Introduction Kulite Semiconductor Products is a privately owned and operated company which was founded in 1959 by Dr. A. D. Kurtz to manufacture Silicon Strain Gauges.

Dr. Kurtz and his team of engineers invented and patented the Silicon Integrated Pressure Sensor in the late 60’s and have developed the miniature test pressure transducer market. Such has been the influence of Kulite on the field of miniature dynamic pressure transducers that the word “Kulite” is frequently used to refer to a dynamic pressure transducer, even when it is not manufactured by Kulite. The next major development from Kulite after the Silicon Integrated Pressure Sensor, was the Silicon on Silicon Sensor design for high temperature operation which Kulite successfully patented.

Recent laboratory testing has demonstrated reliable operation of the silicon on silicon technology up to temperatures in excess of 1000 degrees F/ 540 degrees C. Kulite currently holds over 150 Patents on pressure sensor design and technology and employs over 480 employees worldwide. Sales are typically in excess of $60 Million US. Kulite’s 100,000 square feet world headquarters are located in two modern adjacent facilities in Leonia, New Jersey. Kulite also have four subsidiaries in Europe (UK, France, Italy and Germany) with a world-wide representative network. Kulite is currently manufacturing over 10,000 transducers/month

1.2 Product Overview Kulite is a world leader in the science and engineering of piezoresistive technology pressure sensors and manufactures a wide range of pressure transducers which are used wherever reliability, performance and value are required.

Miniature IS Silicon Diaphragm Pressure Transducers These transducers have found wide acceptance in the aerospace and the automobile industry, for wind tunnel, flight/ road testing and acoustic measurements. They have established the industry standard of excellence for dynamic pressure measurements. The small size of these devices has made them uniquely suited to a large variety of test and production applications in industry, research and development.

Precision Pressure Transducers and Transmitters These transducers are designed and produced for applications that require high performance, accuracy and reliability. They are specifically packaged to perform in severe environment pressure measurement situations. Internal Microprocessor compensation to eliminate the effects of temperature and non-linearity are now provided for the highest accuracy applications.

Aircraft Pressure Transducers There are a very wide range of Kulite Solid State Pressure Transducers which are used in numerous aircraft applications that require high performance and reliability. All Kulite Aircraft Transducers have evolved from four decades of having pioneered the development of miniature static and dynamic pressure measurement devices.


Automotive Pressure Transducers These transducers can be found in the automotive test lab, at the proving grounds, on the racetrack and even on the public highway. They are used to monitor brake lines, fuel and oil pressures, hydraulic pressures and pressures within automotive transmissions. They have been adapted to measure forces and structural vibration.

OEM Pressure Transducers The Original Equipment Manufacture (OEM) pressure transducers are solid state low cost pressure sensors available in mounted chip form to complete transducer assemblies. Strain Gauges

The semiconductor strain gauge may be thought of as a strain sensitive resistor. Generally cemented to a stressed member, its resistance changes as a function of applied strain. This characteristic makes it useful in the field of stress analysis, physical measurements and testing and transducer manufacture. Similar to conventional metallic wire and foil gauges, Kulite Semiconductor Gauges offer the significant advantages of higher sensitivity, smaller sizes, higher resistance, higher fatigue life and low hysteresis. Kulite piezoresistive strain gauges are also used in the manufacture of load cells.

Page 1-2 123 Semiconductor Products, Inc.

Section 2 – Kulite Sensing Technology 2.1 Pressure Transducers A pressure transducer produces an electrical output proportional to the pressure applied. The frequency of pressure fluctuation should be lower than the resonance frequency of the transducer and the electrical output is essentially independent of frequency below one-fifth the resonance frequency (flat frequency response). When pressure is applied, the force on the sensing element due to the pressure results in a deformation of the sensing element. This deformation changes the electrical properties of the element and the electrical output of the transducer. In a well-designed transducer, the deformation and electrical output are directly proportional to pressure over a wide range of frequencies. 2.2 Theory of Operation Pressure is defined as force per unit area. The most common measurements are made in gaseous or liquid media. All pressure gauges and transducers use a force-summing device to convert the pressure into a stress or displacement proportional to the pressure. In transducers, the stress or displacement is then applied to an electrical transduction element to generate the required signal. Kulite piezoresistive pressure transducers combine the force summing device and the transduction element into a micromachined, dielectrically isolated silicon diaphragm. The high stiffness, small size, and low mass of the transduction system provide an ideal combination of wide frequency response, high sensitivity, and immunity to acceleration and strain inputs. 2.3 Transducer Types Pressure transducers are available with four reference pressure options. These are (a) Absolute – (psia); (b) Gauge - (psig); (c) Differential - (psid) and (d) Sealed Reference – (psisg).

Figure 2-1: Pressure Reference Configurations

All pressure transducers use a force-summing device to convert pressure to displacement, but that displacement is then converted to an electrical output by a variety of transduction methods. Kulite specialises in piezoresistive strain sensing technology in silicon and, most recently, in silicon carbide for ultra high temperature applications. 2.4 The Piezoresistive Effect During his experiments with temperature measurements, Lord Kelvin discovered that the resistance of a conductor increased when it was in tension. In his work, this phenomenon was an annoying error source. In the 1930's it was recognised as a useful measurement tool, the metal strain gauge. When a conductor is strained, its length and thickness change (related by Poisson's Ratio). Since electric current is forced to travel a longer path of smaller area when the conductor is stretched by tension, the resistance of the conductor increases.


The resistance change, compared to the original resistance divided by the fractional change in length, is called the gauge factor, G. Different types of gauges exhibit different gauge factors:

Type of Gauge Gauge Factor Unbonded Wire 4 Bonded Foil 2 Thin Film 2 Bonded Discrete Semiconductor 50 to 200 Integral Diffused or Isolated Semiconductor 50 to 200

Higher gauge factor means higher output for the same strain, or higher sensitivity relative to the stiffness and natural frequency of the structure. Semiconductor gauges have much higher gauge factors than metal because, in addition to the lengthening and narrowing of the conductor, the resistivity of doped silicon changes under strain. The change in electrical resistance of a strain gauge with the application of a physical stress is referred to as the piezoresistive effect. For the measurement of strain in an object, the strain gauge is bonded to the object, which in the case of a pressure transducer, is usually a diaphragm. When a load (pressure) is applied, the diaphragm and the strain gauge both deform, causing the resistance of the strain gauge to change. This resistance can be calculated as follows. The resistance of a wire is

ALR ρ= , (2.1)

where L is the length of the wire, ρ is resistivity of the wire material and A is the wire’s cross-sectional area. Differentiating Equation 2.1, we obtain ( ) ( ) ( ) ( )AdALdLdRdR −+= ρρ . (2.2) The relative change of the length dL/L of the strain gauge is known as the axial strain,εa, and similarly the relative change of the diameter dD/D of the strain gauge is known as the lateral strain, εL. The ratio of these strains for a particular material is given by its Poisson ratio, ν:

LdLDdD

al //

−=−= εεν . (2.3)

For the relation between the relative change in the cross-sectional area and the relative change in the diameter we find DdDAdA 2= . (2.4)

Combining Equations 2.2 and 2.4 we obtain

ρρνε /)21( dRdR L ++= . (2.5) The gauge factor, G, of a strain gauge is defined as

( ) ARdRG ε/= , (2.6)

so that we obtain the general expression ( ) AdG ερρν /21 ++= . (2.7)


For metals, the resistivity does not vary with strain so the last term in Equation 2.7 can be ignored. The change in the resistance of metals with strain is due solely to geometric effects. However, in semiconductor materials the strain dependency of the last term of Equation 2.7, the resistivity ρ, is significantly larger than the geometrical piezoresistive effect and results in semiconductor strain gauges having large gauge factors. 2.5 The Piezoresistive Effect in Silicon Kulite generally specifies the use of p-type silicon which possesses a very large gauge factor of up to 200, in comparison with n-type silicon which also has a large, though negative, gauge factor down to –140. Equation 2.7 in section 2.4, shows that a change in resistance is generally dependent on a term which has to do with the geometrical piezoresistive effect and a term originating from the strain dependency of the resistivity ρ. In metals this latter term is zero. If a semiconductor bar is stressed, geometrical changes also occur leading to resistance changes according to the geometrical effect. It can be expected that the contribution of this to the gauge factor will be comparable to that measured in metals. Therefore, the large gauge factor in semiconductors can only be due to the strain sensitivity of the resistivity in semiconductors. An explanation involves the theory of the electronic energy-band structure of semiconductors. In classical as well as quantum mechanics the energy E of a particle can be expressed in terms of the mass m and the momentum p. For a cannon ball the kinetic energy is given by 22mvE = (2.8) where m is the mass and v is the velocity. The momentum is given by mvp = (2.9) When the energy E is expressed in terms of momentum p we obtain mpE 22= (2.10) The plot of E as a function of p is a parabola, as shown in Figure 2-2 (a). This curve also applies for free-moving electrons as for example, those in a CRT. When an electron moves in a solid, an interaction of the electron with the periodic lattice of atoms in the crystal can be expected. This interaction leads to the important result that the energy curve is no longer continuous as for the free electrons, but rather it shows discontinuities at certain values of momentum. The graph describing the relation between E and p is shown in Figure 2-2 (a) as a solid line. This result is obtained for a very simple, one dimensional lattice case. However, actual three dimensional lattices show similar discontinuities in the energy-momentum relations.


Figure 2-2: (a) Energy E as a function of the momentum p for a classical particle (dashed line) and as a function of the wave number k for a particle with wave-like nature in interaction with a periodic crystal lattice (solid lines); (b) Part of the possible solutions indicating the conduction, the forbidden and the valence band.

It is usual in quantum mechanics to replace the momentum by the wave-number k. The relation is π2hkp = (2.11) where h is Planck’s constant. As indicated in Figure 2-2 (a), the discontinuities in the energy occur at ank π= , ,....2,1 −−=n and ,....2,1 ++=n (2.12) where a is the lattice spacing. For values of k far from the discontinuity points given by Equation 2.12, the dashed and solid lines coincide, which means that the electrons behave as free electrons. However, when k is close to anπ , certain energy levels are forbidden, which is due to the strong interaction of the electrons with the lattice. The periodic occurrence of forbidden energy gaps has to do with the wave-like nature of electrons moving in a periodic lattice. Figure 2-2 (b) focuses on the central range. For all k values, we obtain two energy levels. Two bands of allowed energy levels occur. Between these a band is obtained of energy levels that are forbidden for the electrons. In a semiconductor the lower allowed energy band could be the almost empty conduction band. The relationship between E and k as depicted in Figure 2-2 (b) only applies for the very simple case of a one-dimensional lattice. For real three-dimensional semiconductors like silicon the situation is much more complex. Detailed band-structure calculations are rather difficult in the three-dimensional case, so use is made of the fact that the crystal lattice is often highly symmetric. Solutions are often found for directions of high crystalline symmetry such as the [100] and [111] directions shown in Figure 2-3. In Figure 2-3 those parts of the band structure of silicon that are relevant for explaining the piezoresistive effect in silicon are shown.

Heavy Hole

Heavy hole

Light hole

Conduction Band

Figure 2-3: Energy Band Structures of Silicon for (a) the [100] and (b) the [111] directions. In order to understand the piezoresistive effect in p-type silicon, we first must understand the concepts of carrier effective mass and mobility. Figure 2-3 shows that near the maxima, the hole bands (lower two bands) in both the [100] and [111] directions are parabolic in shape. This can be directly compared with the case of a free electron (Equation 2.10) which is also parabolic. The hole band parabola is upside down because the hole charge is opposite that of an electron. By making


comparison with Equation 2.10, we define an effective hole mass which corresponds to the curvature of the parabola in the energy band diagram: the tighter the parabola, the lower the effective mass. When an electric field is applied across a semiconductor, the charged carriers, which are holes in the case of p-type silicon, move in the direction of the field. When the electric field is increased the holes move faster. The ratio between the speed of the holes and the electric field is known as the mobility. In a semiconductor with a high mobility, hole moves faster for the same applied electric field. The mobility is related to the hole effective mass just as if the holes were actually heavier: holes with a larger effective mass move more slowly, and therefore have lower mobilities. Thus, the higher the mobility of the charge carriers, the higher the current resulting from the same applied electric field, and therefore the lower the resistivity. So holes in a p-type semiconductor behave similarly to free electrons with the exception that their effective mass is different. Inspection of Figure 2-3 reveals that silicon has two types of holes with different effective masses, known as the heavy holes and the light holes, and it is this variation in effective mass (and hence mobility) which leads directly to the piezoresistive effect. Actually the silicon bandstructure has a third hole band, known as the split-off hole band, which is not pictured in Figure 2-3 because it does not play a significant role in piezoresistance. The average mobility of the p-type silicon is the average mobility of the individual holes in the silicon and therefore is determined by the proportion of heavy holes to light holes. Piezoresistance in semiconductor works by altering this proportion with the application of stress. When an anisotropic stress is applied, the lattice spacing increases in one direction, while the lattice spacing decreases in the perpendicular direction. As one might expect, the interaction of the charged carriers with the lattice is also affected. The stress causes the holes to move from the heavy-hole band to the light-hole band or vice versa, depending on the direction the stress is applied. Since the ratio of heavy-hole to light-holes is altered, the average effective mass of all the holes changes, and therefore the mobility and the resistivity changes. This is the piezoresistive effect in silicon. Recall that the tightness of the band parabola is directly related to the hole effective mass. As is apparent in Figure 2-3, the difference in curvature and therefore the difference in effective mass is much larger in the [111] direction than in the [100] direction. Therefore, when stress is applied to the silicon and the holes redistribute themselves among the heavy-hole and light-hole bands, the effect of the redistribution on the resistivity is small in the case of the [100] directions but quite large for the [111] directions. Therefore piezoresistors in p-type silicon are aligned in the [111] directions. The gauge factor of p-type silicon is positive. This means that a positive strain (elongation) causes the band with the lowest mass and highest mobility to lower with respect to the low mobility band, so that the holes move to the low mobility band. When, as a result more holes have a lower mobility, the resistivity increases, which leads to the experimentally observed positive piezoresistive effect. Figure 2-4 shows how the gauge factor G depends on the temperature and the doping level. Similar to n-type material, the gauge factor decreases for increasing doping concentration and higher temperature. Both effects can be explained by the fact that when fewer holes distribute themselves over the two bands, the relative change is larger. As can be seen in Figure 2-4, the gauge factor is significantly smaller in more heavily doped material, but it is much more stable with temperature. Therefore, Kulite generally processes the silicon to produce degenerative (highly) doped piezoresistive sensors. A little sensitivity in the magnitude of the output of the sensor is lost in order to create a sensor whose output is much more temperature stable. The loss of sensitivity can be made up by carefully designing the diaphragm under the piezoresistance to maximize output.


G

T

Figure 2-4: Gauge Factor of p-Type Silicon as a Function of Temperature & Doping Concentration 2.6 Sensing Elements The sensing elements are the electromechanical conversion devices which convert the displacement of the force summing device into an electrical signal. They must be designed and fabricated so that the conversion is accurate, linear, stable, and exhibits minimum hysteresis. The more intimate the relationship between the sensing elements and the force-summing device, the better these characteristics will be. Sensing elements may be discrete gauges which are applied to some part of the force-summing device, or they may be an integral part of it. 2.6.1. Discrete Gauges Bonded strain gauges are separate strain gauges applied to the diaphragm, or to some other mechanical element, which is strained by the application of pressure. Silicon semiconductor strain gauges were used in bonded strain gauge transducers before the development of the Integrated Silicon Sensor by Kulite in the early 1960s. The primary advantage of the semiconductor strain gauge over the metal wire or foil strain gauge is higher sensitivity at the expense of greater thermal sensitivity and zero shift. Application of a bonded discrete stain gauge involves finding the best location, properly aligning the gauge, and bonding the strain gauge to the structure with an adhesive. The silicon strain gauge is still used in a small number of pressure transducer designs today, but has largely been superseded by the Integrated Sensor design which is described later. 2.6.2. Integral Gauges The diffused piezoresistive transducer uses a silicon element for the mechanical structure, and the strain gauge is an integral part of the silicon element instead of the strain sensitive elements being bonded to the diaphragm as in the past. The silicon integrated chip is itself the diaphragm. Applied pressure presents a distributed load to the diaphragm, which in turn provides bending stresses and resultant strains, to which the strain gauges react. This stress creates a strain proportional to the applied pressure, which results in a bridge unbalance. With an applied voltage, this unbalance typically produces a 100 millivolt deviation at the bridge output, which is proportional to the net difference in pressure acting upon the diaphragm for a supply voltage of 10 volts. The piezoresistors are formed within the silicon diaphragm by either diffusion or implantation of atoms from the third atomic group (e.g. phosphorus which produces an n-type semiconductor) or the fifth atomic group (e.g. boron which produces a p-type semiconductor). By the use of photolithographic techniques, typically four elongated piezoresistors are created. Two of these resistors are positioned on the silicon diaphragm such that they experience a compressive strain and two positioned where they experience a tensile strain. They are then connected together electrically to form a fully active Wheatstone bridge. Figure 3 is an enlarged view of a silicon diaphragm which illustrates the four piezoresistors and the electrical interconnections.


Kulite later developed and patented a variation to the integrated diffused sensor design in which the four piezoresistive gauges are molecularly bonded to a micromachined silicon diaphragm with an insulating layer of silicon dioxide between. This technology is referred to as either “silicon on silicon”, ”silicon on insulator” or “dielectrically isolated silicon strain gauges”. In diffused silicon sensors, the piezoresistors are electrically isolated from one another and the substrate by reverse-biased p-n junctions which leak current at high temperatures. Dielectric isolation, however, allows for operation at much higher temperatures. This single development resulted in the extension of the maximum operating temperature capability from 150°C up to 540°C, an incredible improvement in performance in one step. Kulite’s capability to design, manufacture and package ultra high temperature piezoresistive pressure transducers is still unmatched even two decades after its first implementation. 2.6.3. The Wheatstone Bridge The most popular circuit for use with all types of strain gauges is the Wheatstone bridge. Bridge circuits can be made using from 1 to 4 strain gauges, at least one of which is active (changes resistance with strain). The popularity of the bridge circuit is due to the fact that it converts the strain-induced resistance change of the gauge to voltage changes which can be measured more directly and accurately with conventional instruments. The Wheatstone bridge is normally energised by applying a regulated voltage across two opposite corners. A voltage output proportional to the product of the excitation voltage and the resistance changes of the strain gauges appears across the signal terminals. For conventional wire and foil gauges, the signal level is measured in terms of a few tens of millivolts whereas semiconductor strain gauges typically produce signals of several hundreds of millivolts. The sensitivity of a constant voltage strain gauge bridge circuit is generally discussed in terms of the ratio of the change of signal voltage to excitation voltage for some fixed strain change.

V+ V-Vi

For a four-arm bridge, as shown above, this can be shown to be

+

−+

−∆++∆+

∆+−

∆++∆+∆+

=−∆ −+

43

3

21

1

4433

33

2211

11)(RR

RRR

RRRRR

RRRRRR

RRV

VV

i

(2.13)

Good circuit design dictates that whenever possible, the two adjacent arms of the bridge should change equally but in opposite directions under strain. This will eliminate temperature induced changes from the output voltage. This condition is achieved, for example, in the special case of a fully active Wheatstone bridge circuit where R1 and R4, the tension gauges and R2 and R3, the compression gauges, are equal. For this special case

GRR

VVV

i

ε=∆

=− −+ (2.14)

where ε is the strain and G is the gauge factor. Thus the bridge output voltage is linear with applied strain for any gauge which exhibits a characteristic of

RR∆ versus ε = constant (2.15)


A Wheatstone bridge is relatively easy to signal condition, since it is excited by a constant voltage (or current) and produces a low output impedance (a few hundred to a few thousand ohms) millivolt output signal. All strain gauges, semiconductor and metal, exhibit two temperature-dependent characteristics:

(1) Their resistance changes with temperature, and (2) Their gauge factor changes with temperature. (3) These variations are generally larger for semiconductor gauges than for metal gauges. In

addition, for semiconductor gauges, another factor must be considered. The thermal expansion of semiconductor materials is much lower than those of metals to which the gauges are usually bonded. Thus, as the temperature changes from that at which the gauges were bonded, the gauges are subjected to a thermal strain in addition to load produced strains. Proper circuitry can do much to isolate wanted from unwanted effects to obtain accurate measurements.

Apparent strain is defined as that strain calculated from resistance changes produced by factors other than load induced strains. Principally, it is the combination of the temperature coefficient of resistivity of the semiconductor plus differential thermal expansion effects. It is given by

GCC KMA /TCR)( +−=ε , (2.16)

where εA is the apparent strain/ºF, G is the gauge factor, CM is the coefficient of thermal expansion of material to which gauge is bonded (inches/inch/ºF), CK is the coefficient of thermal expansion of gauge (1.4 x 10-6 inches/inch/ºF), and TCR is the temperature coefficient of resistance (ohms/ohm/ºF).

The simplest technique for eliminating apparent strain effects is the Wheatstone bridge circuit. Either two strain gauges or four may be used. This techniques utilises the fact that the resistance changes of two gauges in adjacent arms of a bridge will subtract if of the same polarity and add if opposite. Thus, if the two gauges are subjected to the same temperature, their apparent strain contributions will cancel. Of course, it is necessary that one of the gauges be unstrained or both gauges be strained in opposite directions to obtain a load responsive signal. 2.6.4. Key Characteristics of Kulite Piezoresistive Technology The dielectrically isolated piezoresistive technology of Kulite’s pressure sensor has the following strengths which are summarised below:

• Increased reliability – the piezoresistive gauges are molecularly bonded to the diaphragm. • High gauge factor. • High frequency response – the natural frequencies of the diaphragms is 150 kHz minimum. • No hysteresis – the silicon diaphragm has a single crystal structure. • Not susceptible to electromagnetic interference (EMI) as there are no P-N junctions (unlike the

diffused design of piezoresistive technology) • Not electrostatic discharge (ESD) sensitive. • Capable of operation up to 1000°F/ 540°C & down to cryogenic temperatures. • Mature sensor technology with a track record of reliable performance

The active area of the pressure-sensing surface, which is made of silicon, is less than 0.3 mm square. Key to the performance and ruggedness is Kulite’s unique sensor design which incorporates a four-arm Wheatstone bridge molecularly bonded to, but electrically isolated from, the silicon diaphragm. Over many years, Kulite has developed computer aided design tools to design micromachining patterns in the silicon diaphragm which concentrates the stress at the locations of the resistive elements and produces a very linear electrical output against applied pressure characteristic. Additionally, the micro-machining of the silicon diaphragm produces a higher sensitivity for a given diaphragm resonant frequency as well as increasing the robustness of the sensor. The micro-machining process also enables the dimensions of the central boss to be adjusted so as to contact the supporting


Pyrex glass pedestal at a predetermined pressure and provide a dramatic increase in over pressure protection (typically from x3 without stopping to x30 with stopping). In the 1960s and 1970s, Kulite silicon diaphragm pressure transducers were mostly known as dynamic-only pressure measuring devices. They have now gained full acceptance as static pressure measuring devices and typical static error bands (nonlinearity, hysteresis and nonrepeatability) are 0.25% of full scale output, and better than 0.15% when using digital compensation techniques. Kulite piezoresistive silicon on insulator (SOI) technology pressure transducers are currently used in applications where high reliability and high accuracy at an affordable price are required in the aerospace, industrial, automotive and oil industries. Because of the extremely small size of the sensing element (1.6mm x1.6 mm x 0.5 mm), Kulite pressure transducers offer more flexibility in packaging than any other technology. Typical aerospace applications are model and full-scale wind tunnel tests, flight tests, brake and hydraulic system tests, jet engine fuel system tests, and other measurements of turbulent flow. Automotive applications include engine air, oil, cooling and fuel systems, brake systems, transmissions and general laboratory pressure measurements. Kulite amplified pressure transducers are used by the majority of the Formula 1 racing teams in Europe both on the engines and the chassis. Ultra miniature Kulite pressure transducers are used in the wind tunnel testing of the chassis. 2.7. Microphones Microphones are very sensitive pressure transducers which are calibrated in terms of Sound Pressure Level (SPL) rather than common pressure units. Kulite models MIC-062, MIC-093, MIC-152 and MIC-190 are designed as microphones for high-intensity sound measurements. SPL is expressed in dB or decibel notation. "Decibel" for pressure levels, voltages, accelerations, and similar measurements is defined by

dB = 20 log10 P1 / P0. (2.17) Where P1 is the pressure being characterised and P0 is the reference pressure. By international agreement, reference pressure for SPL is 0.000 02 N/m2 (pascal) or 2.9Qx 10-9 psi. Note also that pressures for SPL are always rms pressure levels. 2.8. Integrated Sensor Design The integrated sensor design uses a silicon diaphragm, as the description suggests, as the basic force-collecting and sensing mechanism. The diaphragms for the integrated sensors are fabricated either by the techniques of solid state diffusion and oxide masking, epitaxial growth or, more recently, Diffusion Enhanced Fusion (DEF) bonding of two discrete wafers (pattern and carrier) to produce a strain sensing network that is electrically isolated from the silicon substrate that forms the force collector. Each diaphragm contains a fully active four-arm bridge with the gauge positions chosen to provide; tension and compression. Details of the basic fabrication processes including diffusion, oxide masking, photolithography, contact metallisation and lead attachment as well as the fabrication processes for DEF bonded units are not described in detail in this handbook but are generally available in previously published Kulite technical papers. These devices represented a state of the art advance in the field of miniature pressure transducers when they were first developed by Kulite in the 1960s and have found wide application in such fields as jet engine testing, wind tunnel testing and flight testing for the measurement of high frequency pressure fluctuations. The design of a successful silicon diaphragm pressure transducer requires a detailed understanding of the physics of piezoresistivity and also the relationship between the change in resistivity and the stress inducing the change. Factors such as the type of material to use for a particular application (P or N-


type), the physical form of silicon (monocrystalline, polycrystalline or amorphous silicon), the dopant to be used and the concentration required to fabricate the piezoresistors in the silicon and the orientation of the crystal planes relative to the transducer diaphragm all need to be considered by the designer. In order to construct a pressure transducer from the silicon diaphragm as discussed in the preceding sections, it is necessary to support the silicon diaphragm on a pedestal in such a manner as to enable a pressure differential to be applied across the diaphragm without introducing a mounting strain in the diaphragm. To complete the manufacture of the pressure capsule, a Pyrex glass or silicon pedestal is electrostatically bonded to the silicon sensor. The purpose of the pedestal is to:-

• Mechanically support the silicon sensor • Isolate the sensor from stress • Provide an overload stop when required • To configure the sensor reference pressure (absolute, gauge, differential)

The figure below is a schematic diagram of an absolute pressure capsule diffused integrated sensor.

Figure 2-7: The Diffused Integrated Sensor The use of silicon based components throughout the construction of the pressure capsule i.e. n-type silicon diaphragm, p-type silicon piezoresistors, low expansion “Pyrex” glass results in a mechanical assembly in which the coefficients of expansion of all the components are very closely matched. This aspect of the design ensures minimum internal stresses are generated when the temperature of the pressure capsule is changed and leads to enhanced long term stability. The construction of the pressure capsule also enables semiconductor manufacturing techniques to be employed which produces a very small sensor with a diaphragm which has a very high natural frequency. This enables the sensor to measure both static and dynamic pressures with high accuracy. A dramatic comparison between the relative sizes of an integrated silicon diaphragm capsule (left) and a strain gauged metal diaphragm (right) is shown below.

Figure 2-8: Integrated Silicon Diaphragm v Bonded Strain Gauge Design


2.8.1. Dielectrically Isolated Design (Silicon on Silicon) The diffused integrated sensor design described in section 2.7 was incorporated into the majority of Kulite pressure transducers and enabled the manufacture of ultra miniature pressure transducers with exceptionally high natural frequencies. However, the diffused design of integrated sensor has two major drawbacks. The p-n junction between the p-type piezoresistors and the n-type silicon diaphragm has a resistance characteristic which falls rapidly with temperature. Above temperatures of 150°C to 200°C, the leakage currents within the Wheatstone bridge circuit have effectively bypassed the piezoresistors and the bridge is no longer measuring pressure accurately. In addition, the p-n junctions display a photosensitivity i.e. they are light sensitive. For applications where dynamic pressures are required to be measured in the presence of a detonation or other luminous events, the transducer will generate an electrical output due to the light and which is totally separate from the pressure changes. Many applications to measure pressures within the aerospace, oil and automotive industries require transducers to operate in a temperature environment in excess of 150°C. In order to overcome these limitations of the original p-n junction type sensor, the dielectrically isolated (or silicon on silicon) sensor has been developed by Kulite in which the piezoresistors are electrically insulated from the n-type diaphragm material by the interposition of a layer of silicon dioxide alone or, more recently, in combination with a second layer of glass. A schematic diagram of a dielectrically isolated fusion bonded sensor `together with a photograph of a processed diaphragm is shown below.

Figure 2-9: Schematic Diagram & Photograph of Silicon on Insulator Sensor 2.8.2. Diaphragm Characteristics In order to fabricate integral silicon diaphragms containing a four-active arm Wheatstone bridge and to ensure that both tension and compression gages are fully active, careful attention must be given to the stress distribution across the diaphragm and its effect on gauge placement as shown in Figure 2-10.


Figure 2-10: Connection of Resistors in Wheatstone Bridge

Figure 2-11 shows a classical clamped-edge flat diaphragm with the resulting normal and surface stress distributions shown in Figure 2-12. Use of a diaphragm with this stress distribution would result in a pressure transducer with a very non-linear electrical output with respect to applied pressure. The addition of a specially designed thickened area in the centre of the diaphragm, as shown in Figure 2-13 which Kulite refers to as a “boss”, produces the stress distributions shown in Figure 2-14. The stresses in the surface of the diaphragm are designed to vary equally in the compressive and tensile directions which results in an ideal characteristic for the measurement of pressure. The sensing network is located, in the thin (active) portion of the diaphragm. When the diaphragm deflects under pressure, the surface stress in the region nearest the unetched clamp region is of one sign and the stress next to the boss is of the opposite sign: Therefore, if a sensor network is disposed such that one sensing gauge is adjacent to the clamped edge and the other, sensing gauge is disposed adjacent to the boss by connecting the two gauges together one gauge will increase in resistance and the other will decrease. If the same technique is used on the other side and the gauges are interconnected properly a fully active Wheatstone bridge circuit will result. If a gauge or differential pressure is to be measured, a vent hole is drilled through the pedestal to enable barometric or a reference pressure respectively to be applied to the back of the diaphragm.


Edge to Center DistanceEdge Center

Stress0

Figure 2-11: Clamped-Edge Diaphragm ress Distribution for Figure 2-12: St Clamped-Edge Diaphragm

Edge to Center DistanceCompressive Stress

Tensile Stress

Zero Stress

Figure 2-13: Bossed Diaphragm Figure 2-14: Stress Distribution for

Bossed Diaphragm

Edge to Center DistanceCompressive Stress

Zero Stress

Surface

Normal

Figure 2-15: High Pressure Diaphragm Figure 2-16: Stress Distribution for

High Pressure Diaphragm

The pressure range of a diaphragm is determined by the thickness of the silicon between the central boss area and the clamping area at the edge of the diaphragm. For diaphragms which are designed to operate at high pressures, the diaphragm thickness can approach the radius of a circular diaphragm as shown in Figure 2-15. The resultant stress distributions are shown in Figure 2-16. If the surface stresses are measured as in the lower pressure designs, it can be seen that there is no point on the diaphragm where there is a positive (tensile) stress. This inability to generate negative and positive going gauges for inclusion into a Wheatstone bridge circuit will produce a severely non-linear response to applied pressure. By careful analysis of the stress distributions within a high pressure diaphragm design and by exploiting the surface stresses normal to the diaphragm by the appropriate selection of gauge material and crystallographic orientation, it is possible to produce a gauge which has a positive resistance change. The reference to a detailed report on high pressure transducer designs is to be found in Section 9.3.1.


To ensure optimal performance the etched regions must be very narrow (in the order of 0.010 inches) and the individual gauges must be very short (in the order of 0.001 to 0.002 inches) and extremely narrow (in the order of 0.0001 inches). For low and medium pressure range diaphragms, the gauges are best made such that the gauge length is in a longitudinal direction i.e. perpendicular to the wall of the edge of the clamped region and of the boss. The longitudinal crystallographic direction being in the <110> direction and the transverse in the <100> direction. Use of the Kulite developed DEF bonding technique has produced a group of transducers which can operate up to temperatures well in excess of 1000°F with excellent linearity and thermal characteristics. 2.8.3. Isolated Capsule Design Frequently it is a requirement of airworthiness organisations that pressure transducers which are to be used on board aircraft should have some form of secondary pressure containment incorporated within the basic transducer design. Secondary pressure containment is defined as the ability to contain the pressure media if, for some reason, the primary pressure containment (the silicon diaphragm in the case of silicon based transducers) fails. This is achieved typically by mounting the silicon sensor capsule on a 304 stainless steel header with high-pressure glass-to-metal sealed pins. An isolation diaphragm of 304 stainless steel is electron-beam welded to the front surface of the header and the cavity between the isolation diaphragm and the sensor capsule is filled with silicon oil which acts as a pressure transfer medium. This form of construction is shown schematically in Figure 2-17. The use of the header-silicon capsule-isolation diaphragm design combines the advantages of a metal diaphragm and an all welded construction with the advantages of an inorganically bonded silicon sensor for excellent repeatability and long term stability. The header assembly is frequently welded into an outer case which incorporates the electrical connector. This outer case provides tertiary containment and ensures that the transducer is totally hermetically sealed, which considerably enhances the reliability and survivability of the transducer in humid and contaminated environments.

Figure 2-17: Isolated Pressure Capsule Design

2.8.4. Dual Diaphragm Technology For the measurement of gauge or differential pressures, Kulite has patented a design of pressure transducer which uses two separate, oil filled, absolute pressure capsules (Kulite Patent #4,695,817). This method was developed to eliminate the historical problems of having gauge pressure transducers which were typically vented either through the case, which exposed the internal electronics and components to environmental contamination and frequently resulted in a very short life, or via small pressure tubes which tended to become clogged or filled with water which could freeze. These advantages are also relevant to differential pressure transducers using two pressure capsules. To maintain acceptable levels of accuracy for the two pressure capsule design of differential pressure


transducer, a line pressure limitation of a maximum of 4 times the differential pressure range is generally applicable. This design of gauge or differential pressure transducer produces a transducer, the internals of which are totally hermetically sealed from the external environment leading to exceptional levels of reliability, stability and operational life. Each pressure port of a two capsule differential pressure transducer is associated with a separate measuring element. In the case of a gauge unit, one pressure port is vented to atmosphere to provide the reference pressure. One measuring element is a half bridge array and monitors the main pressure applied to the positive port. The second element is also a half bridge array and monitors the pressure applied to the other port, which is the negative, or reference port. A fully active Wheatstone bridge is formed by two piezoresistors associated with the positive port and two piezoresistors associated with the negative port. By selecting the magnitudes of the resistors associated with each port it is possible to generate one electrical signal which is an analogue of the differential pressure.

The preferred circuit is shown below.

Negative Port Positive Port

- ve Signal + ve Signal

+ ve Excitation

-ve Excitation

Figure 2-18: Schematic Circuit Diagram of Dual Capsule Design

2.8.5. Combination Pressure/ Temperature Transducers. As a result of the small size of the isolated pressure capsule design, which is due in large part to the incorporation of Kulite’s ultra miniature pressure capsule design, pressure transducers can be designed and manufactured which may have two or three pressure sensing capsules inside. Kulite pressure transducers can also be supplied with one or more resistance temperature sensors included within the housing. Such pressure transducers are referred to as “Combination Pressure or Pressure/ Temperature Sensors”. Combination sensors may also be used to provide a redundant capability where two totally independent transducers, with separate electrical connectors within one body, are used to measure the same pressure, an example being the engine oil feed pressure which is critical for the health of a gas turbine. Combination sensors are designed and built for high reliability/ availability and extreme


service environments and feature small size and minimum weight, critical features in aircraft and many other applications. Combined pressure and temperature measurement will give a better indication of process status or fluid health. Redundant sensors give an extra measure of reliability, and maintain process control in the event of a single sensor failure. The installation of redundant or multiple sensors in a single penetration or package makes engineering and installation easier and significantly reduces the weight of the installation. Reducing sensor penetrations and wiring harnesses also decreases installation and life cycle costs. Combination sensors have been developed exclusively by Kulite and are finding many applications in both the Aerospace and Autosport Industries. References to a paper which describes Combination Pressure /Temperature sensors in more detail are given in section 9.3.5. 2.8.6. Kulite Leadless Design Electrical connection is made to the Wheatstone bridge on the silicon diaphragm using four or five 0.024mm (0.001 inch) diameter gold bond wires which are ultrasonically ball bonded to the diaphragm metallisation. The pressure media is in direct contact with the stress-sensing network, leadouts and interconnects which, at high temperatures, in the presence of aggressive chemicals and after prolonged exposure, can deteriorate and fail. The key elements in the design of a ruggedised pressure sensor are the elimination of the gold bond wires and the protection of the sensing elements from corrosive environments at high temperatures, hence the reference to the new sensor capsule as the “leadless” design. The leadless sensor capsule is comprised of two main components, the sensor chip and the cover as shown below in Figure 2-18. The sensor chip and the cover wafer are assembled using an electrostatic bond to form the sensor capsule. Once the two wafers have been bonded, only the metallised leadout pads are exposed whilst all the gauges and electrical interconnections on the sensing side of the silicon chip are sealed by the cover. Thus the active portion of the pressure sensor is hermetically isolated as shown in Figure 2-19.

Figure 2-18: Sensor Chip & Cover Before Bonding Figure 2-19: Sensor Chip Bonded to Cover To avoid the use of gold ball bonds and fine gold wires, a high temperature conductive glass is used to provide the electrical connection between the sensing chip and a specially designed header. The pressure capsule is bonded to the header at a high temperature using a non conductive glass , during which process the conductive glass in the cover wafer holes melts and creates low resistance electrical connections between the header pins and the metal contact pads on the sensor chip as shown in Figure 2-20.


Figure 2-20: Pressure Capsule Bonded to Header

After this firing process, only the non-active side of the diaphragm is exposed to the pressure medium. The small ball bonded gold leads have been eliminated and the entire sensor network and contact areas are hermetically sealed from the environment and the pressure media. 2.8.6.1 Leadless/ Acceleration Compensated Design There are many environments which are very harsh in which to attempt fast response measurements of static pressure, particularly on rotating components. Because of the high rotational speed of most turbomachines, pressure sensors can be exposed to high levels of centrifugal and vibrational acceleration. As has been described, semiconductor pressure sensors function by determining the deflection of a small silicon diaphragm under exposure to a normal stress (pressure), using a Wheatstone bridge network of strain gauges to measure this movement. However, the diaphragm will also deflect under the influence of centrifugal and vibrational accelerations which will generate both offset errors and dynamic errors. Kulite has designed an acceleration insensitive semiconductor pressure sensor which compensates for these deleterious effects and is shown below:-

Figure 2-21: Acceleration Insensitive Pressure Sensor

The device is based upon the leadless technology described in section 2.8.5 and comprises two pressure sensing diaphragms which are manufactured on one silicon chip. On each diaphragm, a half Wheatstone bridge is formed using two piezoresistors in series. The two diaphragms are both exposed to the inertial stresses (vibration and centrifugal acceleration), but only one is exposed to the pressure


to be measured. The two half bridges from each diaphragm are electrically connected to form a full bridge such that for a positive stress applied substantially normal to the diaphragm, the bridge output from one half-bridge will subtract from the other. Thus the signal output is responsive to the pressure as applied to one diaphragm while the signal response to inertial stresses (or any other stress other than that due to pressure) applied to both diaphragms is cancelled out. Reference to a comprehensive report on the acceleration insensitive pressure sensor in given in section 9.3.11.

2.8.7 Temperature Compensation Semiconductor strain gauge characteristics are temperature dependent. In particular, the resistance of an semiconductor strain gauge increases with temperature by typically +10% per 100 degrees F rise in temperature i.e. the Temperature Coefficient of Resistance (TCR) is +10%/100°F. The strain sensitivity, or gauge factor, decreases with temperature by typically –2% per 100 degrees F increase in temperature i.e. the Temperature Coefficient of Gauge Factor or Sensitivity (TCGF or TCS) is –2%/ 100°F. To properly utilise semiconductor gauges for accurate measurements of mechanical strain, it is necessary to compensate the gauge output signals against these undesirable temperature effects. Compensation can generally be accomplished with simple circuit techniques using passive shunt or series resistor elements whose resistance is temperature independent. Piezoresistive strain gauge bridges must be compensated for zero and zero shift with temperature and for the decrease in sensitivity by means of adjustment of the bridge excitation voltage. Relations are derived for calculating the values of the compensation resistors from a knowledge of the strain gage parameters and the measured effects of temperature on signal output. 2.8.7.1 Bridge Zero and Zero Shift Compensation Placing a near zero temperature coefficient of resistance (TCR) resistor in series or shunt with a piezoresistive gauge changes the magnitude of the resistance and TCR of the gauge.

Figure 2-22: Effect of Zero Compensation Resistor Placement Kulite has developed computer based models to calculate the magnitude and location of compensation resistors to nullify the effects of temperature changes on both zero and zero offset values for half bridges, 4-wire and 5-wire full bridges. 2.8.7.2 Bridge Sensitivity Compensation Bridge sensitivity compensation is achieved by the insertion of a span resistor connected in series with the bridge supply voltage.


Figure 2-23: Wheatstone Bridge with Span compensation

(2.19) The equation for the output voltage from a Wheatstone bridge equipped with a span resistor is: The bridge input resistance, Rbridge(T), increases with temperature at it's TCR value. the bridge sensitivity, S(T), decreases with temperature at it's TCS value. By observing the above equation it can be noted that a correct choice of Rspan can compensate the sensitivity of the transducer over temperature. A simple qualitative explanation as to how the span compensation process works is that as the temperature of the bridge increases, the resistance of the bridge also increases. However, the resistance of the span compensation resistor Rspan does not change with change in bridge temperature. Thus as the temperature of the bridge increases, an increasing proportion of the supply voltage will be applied to the bridge which will proportionally increase the bridge output. The resistance of Rspan is chosen such that the increase in bridge output with temperature is exactly offset by the decrease in sensitivity of the bridge with temperature. Because of variations in material properties, processes, and dimensions, the performance of a population of pressure transducers of a given design will scatter about the typical. To provide the lowest effect of temperature, the performance is measured for each transducer during the manufacturing process, and resistance values are chosen to compensate for the changes with temperature. The resistance Rspan in series with the bridge supply is used to reduce the sensitivity variation with temperature. The resistances in series and parallel with one arm of the bridge correct for bridge zero and changes in zero (zero shift) with temperature. The temperature compensation resistors are mounted within the transducer case for the majority of Kulite pressure transducers. Only when the pressure capsules are used without any external casings (chip-on applications) or with ultra high temperature pressure transducers (temperatures in excess of 400°C) are the compensation resistors located remotely from the sensor. This arrangement is possible due to the fact that Kulite pressure transducers only require zero temperature coefficient resistors in their temperature compensation networks, unlike many other manufacturers who also use temperature dependent resistors (thermistors) in their “active” temperature compensation networks. In order to achieve accurate temperature compensation, the temperature of the thermistor must be the same as the temperature of the sensor. Temperature gradients within an actively temperature compensated pressure transducer can significantly degrade the accuracy of pressure measurement, errors of up to 10% full


scale being possible during thermal transients. Each pressure transducer is tested in the manufacturing process and the resistors are selected to optimise performance. 2.8.8. Mechanical Design Kulite’s unique range of sensors are packaged in stainless steel cases of an infinite variety of sizes and shapes. The pressure capsules are mounted in headers which isolate the silicon diaphragms from strains in the casings which may effect the measurement accuracy. Popular designs of miniature pressure transducers include those with threads from sizes M5 to M10 and those with no external threads in diameters ranging from 1.7mm to 3.8mm. The cases for transducers for aerospace applications are frequently designed to meet a specified space envelope. Transducers may incorporate power supplies, regulators and amplifiers with voltage, current or frequency outputs. Cases may be made from titanium, Inconel or Hastelloy in place of stainless steel and are frequently hermetically sealed to protect the unit from external contamination. Kulite’s pressure transducers will operate from -200°C to over 530°C. The silicon diaphragm of many of Kulite’s miniature pressure transducers is mounted at the end of the transducer which enables them to be used in flush diaphragm applications. A protective “B” or “M” screen is usually specified for these units and is designed to have a minimum effect on frequency response. At the front end, the sensing module is isolated from strains in the case, yet it is mounted at the extreme front of the transducer making it equivalent to a flush-mounted diaphragm. Most models include a standard protective screen, designed and tested to minimise effects on frequency response, while providing maximum protection to the silicon diaphragm. At the rear of Kulite’s pressure transducers, the cable is securely anchored inside the case, and sealed with a strain-relieving boot. The vent tube (on gage pressure models) is securely anchored and may be cut, bent or adjusted for specific applications. It may be connected to a reference source for differential pressure measurements. Alternatively, a multi-pin electrical connector may be welded to the case in place of a cable exit an strain relief. For the reference port of Kulite’s miniature differential pressure transducers, the pressure media must generally be clean, non-conductive and non-corrosive. In the case of Kulite’s larger aircraft or industrial pressure transducers which either use Kulite’s patented two sensor design or a true differential “wet- wet” design, virtually all pressure media which are compatible with stainless steel may be safely used. The reference port is the low pressure side in all differential measurements. Differential transducers are designed for specified maximum line pressure. Maximum reference pressure, and maximum case pressure should be (!) specified on the data sheets. 2.8.9 Silicon Carbide Increased performance requirements for pressure transducers for aircraft and for spacecraft demand sensing capabilities at high temperatures. The pressure environments to be measured in these applications require sensing capabilities down to 25 psi. To meet the high temperature requirements, silicon carbide (SiC) has been selected by Kulite as a semiconductor material to be used in fabrication of the sensor chip. SiC material, because of its wide bandgap (3eV), high breakdown electrical field (2.5 x 106Vcm-1) and large piezoresistive coefficients, exhibits excellent thermal, mechanical and electrical characteristics as a sensing material. Kulite have reported on the fabrication, packaging, and testing of a low-pressure 6H-SiC piezoresistive pressure sensor (25_psi range) operational at 600ºC. Sensor fabrication was done using a combination of electrochemical etching and Deep Reactive Ion Etching (DRIE). The sensor is similar in structure with a 1000_psi pressure sensor that was previously reported. The 1000_psi sensor had a diaphragm of about 60_µm thickness, while the 25_psi range required a significantly thinner diaphragm. The sensor reported has a diaphragm of about 20_µm thickness. Fabrication of thin 6H-SiC diaphragms is difficult, because of challenging control of SiC etch depth, and because of the presence of 6H-SiC micropipe defects. These defects are inherent to


currently commercially available 6H-SiC wafers, and their detrimental effect on device yield is increased for thinner diaphragms. As opposed to alternative devices the sensor described has the piezoresistors and diaphragm fabricated from SiC. The utilization of a SiC diaphragm makes the sensor suitable for higher temperature applications (due to excellent mechanical properties of SiC at very high temperatures) and for harsh environments (due to SiC chemical inertness). References to this paper are given in section 9.3.2.

n+ 6H-SiCNPp+ 6H-SiC

Piezoresistors

Diaphragm

Metal

n-typebulk 6H-SiCN

Figure 2-24: Section of SiC Sensor Chip In a departure from Kulite’s preferred use of piezoresistive technology, a paper has been written which describes a silicon carbide dual-resonant-beam pressure sensor capable of operating at ultra-high temperature. Silicon begins to plastically deform above approximately 6000C. However, many applications, such as those associated with combustion in gas turbine engines, require transducers capable of operation at much higher temperatures. At such high temperatures, silicon carbide is the material of choice due to its high temperature of plasticity, large bandwidth and chemical inertness. The device is composed of two pressure-sensing diaphragms each spanned by a beam that is caused to vibrate at its respective resonant frequency. One diaphragm is exposed to the applied pressure, which induces stress in the beam that spans it and therefore alters the beam’s resonant frequency. The other diaphragm is not exposed to pressure; thus its beam’s resonant frequency remains unchanged. The difference in the frequencies of the two beams is then directly proportional to the pressure that is to be measured. As the output of sensor is a frequency, interfacing the signal with a digital system is simple. Because of the close physical proximity of the two diaphragms and beams, any measurement errors induced by such external variables as temperature or acceleration are cancelled out when taking the difference frequency. A sectioned view of the dual resonating SiC beam pressure sensing structure with beat frequency output is shown in Figure 2-25.

Figure 2-25: Cross-Section of a Dual Resonating SiC Beam Structure With Beat Frequency Output References to this paper are given in section 9.3.3.


Section 3 – Performance Characteristics 3.1. Dynamic Range 3.1.1. Definition Dynamic range is the measured values over which a transducer is intended to measure, specified by upper and lower limits. The lower limit, when dynamically (ac) coupled, is a few microvolts of noise generated by the silicon gauges and other internal components. When measuring statically (dc coupled), the lower limit will be determined by the zero measurand output and the long term, very low frequency thermal zero drift. 3.1.2. Range The range of the pressure transducer specifies the recommended maximum peak pressure level for optimum linear response. Most Kulite pressure transducers maintain good linearity up to 3 times the range, and specifications are provided for extended range. This is intended as a safety margin, not for normal use. 3.1.3. Overrange Above the range, nonlinearity increases, but the transducer continues to operate. As a single-degree-of-freedom system the mechanical response of the diaphragm to an applied pressure is frequency-dependent. Do not apply full scale pressure at frequencies above 30% of resonance frequency. This may excite the diaphragm resonance and, cause erroneous data or lead to diaphragm failure. A pressure "snubber," in the form of a small orifice, may be used to attenuate high frequencies and pressure spikes.

Figure 3-1: Instrument System Operating Range

3.2. Sensitivity The sensitivity of a transducer is defined as the ratio of its electrical output to its mechanical input. In the case of piezoresistive pressure transducers, it is expressed as voltage per unit of pressure at the rated excitation. Units of millivolts per psi (mV /psi) are used because Kulite pressure transducers are calibrated and recommended for operation at a specified and fixed excitation voltage of 10.00 volts dc. 3.2.1. Sensitivity Calibration Each Kulite transducer is provided with a sensitivity calibration as measured by a readout device with a high input impedance (loading effects are discussed later). The transducer is operated at rated


electrical excitation. The sensitivity is expressed in mV /psi and is numerically equal to root mean square (rms) mV per rms psi and peak mV per peak psi 3.2.2. Polarity For many measurements, it is necessary to know the polarity of the system output signal relative to the direction of pressure on the transducer. To determine this, the polarity of the transducer output and the input-output phase relationship of the amplifier must be known. Unless otherwise specified, all Kulite pressure transducers produce a positive output signal when the pressure increases. Polarity of the excitation voltage must be applied in accordance with the specifications on individual transducer data sheets. Kulite maintains standard strain gage practice with colour codes of red for positive excitation, black for negative excitation, green for positive output signal, and white for negative output signal. 3.3 Nonlinearity, Hysteresis & Nonrepeatability 3.3.1 Definitions 3.3.1.1 Linearity Non-linearity (sometimes called linearity) is defined as the maximum deviation of the calibration curve (output vs. input) from a specified straight line, expressed as a percent of full scale output, and measured on increasing measurand only. 3.3.1.2. Hysteresis Hysteresis is defined as the maximum difference between output readings for the same measurand value, one point obtained while increasing from zero and the other while decreasing from full scale. The points are taken, on the same continuous cycle. The deviation is expressed as a percent of full scale. 3.3.1.3. Non-repeatability Non-repeatability (sometimes called repeatability) is defined as the ability of a transducer to reproduce output readings when the same measurand value is applied to it consecutively, under the same conditions, and in the same direction. It is expressed as the maximum difference between output readings as a percent of full scale. 3.3.2. Non-linearity Although a piezoresistive transducer is theoretically linear down to zero pressure, a practical lower limit is imposed by its noise level. As in all electrical conductors, the thermally-induced random motions of free electrons cause noise; in addition, the current flow through the diffused gage elements causes some additional noise having the characteristics of Schottky, or shot, noise. As a result, both

Figure 3-2: Input to Output Curve


the diffused and SOI pressure transducers have a wide band noise characteristic of about 5 microvolts RMS, measured at 20°C. This corresponds to about 1 x 10-4 psi for a 2 psi full scale transducer. Because this noise level is very small, the lower limit of dynamic range is usually a function of the noise characteristics of the signal conditioning and power supply equipment used with the transducer. Single crystal silicon is a very good spring material, having essentially no plastic zone to its stress-strain curve and very low hysteresis. Because the input pressure to these transducers is supported only by the silicon element, these transducers become highly non-linear before burst is reached. Although each transducer is identified with a particular full scale range, there is no absolute end to the scale (with the exception of burst). One may elect to use a transducer at some pressure above full scale, or well below full scale, depending on the requirements of the application. Each transducer is tested prior to shipment to a maximum limit for combined linearity and hysteresis to the "defined" full scale level, and for operation to a specified overrange level, typically 2 times full scale. The linearity plotted below and which is shown on the specifications for the transducers is the "independent linearity". This is defined as the maximum difference between the calibration point and the linear regression line (least squares fit) drawn through the points for increasing measurand, zero to + full scale. Numerically, this is usually about one-half the value when using an end-point, or terminal based, linearity definition.

Figure 3-3: Independent Linearity Curve

3.3.3. Hysteresis Because of the excellent elastic characteristics of silicon, the hysteresis of these gages is usually very small, most of the time under 0.1% of full scale, and often as low as 0.03%. As such, the specifications have simply been stated by indicating typical values for linearity and hysteresis, and then indicating a maximum limit for the two combined.

Figure 3-4: Transducer Hysteresis


3.3.4. Non- repeatability Non-repeatability (sometimes repeatability) is the ability of a transducer to repeat output readings when the same pressure is applied to it consecutively under the same conditions, and in the same direction as shown below. It is expressed as the maximum difference between output readings as a percent of full scale output (%FSO). Two calibration cycles are used to determine non-repeatability.

Figure 3-5: Transducer Non-repeatability 3.3.5. Combining Nonlinearity, Hysteresis, and Non- repeatability Combined nonlinearity, non-repeatability, and pressure hysteresis is the maximum RSS (root-sum-square) average of the three independent parameters discussed above. This is the “total error band" calculated as the RSS average of three independent measurements. 3.4. Zero Measurand Output (ZMO) This characteristic is often called zero balance, zero offset, or zero pressure output. Zero Measurand Output is expressed in millivolts at the output of the transducer under room conditions with full rated excitation, but with no pressure applied to the transducer. For an absolute pressure transducer, this means the output when measuring an absolute vacuum. Although the resistance elements in the bridge of a transducer are closely matched and compensated during manufacture, slight differences in resistance will exist. The differences result in a small offset or residual dc voltage at the output of the bridge. This residual voltage is called Zero Measurand Output. Circuitry within associated signal conditioning instruments typically provides compensation or adjustment of the electrical zero. 3.4.1. Mounting Effects Zero offset can be increased by improper transducer mounting of ultra miniature pressure transducers. Any stresses placed on or near the diaphragm will result in changes in the zero offset. However in the case of Kulite threaded transducers, over-tightening has no effect on the zero measurand output and will usually result in physical damage to the threads or the transducer body. Threaded devices have a recommended installation torque specified on the calibration sheet. 3.4.2. Warm-up Warm-up time is the period of time, from application of excitation voltage, required to assure that the transducer will perform within all specified tolerances. The zero offset will move to its final value while the pressure transducer is being "warmed up." Kulite’s unique diaphragm design provides very fast warm-up stabilisation. Kulite’s pressure transducers typically have warm-up times of one millisecond or less to achieve less than 1 % deviation from long term performance. This characteristic enables the power supply to a pressure transducer to be turned on only for the duration of the measurement which, for battery powered applications, is advantageous as it maximises battery life. 3.4.3. Thermal Stability Since both zero measurand output and sensitivity change with temperature, a stable temperature environment assures the most stable measurements. However, Kulite’s design has been shown to have only very small output shifts even under severe conditions.


When making dynamic measurements, the output of the pressure: transducer can be ac coupled to the signal conditioner. This completely eliminates the zero offset, greatly reduces thermal zero shift, and provides a controllable high pass filter. Of course, static or steady state measurements are no longer possible. 3.4.4. Effect of Overpressure Kulite pressure transducers will survive overpressures of up to 2x full scale without any measurable change in calibration. Burst pressure specifications are provided on the data sheets, and are discussed elsewhere in this text. Because the Kulite silicon diaphragms are very elastic until they fracture, if they are not broken, it is unlikely that the transducer has been damaged. 3.5. Phase Shift The transducers themselves are very lightly-damped dynamic systems with essentially no phase shift. Therefore, phase shift of the transducer is negligible. However, the damping characteristics of the measured medium will affect the response of the diaphragm. Also, as discussed previously, connecting plumbing may have damping effects between the measurement point and the transducer, thus causing some phase shift. Electrical characteristics of the transducer's integral compensation and balancing components are all resistive, so they cause no phase shift. Typical signal conditioning will have either 0 or 180 degree phase shift, depending on how many amplifier stages are included and their characteristics. 3.6. Input and Output Resistance The primary uses of these specifications are to calculate excitation current requirements and to assure that the bridge is not open or shorted. The input and output resistances of piezoresistive pressure transducers are specified on the individual data sheets. For an equal-arm four-element Wheatstone bridge, the input and output impedances are equal and are in the order of 1,000 ohms. However, temperature compensating and zero balance resistors are connected in series with the sensing elements. These additional resistors will usually result in slightly differing input and output resistance. These full-bridge transducers have series resistors for thermal sensitivity compensation located external to the bridge, so that input resistance is approximately 1.5 to 2 times the output value. For best results, the readout instrumentation input impedance should be at least 1 Megohm. For an impedance of 20 times the output impedance of the bridge, the sensitivity is reduced by 5%. With an input impedance of 50 times, the reduction is 2% and an input impedance of 100 times yields a 1% reduction. Actual values of input and output impedance are recorded on each calibration sheet and are a very simple and convenient means of verifying the health of any transducer. 3.7. Thermal Sensitivity Shift and Zero Shift Thermal sensitivity shift and thermal zero shift define the effects on sensitivity and ZMO of operation at operating temperatures other than a normal ambient temperature 24°C. Thermal zero shift is specified in terms of the maximum change of ZMO from its room temperature value, as a percent of full scale output. The operating and environmental temperature ranges for piezoresistive pressure transducers are specified on individual data sheets. The operating range indicates the limits within which the transducer will not be damaged. The compensated range indicates the limits within which the transducer will operate with predictable characteristics or for which the transducer has been compensated.


Figure 3-8: Schematic of 4 Gauge Bridge Showing Resistors for Span & Zero Thermal

Compensation

3.7.1. Thermal Sensitivity Shift Sensitivity Shift - The temperature compensation utilised for standard production units typically reduces the thermal sensitivity shift to a maximum of ± 1% of output per 55°C change in operating temperature. Tighter specifications can be met, if required, and the compensated temperature range can be reduced, expanded and moved up or down. Calibration data can also be supplied at any specified temperature within the environmental range, even beyond the compensated range.

Figure 3-6: Typical Thermal Sensitivity Shift 3.7.2. Thermal Zero Shift The changes in resistance of the various elements caused by temperature changes is rarely balanced. Therefore, as temperature changes, the bridge balance changes, resulting in a change in ZMO. The plot below shows an example of typical thermal zero shift. This is usually the largest component of thermal error, especially when measuring to only small fractions of full scale: Thermal ZMO shift is an absolute percentage of FSO, and is therefore, a more significant percentage of the measured value at fractions of full scale.

Figure 3-7: Typical Thermal Zero Shift


Because of variations in material properties, processes, and dimensions, the performance of a population of units of a given design will scatter about the nominal. To provide the lowest effect of temperature, Kulite measure the performance of each transducer during the manufacturing process, and resistance values are chosen to compensate for changes with temperature. The bridge circuit employed in these transducers is shown below. The resistance in series is used to reduce the sensitivity variation with temperature. Note that a resistor may be placed in both the positive and negative power supply lines; this is done to retain balance to aid in rejection of common mode noise. The resistances in series and parallel with one arm of the bridge correct for bridge unbalance and balance change (zero shift) with temperature. 3.7.3. Thermal Transient Response The compensated temperature range is the range in which the pressure transducer will meet the specifications for zero and sensitivity shift as given in the data sheets. Above and below this range, the transducer will continue to operate but the specification will gradually increase from the data sheet values. The transducer is compensated for equilibrium values of temperature and not for very fast temperature changes, pulses or excursions. If the pressure transducer is compensated from 0°C to 300°C and the actual 300°C differential occurs in a rapid excursion, the device must be allowed to come to an equilibrium temperature before it will meet the listed specifications. Compensation is only valid for equilibrium or slow changes in temperature, not for thermal shocks. Thermal transient response is the output of the transducer when subjected to a step-function temperature change from room temperature to the upper limit of the operation range. 3.8. Photo Flash Response Photo flash response is the output of a simple diffused silicon technology transducer when subjected to the flash from a photographic flash bulb approximately two feet in front of the transducer. Since Kulite developed the dielectrically isolated integrated sensor design, photo-flash response has been eliminated. The photosensitivity of silicon quite often will render diffused silicon diaphragm pressure transducers susceptible to radiation. The resulting transient output can be significant for applications where high intensity light can impinge on the diaphragm (such as in explosions or in engine combustion chambers). Several techniques have been devised to minimise this effect, such as covering the diaphragm with an opaque material. However because of the added mass, the diaphragm coating does adversely affects the acceleration sensitivity and frequency response of the transducer. 3.9. Transducer Resonant Frequency Resonant frequency is the frequency of pressure application at which the transducer responds with maximum output amplitude. The resonant frequency of a piezoresistive transducer is a function of its mechanical characteristics. Although it is actually a higher order system, a piezoresistive transducer can be represented as a second order single- degree-of-freedom spring-mass system. The response is shown below as a function of frequency. This curve represents the lowest (first) mode resonant response of the complex structure.


Figure 3-9: Frequency Response of a Pressure Transducer

This curve shows the variation in sensitivity of the transducer with frequency. The silicon diaphragm, because of its small mass and high stiffness, has an extremely high resonance frequency. Kulite characterises the diaphragm resonant frequencies for each design of silicon diaphragm. A rule of thumb for dynamic measurement is to select a device with a diaphragm resonant frequency at least five times the highest frequency to be measured. This is not usually a problem as for a typical Kulite miniature pressure transducer, the resonant frequency for a 5psi range unit is 160kHz rising to 1250kHz for a 2000psi unit. 3.10. Frequency Response Kulite’s piezoresistive pressure transducers are capable of response from steady state to frequencies into the ultrasonic range, and of response to fast rise time transient inputs. The sensing chip is generally mounted at the front of the transducer, making it equivalent to a flush-mounted diaphragm for most applications. The protective screen over the sensing surface has been designed to not degrade performance of the gauge; however, for some applications it may be removed. These applications usually involve high speed fluid flow, high frequency response, and a need to not disturb the surface of the body in the flow. Silicone rubber (RTV) may be added to the front to smooth out the surface. 3.10.1. Rise and Response Times Rise time is the time required for the output to rise from a small percentage to a large percentage of its final value, when excited by a step change in measurand. Unless otherwise specified the percentages are assumed to be 10% and 90%. Response time is the time required for the output to increase from zero to some specified percentage of its final value, when excited by a step change in measurand. The 63.2% (usually rounded to 63%) response time is the 'time constant", (t).

Figure 3-10: Response of First Order System to a Step Change

Technically, rise time and response time refer only to systems which are not underdamped. Piezoresistive pressure transducers are usually under damped, so using these terms is actually not


correct. The under damped system will respond very quickly, but will overshoot and oscillate for some time before reaching its final output value. The time constant is valid for any system. For under damped systems, we can calculate the period, (t), based on natural frequency, which will cause less than 5% distortion. A transient with shorter period than (t) will cause ringing (distortion) greater than 5%.

Figure 3-11: Transient Response of Second Order System with Different Damping

Piezoresistive pressure transducers provide the user with the capability of monitoring extremely rapid rise time pressure pulses. The rise time of the transducer is much faster than the period to which it will respond accurately. A rise time (t) to which the transducer will respond linearly to within +/-5%, can be expressed as a function of the period to which the transducer has a flat response (T).

T = 1 / 0.2 fn

Where fn = resonant frequency and T = t/4 = 1/ 0.8 fn 3.11. Acceleration Sensitivity Acceleration sensitivity is the maximum difference, at any pressure level within the specified range, between output readings taken with and without the application of specified acceleration along specified axes. Kulite’s pressure transducers, by design, are highly insensitive to acceleration inputs. Acceleration sensitivity is the sensitivity of the pressure transducer diaphragm to applied acceleration. The reaction of the device diaphragm to acceleration is a function of its stiffness, mass, thickness and diameter. Acceleration sensitivity is a function of both transducer overall diameter and pressure range. Sensitivities as low as 0.00003 % FS/g are available in the pressure sensitive directions. Therefore, for a 1000 g in-axis (worst case) acceleration input the total error might be as low as 0.03% Full Scale. Cross acceleration sensitivities are generally between a 1/5 and a 1/10 of that in the sensitive directions. Acceleration sensitivity for different pressure ranges are listed on each individual data sheet and are typical values from sample tests. For application in extremely high acceleration or vibration environments, Kulite has designed an acceleration compensated pressure sensor which has a negligible acceleration sensitivity even when subjected to accelerations in excess of 60,000g. The acceleration compensated pressure sensor is based upon the leadless technology and is capable of operation at temperatures in excess of 500°C.


3.12. Burst Pressure Burst pressure is the pressure which may be applied to the diaphragm, and the portion of the space subjected to the pressurised fluid medium, without rupture of the diaphragm. This is a static pressure rating; peak pressure greater than the specified range should not be applied at frequencies greater than 30% of the resonant frequency. The resultant mechanical amplification effect near the resonant frequency may cause erroneous data, or in extreme cases may burst the diaphragm. 3.13. Full Scale Output Full scale output is defined as transducer output from zero to + full scale (maximum range). Kulite pressure transducers have FSO of between 50mV and 200mV, depending on model with 100mV being the typical FSO. Net full scale output is the value of transducer output at the maximum rated load minus the output at the minimum rated load. 3.14. Supply Voltage or Excitation Supply voltage or excitation is the external voltage applied to the transducer for its operation within specified tolerances. The excitation across the piezoresistive elements causes a finite current to flow through each element. The I2R heating results in an increase in temperature of the gages slightly above ambient, which increases the resistance of the elements. Differentials in this effect may cause the output voltage to vary slightly with time, until the temperature is stabilised. With 10 Vdc excitation, stabilisation to within 1% usually occurs within one millisecond when tested at standard barometric conditions. Excitations other than 10 Vdc can be used with a maximum excitation of 15 Vdc usually being specified without damage. 3.15. Input/ Output Resistance Input and output resistance are measured between input leads and between output leads with an ohmmeter using 10 volts or less applied voltage. This measurement is extremely temperature sensitive and may vary significantly with small temperature variations, The main uses of these measurements are to calculate excitation current, and to be sure the bridge is not open or shorted.

3.16. Insulation Resistance Insulation resistance is the lowest value of resistance measured between all leads tied together and the shield, all leads tied together and the transducer case, or cable shield and transducer case. This measurement is made using a megohmmeter with 50 volts applied.


Section 4 – Environmental Limits 4.1. Diaphragm Loading It is important not to subject the silicon diaphragm to concentrated loading. It is designed for distributed pressure loading. Point or concentrated loading, such as pencil tips, tweezers, or other sharp objects pressed against the diaphragm will break the diaphragm! The thinned areas of the sculptured diaphragms vary from 0.2 to 10 thousandths of an inch thick. They are quite rugged when loaded with distributed pressure load, but will not tolerate concentrated point loading. 4.2. Temperature High Temperature Limit – Kulite specify a maximum operating temperature limit of 482°C for the XTEH-10L-190 and XTEH-10LAC-190 units. For applications where units are exposed to extreme temperatures for short time durations, the silicon diaphragm can be protected by ablative RTV coatings. Low Temperature Limit – Kulite specify a minimum operating temperature of -196°C for the CT-375 cryogenic integrated sensor pressure transducer which is normally compensated between the temperatures of -196°C and 38°C. 4.3. Acceleration, Shock and Vibration Environments of steady state acceleration, shock, and vibration may cause spurious output from a pressure transducer. However, Kulite transducers are very rugged and have been successfully used in very severe environments. Most Kulite pressure transducers are qualified for at least 10,000g steady acceleration and 100g peak vibration. However, if the shock or vibration environment contains significant energy at frequencies above 1/3 the resonance frequency of the transducers, it is possible to excite the resonance of the diaphragm. Then, unless the medium provides significant damping, the diaphragm may be broken. Always handle transducers carefully and gently, like delicate instruments. The miniature cables are especially susceptible to damage. 4.4. RF and Magnetic Fields Normally encountered magnetic and RF fields have negligible effect on the piezoresistive strain gage elements.. However, adequate isolation must be provided against ground loops and stray signal pickup. High intensity RF fields may require special shielding of the pressure transducer, cable, and amplifier. Kulite’s four conductor, shielded cables provide more than adequate shielding for most laboratory, field, and industrial environments. To prevent possible ground loops, the transducer case is normally insulated from .the cable shield. The shield should be grounded at the signal conditioner. The case will be grounded to its mounting structure, and to the medium being measured. In applications with a non-conductive mounting structure and low-conductivity media, special case grounding provisions may be required to provide maximum shielding.


Piezoresistive transducers are highly resistant to any effects of magnetic fields, even of very high intensity. They can, therefore, often be used in environments where other sensing mechanisms are not acceptable. 4.5. Sealing and Hermeticity Kulite ultra miniature pressure transducers use epoxy sealing (between sensing diaphragm or module and case) against leakage through the pressure measuring system. For M10 and larger units, true welded hermetic sealing is generally employed with glass to metal sealed assemblies for the isolated pressure capsules and cable/ connector outlets. 4.6. Media Compatibility Most specifications for pressure transducers refer to the "front" (pressure sensitive) end of the transducer. When discussing pressure media, pressure ratings, and temperature, the electrical lead end ("back") must be specified separately. 4.6.1. Pressure Sensitive End Typical silicon “open” diaphragm pressure transducers are compatible with clean dry gases, noncorrosive, medium pH liquids, and common oils. They are not compatible with corrosives, high or low pH liquids, solvents which might attack epoxy, or long exposure to water. Leadless technology silicon diaphragm pressure transducers can be used with any media which is compatible with stainless steel and silicon dioxide. 4.6.2. Electrical Lead End The reference tube or reference pressure side of standard Kulite ultra miniature differential and gauge pressure transducers connects directly through a hole in the pedestal of the pressure capsule. Due to the small diameter of the tubes and passages into the capsule, care must be taken to ensure debris which may be present in the pressure media does not block these small passages. With the leadless design of pressure transducers, only clean, non corrosive, non conductive liquids or gases may be used for the reference pressure media. Aerospace differential and gauge units frequently employ Kulite’s patented dual diaphragm technology which permits the use of pressure media at both the pressure and reference ports which is compatible with stainless steel. 4.6.2.1. Sealing Unless a transducer is designed with a hermetic connector outlet or uses a glass to metal seal header for the cable outlet, the back end of the unit must not be immersed in conductive or corrosive media. The standard Teflon insulated wire used for lead exit does not seal against water intrusion and the electrical end of standard transducers cannot be submerged in water without damage. 4.6.2.2. Reference Port The vent tube, on gauge units, must be kept open to local ambient pressure. In some applications it can be used as a reference port for differential pressure measurements. Maximum (burst) pressures for both pressure side and reference side are specified on individual data sheets. The reference side of gage transducers is epoxy-sealed.


4.6.2.3. Case Case material of most units is 17-4 PH or 300 Series stainless steel. Titanium is an option for some miniature units where weight or resistance to chemical attack is of specific importance. Compatibility should be checked for any medium or environment other than clean, dry, non-conductive gasses. 4.7. Nuclear Radiation Kulite has carried out testing with piezoresistive pressure transducers with some success. Please consult the factory for further details.


Section 5 - Application Information 5.1. Connection Diagrams Kulite piezoresistive pressure transducers incorporate integral compensation elements within the transducer case except for some of the very smallest units, in which case, the compensation resistors are contained within an in-line cable module. Where there is no external compensation module, the leads (cable) may be cut off as short as necessary. Electrical connections require only that correct polarity of excitation (input) and signal (output) be observed. 5.2. Mounting Techniques In order to achieve accurate measurements with any pressure transducers, it is important the pressure transducer diaphragm be allowed approximately 0.254mm (0.010 inches) radial and axial clearance within the pressure measurement volume. WARNING: Any mechanical loading other than distributed fluid pressure loads will cause stress concentrations in the diaphragm and will result in erroneous data and/or possible failure of the device. It is not permitted to press on an exposed silicon diaphragm at any time. Be sure that mounting of the device does not transfer stress to the diaphragm. The case and internal designs of Kulite pressure transducers provide the maximum possible mechanical isolation, considering size limitations. Care must still be exercised when mounting. Bending moments or mechanically stressing the device will usually result in asymmetric bending of the stress sensitive diaphragm and will evidence itself as a large zero shift or highly nonlinear outputs. When mounting the pressure transducer, it is advised that the output of the unit be continually monitored for shifts or deviations of any type. Recommended mounting torques are listed on the data sheets. Consult with Kulite for recommended mounting techniques. If using epoxy or other hard potting compounds for mounting of the transducer, keep the epoxy away from direct contact with the perimeter of the diaphragm. Any contact with the diaphragm may transmit stresses from the test object directly into the diaphragm causing zero shifts, erroneous data and thermal drifts. If it is necessary to fill a void near the diaphragm, use a soft potting material such as RTV silicone rubber which will not transmit shear stress. Flat units are particularly susceptible to bending stresses as a result of their low profile. Be sure to mount these on a flat surface without bending forces. 5.2.1. Strain Sensitivity 5.2.1.1. Threaded Mounting Configurations Detailed mounting dimensions and tolerances are provided on each data sheet. Mounting dimensions with tolerances outside these specifications may cause increased ZMO, increased thermal zero shift, or seal leakage. Avoid any stresses in front (toward the diaphragm) of the mounting threads. When threaded configuration transducers are mounted as specified, they are highly immune to the effects of strain in the mounting structure. 5.2.1.2. Cylindrical Configurations Because of their smaller size, unthreaded cylindrical configurations (Kulite XCQ, XCL etc.)are more sensitive to case strain than threaded designs. They should always be mounted with a relatively flexible adhesive, such as Dow Coming RTV 738. Mounting detail dimensions are provided on each data sheet. It is especially important to avoid application of adhesive near the front of the case. No stresses should be imposed on the front portion of the case.


When cylindrical configuration transducers are mounted as specified, they are highly immune to the effects of strain in the mounting structure. 5.2.1.3. Thin Line Transducers (flat pack) Thin lie transducers are designed to be mounted on the surfaces of airfoils, wings and other aerodynamic components. Transducers are typically mounted using epoxy or silicone rubber materials. If removal of the transducer without damage is desired, the selection of adhesive is very important. The housing of the transducer is very thin and can easily be damaged from bending and prying under the edge. Also solvents for the adhesive may damage the transducer interior if allowed into the pressure inlet area. Silicone RTV adhesives or wax can be used for mounting and can be cut away or removed with the application of heat to free the transducer without damage. Another factor associated with mounting materials is their effect on the transducer when installed on structures which are subject to bending. Structural surface strains which are transmitted to the base of the transducer result in an error signal output. Kulite’s design of flat pack transducers provides for base strain isolation within the assembly. However, its performance can be enhanced if additional strain isolation is provided by using soft mounting materials or by reducing the mounting area. In addition to the effects of mounting materials on base strain sensitivity, the thickness of the structure on which the transducer is mounted affects the strain output. Thin structures such as compressor or turbine blades or an airplane skin typically bend during normal operation. This results in increased error from the base strain sensitivity of the transducer. 5.2.2. Strain Measurement Kulite manufactures a range of semiconductor strain gauges which are available for either application by Kulite to a customer’s component or can be applied by the customer. Kulite provides a Strain Gauge Manual (reference KSGM-3) which is available on request. A more recent application report on the use of Kulite strain gauges is referenced in section 9.3.9. 5.3. Insulation The case of the transducer acts as a mechanical and electrical shield for the sensing elements. It is normally electrically insulated from the elements and is not connected to the shield of the cable. The case is assumed to be grounded to the structure in which it is mounted. Insulation resistance between all leads connected together and the transducer case or the shield is 100 megohms (minimum) at 50 volts dc. In Kulite’s original Integrated Sensor design of pressure transducers, the pressure media is in contact with the piezoresistive strain gauge elements, the metallised connections on the surface of the silicon diaphragm and the 4 or 5 gold wires connecting the diaphragm to the header. This design of transducer is commonly referred to as an “open” diaphragm design and has been used predominantly for ultra miniature units. With all “leadless” design transducers, the gauge elements are insulated from the pressure media by the silicon diaphragm. For all oil filled isolated designs, the isolation diaphragm provides a robust barrier between the pressure media and the silicon diaphragm. 5.4. Cabling The cable which connects a transducer to its matching electronics is an important part of the overall measurement system. It must transmit the transducer signal to the associated signal conditioning equipment without distortion or introduction of noise. Cables must also not affect transducer or test specimen characteristics. Good transducer cables are as small, light and flexible as possible, considering their specific intended application.


5.4.1. Standard Cables Each Kulite transducer is equipped with an integral shielded multi-conductor cable or individual lead wires, typically 30 inches long. The lead wires are colour coded per ISA standards. Individual lead wires and the outer cable jacket are typically Teflon-insulated. Because they are designed for maximum flexibility and micro-miniature size, these cables should be handled with care; they can be damaged if misused. They should not be stepped on, kinked, knotted, etc. When possible, the cable should be tied down within two to three inches of the transducer. Long, unsupported lengths of cable may load the test specimen and lead to cable damage. Good housekeeping should be observed; excess cable should be neatly coiled and tied down. In humid applications, it is good practice to provide a drip loop at the transducer. It may also be advisable to seal the cable to prevent moisture from entering the cable assembly. 5.4.2. Splicing and Extension Cables Leads may be spliced using good instrumentation practice. Care must be taken to minimise the resistance of the splice. The effects of cable resistance on sensitivity and the effects of RC filtering in the shielded cable must be accounted for when accurate effective sensitivity is needed. Soldered or crimped splice and copper extension wire are preferable, to reduce the likelihood of thermoelectric generation of error voltages. For best protection from EMI/RFI induced noise, any extension cable should be shielded. The transducer cable shield should be connected to the extension cable shield, which can then be grounded at the signal conditioner. 5.4.3. Loading Effects An equivalent circuit of a piezoresistive transducer for use when considering loading effects is shown below:-

Figure 5-1: Schematic Diagram of Loading Effects

Where :-

R0 = output resistance of the bridge including cable resistance E0 = sensitivity into an infinite load E0L = loaded output sensitivity RL = load resistance

Using the equivalent circuit above, and the output resistance supplied on the calibration document, the effect of the loading may be directly calculated:-

E0L = E0 ( R1 / (R1 + R0) ) (5.1)

Because the resistance of the strain gauge elements vary with temperature, output resistance must be measured at the operating temperature.


5.4.4. Effects of Cable on Transducer Sensitivity Each Kulite transducer is calibrated and supplied with a specified length of cable. When utilising very long cables in a particular application, three effects must be noted: Excitation voltage drop, signal attenuation, and RC filtering effects. 5.4.4.1. Excitation Voltage Drop Resistance in the input (excitation) wires may significantly reduce the excitation voltage at the transducer, resulting in a loss of sensitivity. The new sensitivity (EiL) is equal to:-

EiL = E0 ( Ri / (Ri + 2Rci) ) (5.2)

where Ri is the input resistance of the transducer and Rci is the resistance of one excitation wire. This effect may be overcome by using remote sensing leads. 5.4.4.2. Signal Attenuation Signal attenuation also results from resistance in the output wires. This attenuation may readily be calculated from the relation:-

E0L = E0 ( RL / (R0 +RL + 2Rco) ) (5.3) where the terms are as defined in Section 5.4.3, and Rco is the resistance of one output wire between transducer and load. 5.4.4.3. RC Filtering RC filtering in the shielded instrument leads may attenuate the high frequency components in the data signal. The stray and distributed capacitance present in the transducer and a short cable are such that any filtering effect is negligible. However, when long leads are connected between transducer and readout equipment, the response at higher frequencies may be significantly affected. Typical instrumentation cable will have capacitance of approximately 30 pF/ft.

Figure 5-2: Schematic Diagram of Simplified Transducer Circuit with Long Cable

The -3 dB cutoff frequency for this system is: fc = 1 / 2π (R0t + 2Rc) Cc (5.4) where:-

Rc = Resistance of cable Cc = Capacitance of cable

Because the resistance and capacitance is actually distributed along the cable, the above circuit only approximates the effect of long wires. It is suggested that each 1000 feet of cable be considered as a


separate RC network. For precise measurements, line filtering action must be determined experimentally as part of the system calibration. 5.5 Measurement of Dynamic Pressures In any dynamic measurement, the frequency response of the transducer and the electronics must be considered. The requirement peculiar to pressure measurement is that one must consider the fluid coupling to the transducer from the measurement point. For example, when a transducer must be placed remote from a measurement point, the response of a pressure line can severely limit the response of the measurement system. When measuring pressure oscillations in the audio frequency domain, the selection and placement of a transducer at the measurement point can be critical. In addition to the dynamic characteristics of a transducer and its placement, the results are a function of certain qualities of the fluid. These are significantly different for gases and liquids. The summaries below review some of the fundamental considerations from the standpoint of a pressure measurement. 5.5.1. Acoustic and Fluid Flow Effects 5.5.1.1. Acoustic Fundamentals Sound Speed in Liquid - The speed of a compressional pulse in any homogeneous isotropic medium is:- (5.5) The approximate speed or sound in two commonly used liquids:

Water = 1440 meter/sec (4724 ft/sec) Alcohol = 1240 meter/sec (4068 ft/sec)

For comparative purposes the speed of sound in steel is about 5500 meter/sec (18040 ft/ sec) 5.5.1.2. Sound Speed in Gas The speed of sound in a gas is:-

(5.6) M

RTγc = where c = speed of sound in a gas

γ = ratio of the two principal specific heats of the gas R = gas constant per mole T = absolute temperature M = molecular weight of the gas

From this equation we may conclude that the speed of sound in ideal gases depends only on the kind of gas and the temperature and is wholly independent of changes in pressure. If we denote Ct as the speed of sound in a given gas at temperature T and by Co the speed in the same gas at temperature T and apply the above equation, we have :-


Ct = Co T/ To (5.7)

At 0°C, the speed of sound in dry air is 331.45 m/s (1088 ft/sec) and the speed increases about 0.6 m/s (2 ft/sec) for each degree centigrade of rise in temperature. Sound velocity increases slightly with increasing humidity. The speed of sound in several commonly-used gases at 15°C is: Air = 341 meter/sec (1119 ft/sec)

Hydrogen = 1270 meter/sec (4167 ft/sec) Carbon Dioxide = 258 meter/sec (846 ft/sec)

5.5.1.3. Organ Pipe Resonance The wavelength of the fundamental wave is equal to four times the length of the pipe for a pipe which is open at one end and closed at the other end. Resonant excitation can be produced by the fundamental frequency, fn, and all the odd harmonics.

Figure 5-3: Organ Pipe Resonance

(5.8) 5.5.1.4. Cavity Resonances (Helmholtz)

Figure 5-4: Cavity (Helmholtz) Resonances


(5.9)

(5.10) 5.5.1.5. Transmitting Tube Connected to Cavity

Figure 5-5: Tube and Cavity


The natural frequency of the tube and cavity system is:- (5.11) 5.5.1.6. Pressures in a Flowing Fluid The inertia of a flowing fluid causes an impact or dynamic pressure to be generated on surfaces perpendicular to the direction of flow. This phenomenon is used in Pitot (or Pitot-static) tubes and stagnation probes to measure flow velocity.

Figure 5-6: Pitot-Static Probe

Stagnation (also "total") pressure (Ps) is measured inside the open end of the tube where the gas stream has decelerated until its velocity is zero. Static pressure (Po) is measured in the; static tube, which has holes perpendicular to the flow direction about 10 tube diameters back from the end of the tube. This is the pressure not caused by flow velocity. Impact ("velocity") pressure is the stagnation pressure minus the static pressure (Ps- Po). It is the pressure caused by the inertial effects of the flowing fluid. For supersonic flow, the configuration is much the same, but the shock wave around the tube does not permit direct measurement of Ps. Instead, the total pressure behind the shock wave is measured and used in a more complicated relationship to calculate the velocity. In addition to Pitot/static tubes, many flow measurement schemes use differential pressure measurements across orifice plates, venturis, or other flow restrictions such as oil, fuel or air filters. In some applications, the normal pressure drop between two points in a system of pipes can be used as an indication of flow. Kulite have developed a Pitot/ Static probe which is commercially available and is fully reported in reference 5.3.10. 5.5.1.7 Pressure Shock Wave Effects Pitot or stagnation probes are often used to measure pressure shock waves generated by explosions, sonic booms, or lightning strokes. These shock waves are large amplitude mechanical (compression-rarefaction) waves travelling at supersonic velocities. When energy is suddenly released into a fluid in a concentrated form, such as by a chemical or nuclear explosion, the local temperature and pressure may rise instantaneously to such high values that the fluid tends to expand at supersonic speed. When this occurs, a blast wave forms, and propagates the excess energy from the point of explosion. If the point of explosion is far from any fluid boundary, the blast wave assumes the form of an expanding spherical shock wave followed by a radially expanding fluid originating from the point of detonation. At measurement locations near the point of explosion, the pressure wave front has an extremely fast rise time, near instantaneous. The amplitude of the wave front and subsequent reflection waves may be


extremely high. This is followed by a long period of decay, a rarefaction or expansion wave and a transition into acoustic waves. Significant pressure variations may continue for a relatively long time. Although the amplitudes are lower, sonic booms and thunderclaps generate similar short rise time, long duration pressure disturbances. In many applications, such as explosive and shock tube tests, the expanding gases carry debris produced by the explosive, or picked up in transit. 5.5.2. Acoustic Limitation of a Pressure Probe Frequency response requirements are often greater than 500 Hz for the measurement of transient total pressure in gas paths of gas turbine engines. To measure this, small pressure transducers such as the Kulite XCQ series of cylindrical ultra miniature units are placed in probes. To protect the transducer from particulate damage and to provide a more thermally benign environment it is sometimes desirable to place the transducer back from the front of the probe. (An extreme example is to measure at the end of a capillary tube.) An example of such device is shown by Figure 1.16.

Figure 5-7: Dynamic Pressure Probe

Simply estimating the lowest resonance of this system by using the equations shown earlier for organ pipes or for tubing/cavity combinations results in an answer of about 3000 Hz. Using a high frequency pressure transducer and the dimensions shown in Figure 1.16, the first resonance is close to 3000 Hz and has a high amplification factor typical of a system with low damping. This is far below the resonance frequency for the pressure transducer diaphragm, which for the Kulite XCQ-072 series of pressure transducers is typically 150 kHz for a 5 psi unit, rising to . 5.5.3. Dynamic Response of Transducer in Liquid System In addition to the limitations from the acoustic characteristics of a liquid system, the mechanical characteristics of a transducer must be considered. Normally one thinks of the dynamic behaviour of a pressure transducer as being a function of its resonance frequency and damping ratio. When making measurements in liquids this oversimplifies the situation. The transducer force summing device may be considered as a spring, mass, and a damper. When this is attached to a liquid system one must effectively add liquid mass and damping to this mechanical system. When doing so the resonance frequency of the measurement system is significantly lowered. This effect can be important to consider when making dynamic pressure measurements in liquid systems and when testing static pressures in a system using liquid filled lines for connection to a remote transducer. If vibration is present on liquid filled lines and transducers, unwanted oscillations can be added to the measurement. Sometimes small diameter orifices (restrictors) can be added to damp out these oscillations. One method to predict performance of a transducer in these types of applications is to express the resonance frequency of the transducer/ liquid system in terms of the volumetric compliance of the transducer's force summing network. Kulite oil filled pressure transducers are normally extremely reliable products. The metal isolation diaphragm along with the rugged silicon-sensing diaphragm together make for a very robust and durable product. However Kulite has come across situations where a seemingly benign application can


lead to unexpected failures. Kulite transducers are often used for measuring fuel pressures after the fuel boost pump in gas turbine engines. In some of these applications Kulite has received units that fail due to gold wire lead breaking and deformation/ cracking of the isolation diaphragm. Gold wire breakage is a very unusual condition because the gold wires have a very high natural frequency, which is well outside the engine vibration spectrum. Under static pressure conditions the isolation diaphragm does not deflect as pressure is applied, and the silicone oil in the pressure capsule transmits the pressure with no relative movement. However a research program within Kulite has demonstrated that very fast, large amplitude dynamic pressure signals can set up uneven pressure waves across the face of the diaphragm due to cavity resonances within the transducer pressure port and pipework. This may cause the isolation diaphragm to be subjected to alternating compressive and tensile stresses, due to transient surface tension effects, which can eventually lead to a fatigue failure of the isolation diaphragm. Additionally, the motion of the isolation diaphragm can cause the silicone oil to flow rapidly back and forth and stress the gold wires, eventually causing fatigue fracture of the wires. Finally, there is a phenomenon which occurs in a liquid when a rapidly moving pressure wave is stopped and generates a pressure spike which can be well in excess of the steady state pressure. This phenomenon is commonly referred to as “water hammer”. The pressure wave can be generated by any rapid change in a liquid filled system, such as valves closing. The pressure spikes which occur can have a large enough magnitude to damage the pressure transducer, associated components and pipework. Laboratory testing at Kulite have demonstrated pressure spikes in excess of 2000psi in a system with a static pressure of only 275psi. In order to protect a pressure transducer from water hammer, a filter can be installed on the front of the pressure fitting. Several tests have been performed with both isolated design and leadless sensors and a variety of filter sizes. The addition of a filter decreases the response time of the transducer to a pressure wave and all of the tested filters completely blocked the water hammer effect. The filter is selected to attenuate the pressure spike to a safe level whilst retaining sufficient response time for the measurement or control function. The table below gives some examples of filter pore sizes and response times

FILTER SIZE THICKNESS RESPONSE TIME 10 Micron .031” 23.95 milliseconds 20 Micron .031” 10.15 milliseconds 40 Micron .039” 3.36 milliseconds

100 Microns .062” 2.08 milliseconds References to technical papers on this subject are given in section 9.3.4 and describe in more detail techniques which can be used to reduce the frequency of the cavity/ pipe resonant systems and attenuate pressure spikes. 5.5.4. Dynamic Pressure Measurements at High Temperatures Kulite pressure transducers are being used to measure dynamic pressures at high temperatures in many areas of research and development, particularly within the gas turbine (aero and industrial), aerospace and automobile industries. Kulite’s SOI and leadless technologies has enabled silicon based pressure transducers to operate at temperature in excess of 540°C (1000°F). Technical papers which have been published in this area are referred to in sections 9.3.7 & 9.3.8. Despite the capability of the latest Kulite piezoresistive pressure transducers which are mounted on the casings of gas turbines to withstand temperatures in excess of 540°C (1000°F), there are regions of a gas turbine where dynamic gas path pressures are required to be measured which are hotter i.e. a


modern high pressure compressor outlet temperature typically exceeds 650°C (1200°F) and the gas temperature in the high pressure turbine can exceed 1400°C (2550°F). In these ultra high temperature environments the preferred way of measuring these small dynamic pressures reliably is to use either a flush-diaphragm transducer mounted in a water or air cooled jacket, which requires the supply of cool water or air and, although very effective, may be impractical, or the use of a non-resonant semi-infinite tube (SIT) system which removes the pressure transducer from the very hot environment by a distance of up to 1 metre. The diagram below is a schematic representation of an SIT system.

Figure 5-8: Diagram of a SIT Dynamic Pressure Measuring System

LM is typically 3 feet maximum and LL is 18 feet minimum. The theory behind the SIT design is that for a sufficiently long tube (semi-infinite), pressure fluctuations at the measuring station will have attenuated to small enough values at the far end that their reflection back to the measuring station will be very small, giving negligible measurement errors. The system does not possess a measurable resonant frequency. The far end termination is usually a closed end to the pipe. Alternatively a low response pressure transducer can fitted to measure the static pressure value. Such SIT systems are capable of measurement bandwidths of many kHz but require precision manufacture, minimum discontinuity at the transducer matching position, no steps or discontinuities in the bore of the pipe and sufficient length of backing tube for the test conditions. Olsen [“Acoustic Engineering”, 1957] describes the use of non-resonant dynamic pressure measurement systems and gives a simple equation to calculate the attenuation against frequency characteristics of the tube to the transducer. Since the system is particularly prone to resonance problems caused either by manufacturing defects or lack of care when handling/installing the system, it is usual to calibrate the system and not rely upon the theoretical predictions used for design.


Section 6 - Electronics 6.1. Power for Excitation Piezoresistive transducers are passive devices and require an external power supply to provide the necessary current (Ix) or voltage excitation (Ex) to operate the transducer. These energy sources must be well-regulated and stable, since they may introduce sensitivity errors and secondary effects at the transducer which will result in error signals at the output. The excitation across the piezoresistive elements causes a finite current to flow through each element. The I2R heating results in an increase in temperature of the elements slightly above ambient which increases the resistance of the elements. The power supply compliance voltage and regulation must be able to maintain constant voltage excitation on this varying resistance. Most Kulite transducers require 10.00 Vdc excitation, but can be operated at higher or lower voltages. When exciting an unamplified pressure transducer, you may choose to ground one side of the excitation source, but do not ground either of the output leads. DO NOT GROUND BOTH INPUT AND OUTPUT LEADS. GROUNDING BOTH SIDES WILL SHORT CIRCUIT ONE STRAIN GAGE, PRODUCING ERRONEOUS OUTPUT SIGNALS. If floating both input and output of the pressure transducer ensure that common mode voltage of the power supply does not exceed 50V. Accidental short term application of excitation voltage to the output leads will not damage the transducer, but it should not be operated while connected backwards. 6.1.1. DC Power Supplies Most Kulite piezoresistive pressure transducers require a constant-voltage supply for excitation. A constant-current supply should not be used unless the transducer is specifically designed or compensated for operation in this mode. Because the typical four-element transducer may not be perfectly balanced or matched, variations in excitation voltage or current, including ripple, will result in an error output signal. It is necessary, therefore, that a stable and well-regulated power supply be employed. A number of important characteristics must be considered in the selection of a suitable power supply. Among these are: • Line Regulation • Load Regulation • Ripple and Noise • Temperature Stability • Time Stability • dc Isolation The output of the transducer is differential, so the signal conditioner input should not be grounded. This requires that the power supply be well insulated from ground. Not only must the power supply be well insulated to prevent dc leakage currents flowing through the transducer, but in addition ac coupling to ground and power line must be minimised to prevent line transients and dynamic around loops from generating error signals. Recommended grounding point is at the signal conditioner output. To calculate power supply requirements, the required current is calculated from:

Ii = Vi / Ri (6.1) Where:


Ii = Input current Vi = Excitation voltage Ri = Input Resistance

When powering more than one unit with a single power source, use the parallel combination of input resistance for all units used.

Ii = Vi / Rc (6.2) Rc = Parallel combination of Input Resistance Typical current requirement is 8 mA per transducer. 6.1.1.1. Constant-Current Power Sources In many applications, the effects of long-line resistance and/or extraneous inputs are not negligible. The resistance of a long line will change with temperature, and the voltage drop along the line will vary as the transducer resistance or load changes. For these applications, constant-current excitation provides an output that is less dependent on these effects than is voltage excitation. In addition, current excited bridges are more linear than voltage-excited bridges when the percent variation of bridge resistance is relatively large. The bridge output tends to be proportional to absolute resistance variations when the excitation source is current; and proportional to a unit resistance variation when the excitation is voltage. Thus, resistance gages or transducers which are to be used in a constant-current system must be compensated and calibrated with constant-current excitation over their full range of operation. Piezoresistive pressure transducers, specifically designed for operation with constant-current systems, are only available as specials from Kulite. 6.1.1.2. External Sensing The voltage drop along long lines between a constant-voltage supply and transducer results in a reduced and sometimes unpredictable voltage at the transducer. Errors and spurious signals may appear 'at the transducer output due to variations in the resistance of these lines caused by temperature changes.

Figure 6-1: 6-Wire Connection to Wheatstone Bridge

Many constant-voltage supplies provide for external voltage sensing leads which connect directly to the transducer, independent of the power or excitation leads. Low current in the sensing leads reduces


the voltage drop along these lines and the effects of changes in resistance. Thus, the voltage across the transducer is maintained constant and independent of resistance and current variations on the power leads. Input resistance of a pressure transducer may vary significantly over its operating temperature range. This change results in a relatively large change in input current, and proportional change in power-line voltage drop. With external sensing wires, the power supply controls and maintains the voltage at the transducer at a constant level. 6.1.2. AC Excitation Kulite piezoresistive transducers may be excited with an ac carrier signal. The amplitude of the signal must be stable and the frequency should be five to ten times the maximum frequency of interest. Kulite piezoresistive transducers may be operated with up to 150% rated excitation voltage. With sinusoidal excitation voltages, the peak carrier signal will almost reach this limit. Therefore, it is recommended that the rms value of the carrier voltage be limited to the dc rated excitation voltage or less. 6.2 Signal Conditioning The millivolt output pressure signal of the typical pressure sensor can be conditioned to interface with virtually any data acquisition system ECU, FADEC, EICAS or any other control or monitoring device. A voltage, current, frequency or digital output can be provided as an option. Kulite has developed an in-house capability to design and produce microcircuitry, ASIC based designs and hybrid electronic modules as well as discrete component designs, including surface mount devices, to temperatures in excess of 365°F (185°C). Kulite also has the capability design mechanical packages that place the solid state electronics in a stress free environment, thus allowing customers to use transducers in severe temperature and vibration environments. 6.2.1 Analogue Amplifiers Many pressure transducers are available with integral electronics modules which amplify the millivolt output from the sensing Wheatstone bridge to a higher voltage level. The electronics can be supplied to operate from either a regulated or an unregulated supply. A typical regulated input voltage is 10 volts dc and frequently specified unregulated voltages are 12 volts dc ± 4 volts, 28 volts dc ± 4 volts and 10 to 40 volts dc. The output voltage range is usually 0 to 5 volts but options include output ranges of 0.5 to 4.5 volts, 0 to 10 volts, etc. 6.2.2 Digital Corrected Analogue Output For applications which require the highest accuracy, Kulite have developed a range of pressure transducers which incorporate a microprocessor which provides digital compensation for the effects of temperature, as opposed to the passive compensation techniques described earlier. The total error band for a digitally corrected transducer is typically better than 0.15% full scale, which represents an improvement over a passive compensated pressure transducer of at least 5 times. 6.2.3 Digital Output Kulite provides microprocessor based electronic packages which amplify the millivolt sensor output, digitise the amplified voltage and output the digital data stream in one of several industry standard formats, such as RS485, CanBus, etc. 6.2.4 Pressure Switch Output Pressure switches usually employ electromechanical technology of bellows or bourdon tubes which are connected via a mechanical linkage to a microswitch. Kulite have designed a range of solid state pressure switches which are based on the integrated silicon pressure sensor design connected to an electronic switching module. The reliability of the solid state switch is often an order of magnitude


greater than the electromechanical equivalent. For aeroengine applications which frequently require pressure switches to operate at temperatures in excess of 200°C, unamplified pressure transducers can be used which input to the engine electronic control system (EEC). The software within the EEC can be designed to set the switch point and the required level of hysteresis. Thus one pressure transducer can be used for many switch applications with different switching characteristics which are programmed in software. 6.2.5 Solid State Replacements for Electro-Mechanical Pressure Transducers In the past such practical implementations of pressure transducers used potentiometers, LVDTs (Linear Voltage Differential Transformer), synchros, variable reluctance systems, etc. All these pressure transducers used a Bourdon tube moving a mechanical part of a system, which resulted in a change of the electrical output. Some of these early transducers employed mechanical designs comparable in complexity and ingenuity with the most expensive Swiss watches. Unfortunately, none of these types of transducers escaped the inherent disadvantages of mechanical systems with moving parts. Kulite transducers employ a different approach, as explained in earlier sections of this handbook. The pressure-sensing element is a solid-state component, to which an electronic circuit is added which produces a normalized, compensated output. The piezoresistive bridge is arguably the most widely used, reliable and versatile sensing element available. Kulite has developed a range of electronic interfaces which operate with the Kulite piezoresistive silicon pressure sensing bridge to replace all the various obsolescent technology, electro-mechanical pressure transducers. These solid state replacement pressure transducers can be designed by Kulite to be form, fit and functionally identical to the old pressure transducers but have the reliability, performance and cost advantage of Kulite’s new generation of transducers. References to a paper which gives more details about Kulite’s developments in this area are given in section 9.3.6 6.2.6 Wireless Transmission In applications where a cable connection to a pressure transducer is either undesirable or impractical, Kulite have developed a range of pressure transducers which transmit the pressure data via an rf link to a ground station. The communications standards which can be employed include IEEE 802.11b (WiFi), IEEE 802.15 (ZigBee), Bluetooth, ISM frequencies 868/ 915 MHz, 2.4 GHz. Power for the pressure transducer and the processing electronics and transmitter can be provided by replaceable or rechargeable batteries, inductive coupling or optical power transmission. 6.3. Readout and Recording Devices A detailed discussion of readout and recording devices is beyond the scope of this handbook. However, some characteristics of these devices are important to overall system performance and data quality. The category of readout and recording devices encompasses all types of meters, oscilloscopes, analysers, recorders, voltage controlled oscillators, and shaker control systems which receive their inputs from the transducer/amplifier system. Meters may have analogue or digital displays, and recording devices may use direct, AM, FM, or digital recording. Some analysers and control systems digitise and process data in "real time" while some process delayed, sampled, and/or hatched data. However, regardless of the readout, recording or analysis techniques used, all of these "downstream instruments" have input impedance, frequency response, dynamic range, noise, and overload characteristics which may significantly alter the data. Instruments which digitise the analogue data and then process it digitally provide additional opportunities for Murphy's law to come into play. Some of these characteristics will be discussed in this section. 6.3.1. Input Characteristics Input impedance, frequency response, dynamic range, noise, and overload response characteristics of downstream instrumentation are sometimes overlooked when putting together the total measurement, recording and analysing system.


The input impedance must be high enough to prevent overloading, slew rate limiting, and distortion. A rule of thumb is that input impedance of any instrument should be at least 100 times the output impedance of the preceding device. Most instruments in use today have input impedances of a megohm or more. Preceding instruments have output impedances of 1 k ohm or less, so this is seldom a problem. Frequency response of the meter, analyser or recorder is often different from that of the transducer/amplifier system. If it is wider, then all frequencies will be processed undistorted. However if it is narrower, or if it distorts at some frequencies, its frequency response over the frequency range of interest must be known. Also, the user must know how the instrument treats frequencies outside its flat frequency band. Does it roll them off? How steeply? Or, does it amplify or distort them? Or, does it fold them back into the pass band and create new frequencies ("aliasing")? Dynamic range, noise, and overload response characteristics are all important to the amplitude accuracy of the data. The noise level should be less than half of the lowest expected signal level, and preferably even less than that. Overload "headroom" should be sufficient to accurately process any possible overrange signal in the frequency band of interest. Finally, overload response must permit any distortion (such as clipping) of the overrange signal to disappear as soon as the signal is again within the dynamic range of the instrument. 6.3.2. Meter Characteristics Most meters (regardless of their scaling) sense and respond to either the average or the rms value of the input. They therefore have some time constant or averaging time associated with the reading. Even direct reading galvanometer-type meters cannot respond instantaneously; they have a response time. This averaging time or response time gives the effect of a low pass filter by responding less to higher frequency inputs. Some meters incorporate peak detecting circuitry; they provide an output proportional to the peak signal detected during some time interval. Other meters provide a sample-and-hold feature which allows manual or automatic time period sampling and readout of the average or peak detected during the sampling period. The two greatest error sources when using a meter are (a) using it for readings in the lower (least accurate) part of its scale, and (b) not using a true rms meter for reading the rms value of a non-sinusoidal signal. 6.3.3. Errors in Digitising Whenever an analogue signal is digitised, the possible digitising errors are added to all of the other potential errors in the system. The two most common digitising errors are aliasing and accuracy errors. Aliasing refers to the phenomenon of creating new frequencies during the digitising process. This happens when the sampling (digitising) frequency is not high enough relative to the highest frequency present in the analogue signal. The process of sampling generates heterodyne frequencies equal to the sum and the difference of the data frequency and the sampling frequency. If the difference frequency, fs – fd falls in the frequency range of interest, it is called an alias frequency. In order to avoid alias frequencies, fs must be at least 2 x fd where fd is the highest frequency present in the input.


Section 7 - Measurement of Transient Pressure Pulses Special problems are encountered in transient pressure pulse measurement, which place stringent requirements on the measuring system. Some of these problems are:

High levels. Wide frequency content of pulses. Transient characteristics of instrumentation.

As a result, each part of the instrumentation system should be evaluated and selected for:

Adequate linear dynamic range (including safety factor). Adequate linear frequency response over a wide range. Ability to respond to transient inputs. Negligible phase-shift errors over the frequency range of interest.

Kulite pressure transducers provide a unique combination of high performance in all of these critical characteristics. 7.1. Dynamic Range The transducer should be selected for its ability to meet the linear dynamic range required. All Kulite transducers are rated for both linear dynamic range and for burst pressure (maximum static input without damage). Care should be taken that the signal output does not overload the associated electronics. For a transducer of known sensitivity and a given pressure input, the signal which the amplifier must handle can be computed. The gain of the related amplifier must be constant over the entire dynamic range of the input pressure and be adequate to provide full-scale output for the expected input. 7.2. Low Frequency Response Inadequate low frequency response in the measurement system will result in failure to accurately reproduce the transient pulse. Piezoresistive transducers respond to steady-state or zero static pressure. When they are connected to dc amplifiers or dc readout instruments, there is no limit to the duration of a pulse they can measure. In some applications, where true steady-state measurements are not required, and where low frequency drift may be a problem, ac coupling can be used. When ac amplifiers or other equipment with limited low frequency response are connected in a system, the pulse wave shape will not be maintained. The nature of this inaccuracy can be seen by examining the effect on a rectangular pulse of duration T and amplitude A applied to the input of a signal conditioner which does not respond to dc (steady state) signals. If this transient is passed through an ac system with first-order low frequency response, the resultant output will be as shown in Figure 7-1. A first-order system has the same low frequency response as a single resistor-capacitor high-pass filter whose time constant in seconds is equal to RC (ohms x farads). Such a system exhibits a low frequency cutoff equal to 1/ 2π RC and is 5% down at a frequency of 3/ 2π RC. A critically damped mechanical system has a first-order response.

(7.1)


The output does not remain at the peak value for the full pulse duration, but decays exponentially, "droop." The output amplitude at any time, t, (during the pulse) can be expressed as: where RC is the system time constant. At the termination of the pulse, the output does not return to zero, but overshoots in a negative direction. Recovery from this "undershoot" occurs at the same exponential rate as droop.

Figure 7-1: Response of System With First Order LF Response to a Rectangular Pulse

The ratio of total pulse height to droop is a function of the ratio RC/T. The larger this ratio, the less error (and the less undershoot). For example, if this ratio is 20, there will be approximately 5% error in the rectangular pulse amplitude; if the ratio is 50, there will be only a 2% error. Although slightly more complex to analyse, it can be shown that similar low frequency effects occur for other pulse shapes. If the requirement for adequate RC/T is not satisfied, it is possible to predict the degree of error for these pulses and apply appropriate correction factors to the data obtained. 7.3. High Frequency Response Consider again a rectangular pulse of duration T and amplitude A. If this transient is passed through a system with first-order high frequency response (corresponding to a single RC low-pass filter combination), the resultant output will be as shown as shown in Figure 7-2. The effect of the high- frequency rolloff is to slow the rise and fall time of the pulse, thus rounding both the leading and trailing edges.

Figure 7-2: Response of System with First Order HF Response to a Rectangular Pulse

It is also of interest to note the effect of passing the rectangular pulse through a system possessing second-order high frequency response (Such a system corresponds to the electrical frequency response of a single LC low-pass filter combination, or a damped mechanical system). Figure 7-3 shows the resulting output, A high-frequency ringing at approximately the resonance frequency is superimposed on the pulse. The amplitude and duration of the ringing depends on the damping factor.


Figure 7-3: Response of System with Second Order HF Response to a Rectangular Pulse

Fourier analysis show that short transients ,contain significant high-frequency components. Faster rise-time transients contain higher frequency components. Both the transducer and associated systems must have adequate high frequency response to avoid undesirable measurement distortion. Although the diaphragm is actually a higher-order structure with multiple high-frequency modes, the high frequency response of a piezoresistive transducer with very little damping is approximately a second-order function and is determined by the transducer fundamental (first) modal resonance frequency. The use of such devices provides desirable high frequency response along with minimum phase shift in the frequency range of interest. Transients, however, may excite such a transducer's first mode to resonance; natural frequency "ringing" will then be superimposed on the basic input transient. In the case of short, rectangular or other transients with essentially zero rise time (very short rise time in proportion to the natural period of the transducer), almost 100% overshoot on the transient may occur along with subsequent excitation of transducer natural frequency. To minimise or prevent these distortions, the transducer should have a natural period (the reciprocal of the natural frequency) one-third the expected rise time or less. Resonance frequencies should be as shown in the following table in order that the natural period be one-fifth the pulse duration for undamped transducers measuring half-sine or sawtooth transients:

Required Pulse Width (microseconds)

Required Natural Period (microseconds)

Resonant Frequency (Hertz)

500 100 10,000 200 40 25,000 150 30 33,000 100 20 50,000 75 15 67,000 50 10 100,000

If the matching amplifier possesses high frequency response flat. to at least one-half the pressure transducer resonance frequency, no appreciable error will be introduced by amplifier rolloff characteristics. The most common types of readout devices are: (1) Oscilloscope, (2) Computer based dynamic data capture devices. The storage oscilloscope is probably the most versatile and easy to use. A good quality scope will have a response from dc to above a gigahertz so that it will not introduce significant errors.


7.4. Phase Shift Faithful reproduction of transients requires that the measurement system be free of phase distortion. As we have seen, undamped Kulite pressure transducers exhibit 0° phase shift over their useful frequency range, and thus are not a source of this type of error. Matching amplifiers should be chosen for acceptable phase characteristics. Recording galvanometers, when properly damped, exhibit linear phase shift over their usable frequency range. Other readout devices do not, in general, introduce phase distortion. If a virtually undamped transducer is subjected to a transient which excites its resonance frequency, the magnitude of the output does bear a fixed relationship to the mechanical input. This relationship, however, is so sharply dependent upon the input pulse duration and rise time that the output signal is rarely of practical value. It is usually possible to select a transducer with a high enough natural frequency that it will not resonate for a given transient input. In some cases, when rise times are extremely short, it may become necessary to resort to electrical filtering. The data in the pass band of a low-pass filter will have quantitative value, even if the transducer is resonating, as long as the filter has a linear phase-shift characteristic (constant delay). When filtering is used in the measuring system, the actual and recorded transients may differ widely. For this reason, it is strongly recommended that the unfiltered response from the transducer be recorded, as well as the filtered signal. The introduction of filtering into the measurement chain can cause phase shifting of the data. In many transient measurement applications, only certain frequencies (within the Fourier spectrum of a pulse) are of interest. Unwanted noises or system transients of one kind or another must be eliminated. In these cases filtering is introduced in the system to allow only the desired frequencies to pass. Even though the filter passes the desired components without attenuation, the phase relation of a signal at one frequency with respect to a signal at another may be changed; this causes an error in the composite wave shape. To minimise this, the filter should be designed to have a constant time delay within its passband. In this way the phase relation between the frequency components of a transient is maintained and only a time shift of the entire composite wave shape will occur. 7.5. Special Considerations For Air Blast Measurements To determine the effects from explosive detonations, pressure measurements are taken of the airblasts. Four important pressure measurements are associated with this: (1) static overpressure, (2) reflected overpressure, (3) total pressure, and (4) impact, or dynamic pressure. Each measurement requires different techniques, however, all have similar peculiarities because of the measurement environment. As an example, Figure 7-4 below shows a pressure measurement record from a typical air blast.

Figure 7-4: Pressure Transducer Response to an Air Blast


In essence, air blasts involve shock waves. A shock wave is defined as a pressure wave characterised by a very steep, almost discontinuous, rise in pressure which occurs when a region of high pressure overtakes a region of low pressure, with a consequent rapid compression of the medium. The duration of a shock wave is distinguished by two phases. First, there is the positive (or compression) phase during which the pressure rises very sharply to a value that is higher than ambient and then decreases rapidly to the ambient pressure. The positive phase for the dynamic pressure is somewhat longer than for overpressure, due to the momentum of the moving air behind the shock front. The duration of the positive phase increases and the maximum (peak) pressure decreases with increasing distance from an explosion of given energy yield. In the second phase, the negative (or suction) phase, the pressure falls below ambient and then returns to the ambient value. The duration of the negative phase is approximately constant throughout the blast wave history and may be several times the duration of the positive phase. Deviations from ambient pressure during the negative phase are never large, and they decrease with increasing distance from the explosion. A transducer that is exposed to a step pressure change will be excited at its resonance frequency if the rate of pressure rise is sufficiently fast. The rise time for a shock wave will vary, depending upon the blast source and intensity, distance from source and gas dynamic properties of the shocked medium. Nonetheless, rise times can sometimes be less than a microsecond. Damped oscillations will show up on the output of virtually any air blast pressure measurement that is not overdamped since very few (if any) sensing elements have natural frequencies that would not be excited by high frequency energy in the shock wave. Many Kulite pressure transducers have natural frequencies of 500 kHz or higher. Referring to the example above, this ringing response is clearly evident in the initial response of the transducer. Sources of this oscillation may be mechanical resonance of the transducer sensing element, acoustical resonance of ports and cavities, or aerodynamic oscillations in the flow set up by diffraction of the initial and succeeding shocks and rarefaction waves over the probe. Three major problems are common to all pressure measurements in the extreme close-in region to air blasts: (1) survival of the gage and cabling, (2) thermal isolation of the sensor, and (3) minimising effects of mechanical shock motion on the transducer. Survival needs include protection from small, and even large, particles which are moving at high speed towards the pressure transducer. To provide the high frequency response, the transducer must be coupled to the pressure in close-to-flush mounted configuration, but the diaphragm :must also be protected from high speed particles. This is sometimes accomplished by using a mechanical screen or baffle (a plate with offset holes in front of the diaphragm ). Thermal protection of the diaphragm is also required for a short time period. Carbon-filled vacuum grease is often placed in front of pressure transducer diaphragms to accomplish this. Mechanical shock motion can be at thousands of g'S. To minimise errors, acceleration sensitivity of the transducer must be low along with a high resonance frequency. 7.5.1. Rise and Response Times Pressure transducers provide the user with the capability of monitoring extremely rapid rise time pressure pulses. The rise time of the transducer is much faster than the period to which it will respond accurately. 7.5.2. Spatial Averaging of Pressure Across Diaphragms Static overpressure is the pressure a fluid exerts normal to the surface along which it flows. The stagnation pressure is the sum of the static overpressure and the pressure attributable to the kinetic energy of the flowing fluid. Measurement of the stagnation pressure usually involves orienting the diaphragm of the pressure transducer normal to the direction of fluid flow. The dynamic model historically presented to describe a flush-mounted circular diaphragm pressure transducer measuring stagnation pressure is that of a spring-mass-damper system mathematically represented by an ordinary linear second-order differential equation possessing constant coefficients.


In reality, diaphragms of the type being discussed typically deflect in accordance with theory applicable to rectangular plates. This theory does not predict only one natural frequency, but an infinite number of natural frequencies for the various plate vibration modes. Practically, only a finite number of plate natural frequencies are of importance due to damping in the plate structural material. When a pressure transducer is flush-mounted, if static overpressure measurements are for a travelling pressure wave, the spatial averaging effect of the transducer diaphragm can result in a dynamic response model different from that represented by just the mechanical system. This can be visualised by referring to Figure 7-5 below. At high frequency the wave lengths of an acoustic pressure wave become quite short. For example, given a sound speed in air of 300 m/s and a frequency of 10,000 Hz, the corresponding wave length is 0.033 m (or 1.3 inch). To prevent extreme pressure averaging one can see that the diameter of the pressure sensor should be many times less than the wave length.

Figure 7-5: Pressure Wave Averaging Over Pressure Transducer

Calculations have been completed which permit estimates of the averaging effect. Figure 7.7 shows this transfer function, The abscissa of Figure 7-6 must be multiplied by the velocity of propagation of the pressure wave traversing the diaphragm of the transducer to dimensionally change its scale factor to Hz. When using this for the above example, the averaging results in a -5% error. for a transducer diameter of about 0.13 in.

Figure 7-6: Transfer Functions when Analysed as Spatial Averaging Transducers

7.5.3. Mechanical Protection


In addition to protecting the diaphragm from flash and from particle impingement, the transducer case and cable must be protected from physical damage. This is best done by proper precautions in mounting and cable routing. The design of the mounting structure should allow for its dynamic mechanical response to the shock wave. Mounting threads should be fully engaged, but contact with the relieved area at the front of the transducer should be avoided. Installation torque recommendations should be observed as well as recommended specifications for mounting hole dimensions and perpendicularity. Only the 0-rings or crushwashers provided or recommended by Kulite should be used for sealing. The cable should be restrained to prevent whipping and flexing. Cable tension should be avoided by providing strain relief cable routing and/or clamping. The cable should also be protected from crushing. Where the cable might be subjected to high pressure pulses or shock waves, it should be protected in a rigid conduit.


Section 8 – Calibration Complete calibration requires more than just determining the sensitivity at one or more values of input pressure. Kulite tests 100% of production, and supplies calibration data on the most important static input, electrical, and thermal characteristics. Dynamic characteristics are established from periodic sampling tests. All specifications and calibrations are in accordance with applicable ANSI and ISA standards. 8.1. Temperature Calibrations Although Kulite’s passive compensation technique reduces thermal zero shift and thermal sensitivity shift to very low values, there is some change of performance over temperature. If these inaccuracies are quantified, data can be corrected for improved accuracy. Kulite provides a test report giving measured data for each transducer. Not only is the maximum value tabulated, but curves plotting several temperature points can be provided to allow the user to correct zero and sensitivity data at any temperature in the operating range. 8.2. Electrical Calibrations Bridge resistances and isolation resistance are excellent indicators of the health of a transducer (see Quick Checks in Section 10.1). Input and output resistance and isolation resistance are measured on each unit and included on the test report. The values provided are at room temperature; input and output resistance may shift significantly with temperature variations. Therefore, variations of as much as 25% from the room temperature value on the test report should be considered acceptable. 8.3. Static Calibrations Ranges up to 1000 psi are calibrated on computer-controlled automatic test facilities, using gas as the pressure medium. Ranges above 1000 psi range are calibrated manually using oil or alcohol media. The calibration standards all have accuracies better than ± 0.01%. Computer-controlled calibrations are performed at zero and 20% .increments of full scale to 100%. Manual calibrations are performed at zero and 25% increments of full scale. For all transducers the test data is fed into a computer which calculates the test parameters, plots nonlinearity, thermal zero shift and thermal sensitivity shift, and prints out the test report. The computer then compares the test data with stored specification limits and accepts or rejects each transducer. Measurements are first made at zero pressure, then 2 times full scale, then zero, to establish zero shift after 2 x FSO. Then two complete cycles from zero to full scale and return are performed, measuring output at 20% (or 25%) increments. These data points are used for calculation of nonlinearity, hysteresis, and non-repeatability. All output measurements are made typically with 10.0 Vdc excitation applied. For aircraft pressure transducers there will be an agreed Acceptance Test Procedure (ATP) which is carried out on each transducer. 8.3.1. Dead Weight Testers Dead weight testers are primary standards that are used to generate precise pressures for calibrating pressure measuring instruments. They are primary standards because the factors influencing the accuracy of the generated pressure are traceable to standards of mass, length, and time. The accuracy of the pressure generated by a dead weight tester is dependent only on the accuracies of the weight and area measurements at operating conditions.


Corrections are often applied for 1) local acceleration of gravity, 2) air buoyancy, 3) change of piston area caused by temperature, 4) change in effective area due to pressure, and 5) difference in height of test instrument. Consideration of the magnitude of these corrections under actual operating conditions may indicate that some, or all, can be ignored. 8.4. Dynamic Calibrations Most pressure measuring instruments are calibrated statically in order to achieve the greatest possible accuracy. However, the accuracy available in dynamic measurements cannot be extrapolated from only static calibrations. The prime reasons for conducting dynamic pressure calibrations are to evaluate the frequency response of a measurement system used in a dynamic application. To complete this calibration, either a continuous wave (periodic) input or transient input (step, pulse, etc.) can be used. Calibration uncertainties for both of these approaches are much larger than for the static calibration approaches, so they are not generally used for acceptance testing of static responding transducers. Several of the more commonly used methods of dynamic calibration are described below: 8.4.1. Oscillating Pressure Calibrations Comparison pressure calibrations can be performed dynamically with a range of sinusoidal pressure generators, four of which are described briefly here: 8.4.1.1 Hydraulic Pressure Generator This device incorporates a compression spring, piston and seismic mass assembly, hydraulic oil-filled chamber, and mounting cavities for the reference standard and test transducers. Sinusoidal vibratory motion applied to the generator housing imparts sinusoidal pressure to oil which is exposed simultaneously to the sensing surfaces of both transducers providing a direct comparison calibration capability.. 8.4.1.2. Vibrating Liquid Column Sinusoidal vibration of a vertically-mounted liquid column provides a dynamic pressure which is applied to a test transducer mounted at the bottom of the tube. By attachment to an electrodynamic vibrator, short liquid columns can be used to provide about ±5 psi from about 50 Hz to 2000 Hz. Amplitude is limited since the approach is only linear over ±1%. Low frequency is generally limited by the vibrator and the high frequency is limited by the resonance frequency and damping of the liquid column. Calibration errors are better than ±4%. One advantage of this technique is that it is a primary calibration (since it is not simply a comparison to another transducer). Because of its limited amplitude and frequency range, the vibrating liquid column is seldom used. 8.4.1.3 Inlet Modulated Pressure Generator The Inlet Modulated Pressure Generator (IMPG) consists of a wheel with holes drilled through its periphery which is rotated at high speed. Air is blown through the holes from one side of the wheel and there is a cavity on the opposite side of the wheel in which are located the transducer under test and a reference transducer. The frequency of the signal generated is directly proportional to the speed of rotation. Frequencies of up to 12 kHz can be generated with pressure amplitudes of 1 Bar at 1 kHz falling to 0.1 Bar at 12 kHz. Static mean pressures can be generated of up to 7 Bar. 8.4.1.4 Gulton Whistle The Gulton whistle consists of a tube which is sharp edged at one end and is closed with a moveable piston at the other. Air which is blown over the edge of the tube excites the first organ pipe resonance. The resonant frequency is adjusted by the position of the piston within the tube. Mounted in the piston are the reference pressure transducer and the transducer under test. Figure 8-5. Frequencies of up to 4.5kHz can be generated with pressure amplitudes of up to 0.1 Bar. Static mean pressures can be generated of up to 20 Bar.


8.4.1.5 Gas Pistonphone Periodic pressure oscillations are generated by a piston moving in and out of a small gas cavity. Using a motor to drive a piston, pressures of 124 dB SPL are obtained at 250 Hz. Accuracy is ±2.5% when calibrated optically to determine piston displacement. The dynamic pressure is a function of the relative volume change. 8.4.2. Step Pressure Generators 8.4.2.1. Fast Acting Valves Liquid: To provide a dynamic calibration at high pressures a convenient and safe method is to quickly release the pressure in a closed liquid system. Figure 8.9 shows such a method using an oil dead weight system and a fast acting ball valve. Pressure can be ramped from one value to another within about I ms using this method. For this approach to be accurate the measurement system must have flat frequency response for the frequency content of the pulse. The starting and ending static pressures are accurately known. Care must be taken to protect the dead weight tester. Also, note that the pressure change is negative-going, which is not representative of many usage applications. Valve action may not be highly repeatable. However, this method provides a good comparison or evaluation test. Gas: Step pressures in gas can also be achieved by opening fast-acting valves between gas pressure vessels. This method may be preferable for transducers which are incompatible with liquid media. However, greater care must be taken in designing and operating a high pressure gas system because of the greater danger of explosion. 8.4.2.2. Gas Shock Tubes Small shock tubes are often used to provide rise time and frequency response characteristics for transducers. Because of difficulties in determining the pressure level in the step, shock tubes are not usually used for pressure sensitivity calibration. Pressure rise times of about 1 microsecond are practical, which permits transducer characterisation to frequencies beyond 100,000 Hz. Kulite uses shock tubes to determine the frequency response of all diaphragm designs.

Figure 8-10: Pressure Shock Tube

The excitation source is a shock tube which has a 2.5-inch diameter cylindrical cross section with a 15-inch driver end and 60-inch driven end. The tube sections are mechanically coupled by a bolted flange which can be disengaged to enable insertion of a diaphragm material as a separator. The driver section is pressurised with a gaseous medium until the test pressure is reached. The membrane is then punctured by a pneumatically actuated needle from outside the shock tube. This results in the sudden release of the pressurised air into the lower pressure driven compartment and produces a hypersonic shock wave front which impinges the end plate in which the pressure transducer is flush-mounted. The transducer diaphragm is thus exposed to a very fast rise time pressure step which has significant high frequency content, such that extended frequency response information is available.


Section 9 – Glossary, Unit Conversions & Kulite Reports 9.1. Glossary of Terms

A Acceleration Sensitivity (Error) The maximum difference at any measurand value between the output with and without the application of a specified constant acceleration along specified axes. Normally specified in transverse and perpendicular directions. (% FSO/g) Accuracy The ratio of error to Full Scale Output often expressed in percent of Full Scale output. (%FSO) Acceptor A material added as a dopant to a semiconductor to make it p-type by accept-in valence electrons and leaving holes behind that can conduct electric charges. Normally Boron is used for doping Silicon. A/D Abbreviation for analogue to digital A/D Converter An electronic device used to convert an analogue voltage to a digital signal. Altitude The vertical distance above a specified reference datum such as sea level. Ambient Conditions The conditions (Pressure, temperature, etc.) of the device's environment. Analogue Output An output of analogue voltage derived from processing of digital and/or analogue input to circuitry within an (electronic) device. Normally a continuous function of the measurand except as modified by device resolution. Angstrom A unit for Ultra-high frequencies equal to ten raised to the power -10 meters. Annealing A heating process that reverse damage to the crystal structure or to activate dopant. Attitude Error The error due to the orientation of the device relative to direction of the force of gravity on the transducer. --------------------------------------------------------------------------------

B Bandwidth The highest frequency signal component that can pass through input amplifiers without being attenuated. Best Fit Straight Line A line midway between two parallel straight lines closest together and enclosing all output points of the instrumentation between zero load and full scale. Breakdown Voltage Rating The voltage (AC or DC) which can be applied across the transducer insulation without causing arcing or conduction above a specified current. Bridge Resistance Resistance of the transducer bridge elements ( input or output impedance). Burst Pressure Rating The pressure which may be applied to the sensing element or the transducer case just prior to rupture of either. A minimum number of applications and time duration are also often specified. --------------------------------------------------------------------------------

C Calibration A test to determine the output signal of the device with a steady input excitation and known values of the parameter under controlled conditions Calibration Curve A graphical representation of the device output vs. measurand under controlled conditions.


Calibration Cycle The application of controlled values of a parameter proving an output signal over the full range of the instrument in ascending and descending order. Compensated Temperature Range Range of Temperatures over which the transducer has been corrected by the addition of a circuit to correct the output for errors induced by changes in bridge resistance due to temperature. Compensation Addition of supplemental device, circuit, or special materials to reverse known source of errors. Combined Linearity and Hysteresis Square of the sums of squares of error due to non-linearity and that of and hysteresis non-linearity. --------------------------------------------------------------------------------

D D/A Abbreviation for Digital to Analogue. D/A Converter A device for converting a digital signal to an analogue signal. Damping An energy absorbing factor that in conjunction with the natural frequency determines the limit of frequency response and the response time characteristics of a transducer. In response to a step function of the parameter a periodic (underdamped) system oscillates about the voltage level before stabilising at its final steady output; an aperiodic (overdamped) system comes to the final steady output without overshooting; and a critically damped system is defined as one that is at the point of changing from a periodic to a aperiodic system. Dead Volume Total volume of the pressure port cavity of a transducer with room barometric pressure applied. Decibel A unit of logarithmic measure based on the ratio of power related quantities such as sound, volts, or watts to a specified reference in same units. Deposition The procedure of deposit materials onto a substrate by means of vacuum, electrical, screening, or vapour techniques. Dice or Die A section of a processed wafer, usually rectangular, which contains one functional circuit. Dielectric An insulating layer. A material that has high resistance. Dielectric Strength Same as breakdown voltage. Diffusion A process used in semiconductor production by adding small amounts of impurities or dopants to a semiconductor. Digital Output Transducer output that represents the magnitude of the parameter measured in terms of discrete quantities or codes in a system of notation. 0 and 1 are commonly used. Digital and analogue are common output types. Donor A material added as a dopant to a semiconductor to make it n-type by donating valence electrons which can conduct electric charge. An example is phosphorus. Dopant A material added in minute quantities to a semiconductor to alter it's electrical conducting characteristics. They may be donors or acceptors. Doping The process of adding a dopant to semiconductor material. Drift A undesired change of a reading with no charge in the input signal or operating conditions. Dynamic Characteristics The characteristics of a transducer which describes its response to variations in measurand pressure over time. --------------------------------------------------------------------------------

E Electrical Connection The portion of the transducer assembly used to connect, disconnect, and reconnect the electrical wiring that carries excitation voltage, signal or current to and from the transducer.


Environmental Conditions Specific external conditions, such as shock, vibration, temperature, moisture, etc. to which a transducer may be exposed during normal operations. End Point Output signal at upper and lower limits of the transducer range. End Point Line Line drawn between the end points of a transducer calibration. Environmental Conditions Specified external conditions ,such as shock, vibration, temperature, moisture, etc. to which a transducer may be exposed during normal operations. Environmental Conditions, Operating Specified external conditions ,such as shock, vibration, temperature, moisture, etc. when a transducer is exposed to must perform as specified. Error The mathematical difference between the indicated value and the true value of the parameter signal. Error Band The band of maximum deviations of output values from a specified reference line or curve due to those causes attributable to the transducer. Usually described as plus or minus some value of transducer full scale output. Excitation The external electrical voltage and/or current applied to a sensor assembly to initiate the proportional output. Usually expressed in ranges the transducer may subjected to without damage. The value of this voltage is set by calibration and any serious deviation from this value will negate the calibration. Extrinsic Semiconductor A semiconductor that has been doped either n-type or p-type. Electrons and holes are present in unequal proportions (by 4 to 8 orders of magnitude). --------------------------------------------------------------------------------

F Frequency Modulated Output An output in the form of frequency deviations from a centre frequency, where the deviation is a function of the measured parameter. Frequency Output An output in the form of frequency which varies as a function of the applied measurand. Frequency, Natural The frequency of free (not forced) oscillations of the sensing element of a fully assembled transducer. Frequency, Resonant The input frequency at which a transducer responds with maximum output amplitude. If there is more than one frequency the lowest is the resonant frequency Frequency Response The change with frequency of the output/parameter amplitude ratio( and of the phase difference between the output and the parameter) for sinusoidal varying measurand applied to a transducer within a specified range. It is normal consider at plus or minus 3 dB and is given by the approximation. F.R.= 1/2pt Freq. Response = 1 divided by 2 times Pi times time constant Full Scale Output The value of transducer output at the maximum rated load minus the output at the minimum rated load. --------------------------------------------------------------------------------

G Gage Factor A measure of the ratio of the relative change of resistance to the relative change in length of a resistance strain transducer (strain gage). --------------------------------------------------------------------------------


H Hysteresis The maximum difference in output, at any measurand value within the specified range, when the value is approached first with increasing and then decreasing measurand. Normally expressed in % FSO. --------------------------------------------------------------------------------

I IEEE Abbreviation for Institute of Electrical and Electronics Engineers. Input Measurand signal (and/or exciting voltage or current). Input Impedance The impedance (presented to the excitation source) measured across the excitation terminals of the transducer. Unless otherwise specified the impedance is measured at room temperature, etc. Insulation Resistance The resistance measured between specified insulated portions of a transducer when a specified DC voltage is applied at ambient conditions- room temperature etc. Ion The result of an atom losing an electron and becoming positive or gaining an electron and becoming negative. --------------------------------------------------------------------------------

J, K, L Leakage Rate The maximum rate at which a fluid is observed or permitted to leak through a seal. The type of fluid, differential pressure across the seal, and the direction of flow should be specified. Units are normally volume or pressure drop per unit of time. Least Squares Line The straight line for which the sum of the squares of the residuals( deviations) is minimised. Life, Cycle The specified number of full and/or partial range excursions over which a device will operate within specified performance criteria. Life, Operating The length of time over which device will operate to a specified performance. Life, Storage The length of time over which device can be stored at specified conditions and still operate to a specified performance. Linearity The closeness of a calibration curve to a specified straight line expressed as % FSO. Load Impedance The impedance presented to the output terminals of the transducer by the external circuitry connected to the device. --------------------------------------------------------------------------------

M Maximum Ambient Temperature The value of the highest (and the lowest) ambient temperatures that a (Minimum) transducer can be exposed to with or without excitation applied, without being damaged and subsequently suffering performance degradation. Maximum Excitation Value of excitation voltage or current that can be applied to the transducer at room conditions without causing damage or performance degradation. Measurand (Parameter) A physical quantity, property, or condition, which is measured, Sometimes called input, parameter, or variable. Measured Fluid The fluid that comes in contact with the sensing element. Chemical and/or physical properties of this fluid may be specified to insure proper transducer operation and life.


Mounting Error An error due to mechanical deformation of the transducer caused by mounting the device and/or making the electrical connection. --------------------------------------------------------------------------------

N Natural Frequency See Frequency, Natural Non-Linearity See Linearity. Normally expressed as % FSO. Non-Operating Conditions Any conditions outside of operating conditions that might cause transducer to malfunction. Non-Repeatability Breakdown of the transducer characteristics of repeatability a small amount of which is contained in an acceptable error band. Usually expressed as % FSO. Null A condition, such as of balance, which results in a minimum absolute value of output. --------------------------------------------------------------------------------

O Operation Mode A description of how the transducer is used to provide a usable signal representing a measurand variation. Operating Conditions See environmental conditions. Operating Temperature Range Temperature range the operating transducer will be subjected to. Output The electrical quantity produced by the transducer which is a function of the applied measurand. Output Impedance The impedance across the output terminals of a transducer presented by the transducer to the associated external circuitry. Output Noise The RMS or peak to peak, as specified, ac component of a transducer's DC output in the absence of measurand variations. Output Range Design band of output for specified input. Overload The maximum value of a measurand that can be applied to a transducer without damage or change in performance beyond a specified tolerance. Over Pressure See Overload. --------------------------------------------------------------------------------

P Pressure Range Lowest to the highest pressures to be measured by a specific transducer. Pressure Media See measured fluid. Proof Pressure The maximum pressure, which may be applied to the sensing element without a change in the transducer performance beyond specified tolerances. Differential pressure transducers must have the reference pressure specified and whether the reverse pressure is applicable. --------------------------------------------------------------------------------

Q, R Range The measurand values over which a transducer is designed to measure. Indicated by upper and lower values.


Rated Electrical Excitation The electrical voltage supply the transducer sensing element for normal operation given the specified output with the application of a known measurand. Reference Pressure The pressure applied to the opposite side of the sensing element- ambient pressure for gage and perfect vacuum for absolute designs. Reference Pressure Error The error in transducer output resulting from errors in the reference pressure value within a specified reference pressure range. Reference Pressure Range The range of pressures that can be applied to the backside of a sensing element without changing the pressure transducer's performance beyond specified tolerances. If no tolerance is specified, none is allowed. Reference Pressure Sensitivity Shift The sensitivity shift resulting from variations of a differential pressure transducer's reference pressure within specified limits. Repeatability The ability of a transducer to reproduce output values when the same measurand value is applied repeatedly under the same conditions and in the same direction. Reproducibility See Repeatability Residual Unbalance Zero measurand sensing element output Resolution The magnitude of output step changes as measurand is continuously varied over the range. Resonance Amplified vibrations of transducer components, within a narrow frequency band, observable in the output as a vibration applied along a specific transducer axis. Resonant Frequency The input frequency at which a transducer responds with maximum output amplitude. If there is more than one frequency the lowest is the resonant frequency Response Time The length of time required for the output of a transducer to rise to a value normally specified as 98 % of the value of a step change in measurand expressed in milliseconds. Rise Time The length of time required for the output of a transducer to rise from a small specified percentage of it's final value to a large specified percentage of it's final value as a result of a step change in the measurand. Room Conditions Normal specified ambient conditions. Normally 77 F + or - 18 F, 90 % RH, and 29 + or - 3 in Hg. --------------------------------------------------------------------------------

S Self-Heating Internal heating as a result of electrical energy dissipated within the transducer. Sensing Element The part of the sensing element that responds directly to the measurand. Sensitivity The ratio of the change in transducer output to a change in the value of the measurand. Sensitivity Shift A change in the slope of the calibration curve due to a change in sensitivity. Sensor Instrumentation device, such as transducer. Sound Pressure Level (SPL) A unit that is 20 times the logarithm to the base 10 of the ratio of the pressure of the measured sound to the reference pressure of 20 micronewtons per square meter. Span The algebraic difference between the limits of range. Stability The ability of a transducer to retain its performance characteristics for relatively long period of time. Normally expressed in % FSO. Static Calibration A calibration performed under room conditions and in the absence of any vibration, shock, or acceleration. Static Error Band See Error Band. --------------------------------------------------------------------------------

T Temperature Error Band The error band applicable to increased or decrease environmental temperature usually expressed in % FSO/ 100 F.


Temperature Range, Compensated See Temperature range, operating. Compensated temperature range is the interval of temperature range that was considered when designing compensation module for a specific transducer. Temperature Range, Fluid The rang for temperatures of the measured fluid, when it is not the ambient within which operation of the transducer is intended, and all specific tolerances for the temperature error band apply. Temperature Range, Operating The range of temperatures, given by their extremes, within which the transducer is designed to operate with no permanent damage to the transducer. Thermal Coefficient of Resistance (TCR) The relative change in resistance of a conductor or semiconductor per unit change in temperature over a stated range of temperature normally expressed in ohms per degree F or C. Thermal Compensation The addition of circuitry to alter output changed by temperature error back toward specified values at room temperature. Thermal Sensitivity Shift The sensitivity change due to changes of the ambient temperature from room temperature (design conditions) toward the limits of the operating temperature range. Thermal Zero Shift The zero measurand transducer output shift due to changes of the ambient temperature from room temperature to the specified limits of the operating temperature range. Transducer A device which provides a usable output in response to a specified measurand. Transient Response The response of a transducer to a step change in measurand. It is indicated by Time Constant, Ringing Period, and Response Time. Transverse Acceleration An acceleration perpendicular to the sensitive axis of the transducer. --------------------------------------------------------------------------------

U, V Vibration Error The maximum change in output, at any measurand value within the specified range, when vibration levels of specified amplitude and range of frequencies are applied to the transducer along specified axes. Vibration Sensitivity See Vibration error. --------------------------------------------------------------------------------

W, X, Y, Z Warm-up Period The period of time required from the time the excitation voltage is applied to the transducer until the output of the unit is within the specified tolerances for the applied measurand. Weight Weight of the transducer normally not including any portion that varies such as cables.


9.2 Pressure Unit Conversions 9.2.1 Units of Measurement

9.2.2 Decibel Formulae


9.3 Kulite Reports 9.3.1 Andrew Bemis “High Pressure Transducers” 9.3.2 Dr. A. D. Kurtz, A. A. Ned, Glenn Beheim “6H-SiC Pressure Sensor for High

Temperature, Low Pressure Applications”. 9.3.3 Dr. A. D. Kurtz, Boaz Kochman “Dual Resonant Beam Pressure Transducer”. 9.3.4 Joe VanDe Weert “Fuel Pump Transducer”. 9.3.5 Adam Kane “Miniature combination Pressure/ Temperature Sensors with Redundant

Capability”. 9.3.6 Wolf Landmann “Solid State Replacements for Electro-Mechanical Pressure

Transducers”. 9.3.7 Dr. A. D. Kurtz, Dr. J. W. H. Chivers, A. A. Ned (Kulite) & Professor A. H. Epstein

(MIT) “Sensor Requirements for Active Gas Turbine Engine Control” 9.3.8 Adam Kane “High Accuracy DC Combustion Transducers” 9.3.9 Andrew Schwartz “Subminiature Load Cells”. 9.3.10 Dr. A. D. Kurtz, Joseph VanDe Weert & Boaz Kochman “Static Pitot Transducer”. 9.3.11


Documents

Kulite Pressure Transducer Handbook