HP 71500A MicrowaveTransition AnalyzerGroup Delay Personality

Product Note 70820-10HP 71500A/70820AProduct Note Series

Measure Group Delay and Phasethrough Frequency Converterswith Internal Local Oscillators

• Accuracy to .2 ns• Uses envelope delay technique, with AM or FM • Access to internal LOs not required• Coverage to 40 GHz, without need for reference

mixers or external converters

HP 71500A Group Delay Personality

Measurement Block Diagram

HP-IB

DigitalPhaseComparator

HP 71500A with IBASIC

CH1

CH2

RF carrier

Modulator AM or FM

Receiver

InternalLO

Modulation source

(AM shown)

DigitalDemod

This literature was published years prior to the establishment of Agilent Technologies as a company independent from Hewlett-Packard and describes products or services now available through Agilent. It may also refer to products/services no longer supported by Agilent.We regret any inconvenience caused by obsolete information. For the latest information on Agilent’s test and measurement products go to:

www.agilent.com/find/products

Or in the US, call Agilent Technologies at 1-800-452-4844 (8am–8pm EST)

Table of Contents

Introduction 3Product Description 3Why Measure Group Delay? 3Group Delay Techniques 3Direct Phase 3Modulation Delay 4Key Features and Benefits 4

Measurement Overview 5Measurement Block Diagram 5Modulation Source 5Synthesizer Support 5Measuring in the Frequency Domain 6Direct Demodulation Mode (Zoom FFT Off) 6 Zoom FFT Mode 7 Marker Options (Zoom FFT on) 8Normalization 9How to do normalization 9What frequency span to use? 9Unambiguous delay range 10Negative delay 10Measurement Times 11

Using the Personality 11Obtaining the Program 11Firmware Upgrades 11Downloading the Program 11Loading from an external computer 12Storing to a memory card 12 Loading from a memory card or disk drive 13Menu Overview 13 Setting Up and Doing a Measurement 14Printing/Plotting the Results 16

Accuracy Considerations 17Typical Accuracy 17Noise Reduction 17Averaging 17Smoothing 17Signal-to-Noise Ratio 18Mismatch Error 18 LO Stability 18

Modifying the Program 19Manuals 19 Changing Defaults 19Adding Sources 19

Appendix A 20 Comparison to Other Instruments 20Vector Network Analyzers 20Scalar Network Analyzers 20Microwave Link Analyzers 20HP 11770A Link Measurement Personality 21Summary 21

3

Introduction

Measuring absolute group delaythrough frequency translatingdevices such as downconvertersor receivers is difficult at best,particularly at microwavefrequencies. Vector networkanalyzers are excellent tools formeasuring group delay relativeto a “golden device”, but theyare difficult to use for absolutemeasurements. For microwavedevices without access tointernal local oscillators (LOs),the problem is especiallydifficult as vector networkanalyzers cannot be used at all.For these cases, a modulationtechnique must be used, whichoften involves a custom solutionwith external mixers, filters,LOs, and discriminators ordetectors — in other words,complex and expensive.Customers have been asking fora better solution for a long time.

The HP 71500A microwavetransition analyzer coupled witha downloadable InstrumentBASIC (IBASIC) personalityoffers a cost-effective and simplesolution for microwave groupdelay measurements offrequency translating devices.

Product DescriptionThe group delay personality is adownloadable IBASICapplication program that runsinternally in the HP 71500A. It is targeted for measurementsof microwave frequencyconverters and receivers withinternal LOs, or any microwavefrequency converter whereabsolute group delay is desired.It is intended for measuringsub-systems or systems wherethe source and the HP 71500A

can be located near one another.It is not intended for end-to-endmeasurements separated by along-distance, due to the needfor the HP 71500A to measureboth the RF and the basebandmodulation source.

The program uses the modulationdelay technique for measuringgroup delay, using either AM or FM. The modulated RFcarrier is measured in channel1, and the baseband modulationsignal is measured in channel 2.

Why Measure GroupDelay?Ideal networks pass modulatedsignals containing informationwithout inducing any additionaldistortion. They arecharacterized by both flatmagnitude response and linear(constant slope) phase shiftversus frequency. Since groupdelay is the derivative of phasewith respect to frequency, thisimplies a flat group delayresponse. All of the frequencycomponents of the signalspassing through the device-under-test (DUT) willexperience the same amount oftime delay. The real world,however, is composed of non-idealnetworks, which causedistortion due to amplitude andphase nonlinearities. Groupdelay is one measure of phasenonlinearity. It is particularlyimportant to characterize groupdelay for systems where sometype of phase modulation isused, such as FM, QPSK, or64QAM.

Group DelayTechniquesGroup delay is definedmathematically as the negativeof the derivative of a network’sphase versus frequencyresponse:

Group delay td =– ––dØdw

(td in seconds, Ø in radians,w in radians/second)

ortd=–––1

360x dØ––

df(td in seconds, Ø in degrees,

f in hertz)

There are two basic techniquesfor measuring group delay of acomponent or sub-system.

Direct Phase

For linear, non-frequencytranslating devices, the mostcommonly used technique is thedirect phase method. A vectornetwork analyzer is used tomeasure the phase response of a DUT, and a numericaldifferentiation is done to derivegroup delay. The derivative isapproximated by using ,where delta phase is the phasedifference between adjacenttrace points, and deltafrequency is the frequency stepsize (total frequency spandivided by the number of tracepoints). Since delta frequency isa nonzero quantity, there isaveraging occurring over thisfrequency interval. This intervalis known as the measurementaperture. Larger apertures giveless group delay resolution(more averaging occurs), butless trace noise. Conversely, asmaller aperture implies moreresolution at the expense of anoisier measurement.

∆Ø∆ƒ

4

DigitalPhaseComparator

HP 71500A with IBASIC

CH1

CH2

RF carrier

Modulator AM or FM

Receiver

InternalLO

Modulation source

(AM shown)

DigitalDemod

HP-IB

Figure 1: Measurement block diagram.

The direct phase techniquerequires that both the test andreference signals be at the samefrequency in order for phase tobe measured. For measuringfrequency translating devices, areference DUT or externalmixer is needed to ensure thiscondition.

Modulation Delay

The second technique measuresmodulation delay. A modulatedcarrier is passed through theDUT, and then demodulated.The phase of the demodulatedsignal is compared to theoriginal modulating signal todetermine the delta phase atthat particular frequency. Thecarrier must be stepped acrossthe frequency band of interestfor the complete group delaymeasurement. The modulationtype can be AM, FM or PM.Group delay is calculated as:

td = –––––––Øe

360x fm–

(td in seconds, Øe{envelope phase shift} indegrees, fm {modulationfrequency} in hertz)

The HP 71500A group delaypersonality uses this technique.The aperture of thismeasurement is equal to twotimes the modulation frequency.

Key Features andBenefitsThe key features and benefitsfor the HP 71500A solution aredescribed below. A moredetailed comparison to othergroup delay techniques isprovided in Appendix A.

Accuracy to better than 1 ns.Using normalization and zoomFFT processing, accuracy downto +/- 0.2 ns can be achieved.This is adequate for mostapplications.

Access to internal LOs notrequired. Since it is themodulation on the carrier thatis of concern and not the carrieritself, access to internal LOs orfrequency references isunnecessary. The vector networkanalyzer method measuresphase of the RF signals directly,and therefore requires LO orreference access.

No need for externalreference mixers. The vectornetwork analyzer approachneeds an extra external mixeror DUT to provide a referencesignal that is the same frequencyas the test signal. While thismethod provides excellentrelative measurements(referenced to a golden standard),absolute delay measurementsare difficult due to the need tocalibrate the mixers. The HP

71500A solution measuresabsolute group delay directly,providing a simpler and moretraceable measurement. Frequency coverage extends to40 GHz. No extra upconvertersor downconverters are needed tocover most applications, such as satellite transpondermeasurements. This lowers themeasurement system cost andcomplexity.

FM group delay. The FMdemodulation routine internalto the HP 71500A eliminatesthe need for an external FMdiscriminator. Receivers thathave internal limiters orsaturated amplifiers require theuse of FM, since any AM isstripped away. An example of areceiver with a saturatedamplifier is a typical Ku bandsatellite transponder.

Amplitude and phase. Theprogram also provides theamplitude and phase responseof the DUT with no additionalmeasurement time.

5

70004A 70820A

HP-IB

10 MHz Ref

CH 1

CH 2

HP 71500A with IBASIC

Attenuator20 dB

RF Out

Internal Modulation Opt. 002

Frequency Translating DUT

Internal LO

AM/FM Output(rear panel)

836xxA

Figure 2: Measurement setup using internal modulation.

Simplified user interface.Custom menukeys tailor theMTA’s user interface for makinggroup delay measurementswithout the need for an externalcomputer.

Limitations

The key limitations of thisapplication are that it is slowerthan most other solutions, andthe relative accuracy, whileadequate for most measurements,may not be as good compared to what customers have come toexpect from instruments suchas the HP 8510 vector networkanalyzer.

MeasurementOverview

Measurement BlockDiagramA block diagram view of themeasurement is shown in figure1. The modulated RF carrier ismeasured in channel 1 of theHP 71500A, while the basebandmodulation signal is measuredas a reference in channel 2. The modulation source can be astand-alone instrument orinternal to the RF source.

Modulation SourceThere are two basicmeasurement setups. The firstsetup (figure 2) uses the internalmodulation capabilities of thesource, such as (optionally)provided by the HP 8360 familyof synthesizers. The secondsetup (figure 3) uses an externalmodulation source, controlledeither manually or via HP-IB.Many non-HP-IB modulationsources have poor frequencyaccuracy. In some cases, it maybe necessary to fine-tune the modulation frequency of the

source to match the value usedin the program.

Note: when connectingsynthesizers or modulationsources that are to be controlledvia HP-IB, use the HP-IBconnector on the rear panel ofthe HP 70820A module, NOTthe HP-IB connector on the HP 70004A mainframe/display.Connecting the 10 MHzreferences of the HP 71500Aand the sources together isrecommended, but not required.

70004A70820A

HP-IB

10 MHz Ref

Attenuator

CH 1

CH 2

HP 71500A with IBASIC

RF Out

Frequency Translating DUT

Internal LO

836xxA

Modulation Input

Modulation Source

Figure 3: Measurement setup using an external modulation source.

Synthesizer SupportThe group delay programsupports the HP 836xx, HP 834x,HP 70340A, and HP 3325Asynthesizers. The program caneasily be modified to drive othersynthesizers if necessary (seeAdding Sources on page 19).

The group delay personalitydoes not require 1 Hz frequencyresolution of the RF source, butthis feature is very desirable forother measurements (such aspulsed-RF characterization) andis highly recommended if a newsource is being ordered.

6

Figure 4: Screen display using direct demodulation mode (zoom FFT off).The upper trace shows the demodulated RF carrier. The lower trace showsthe modulation signal. Both signals are in the frequency domain.

Measuring in theFrequency DomainThis program makes extensiveuse of the Fast FourierTransform (FFT) to perform thegroup delay measurement.FFTs are used to demodulatethe carrier and compute thephase of the baseband signal.Zoom FFTs are also used as anoise reduction technique. A zoom FFT takes a larger timerecord to get better frequencyresolution, and displays only aportion of the frequency domain data.

The program operates in twobasic measurement modes. Therecommended practice is to usethe zoom FFT mode. This mode yields the most accuratemeasurements. The user canselect the trade-off betweenaccuracy and measurementspeed. The greater the amountof zoom used, the more accuratelythe carrier and modulationfrequency must be known as well.

The direct demodulation modedoes not use zoom FFTs and is generally faster than the zoomFFT mode, but not as accurate.It is less sensitive to frequencyinaccuracy, and should be used when the carrier andmodulation frequencies are notcontrolled well enough to usethe zoom FFT mode.Trace averaging can be usedwith either mode to furtherimprove measurement accuracy. For a given level ofperformance, the zoom FFTmode is faster than using thedirect demodulation mode withaveraging.

Direct Demodulation Mode(Zoom FFT Off)

This mode uses the built-in FMdemodulation routine of the HP71500A. Trace 3, which showsthe demodulated carrier in thefrequency domain, is defined asFM(CH1), and trace 4, whichshows the modulation signal inthe frequency domain, isdefined as CH2. Both traces aretransformed into the frequencydomain using the FFT. Thedisplayed signals appear at thefar left-hand portion of thescreen (figure 4). The markerstake one reading on each trace,using the local peak function to ensure the signal peak ismeasured.

The RF signal is sampled suchthat the carrier is centered inthe IF, and the modulationsidebands are unaliased (notfolded about DC or the Nyquistfrequency). If they weredisplayed, they would appear asthey would on a spectrumanalyzer (figure 5).

When using the directdemodulation mode (zoom FFToff), phase versus frequencydata is available, however,amplitude versus frequencyinformation of the DUT is notmeasured. Amplitudeinformation is available whenusing the zoom FFT mode.

7

Figure 5: Screen display showing an FM modulated RF signal as it wouldappear in the IF using direct demodulation mode, but without the FMdemodulation function applied.

Zoom FFT Mode

With this mode, the userspecifies the amount of zoomused, from 1 to 128. Higherzoom values result in betteraccuracy, but slowermeasurement times. Largerzoom values are analogous tosmaller resolution bandwidthsin a spectrum analyzer. A typical value whichrepresents a good trade-off ofaccuracy versus measurementtime is 16. Each factor of twoincrease in the zoom FFT valuewill slow the measurement timedown by two as well. In thismode, trace 3 (the carrier) isdefined as CH1, and trace 4 (themodulation signal) is defined as CH2.

In the zoom FFT mode, the RFsignal is sampled so that thecarrier gets mixed to the IFnear DC, and half of themodulation sidebands are foldedaround DC such that sidebandpairs of the same order end upvery close together (for AM,there is only one pair ofsidebands). If this result wasdisplayed without the zoomFFT, it would look like figure 6.The zoom FFT is used to expandthe first order sidebands. ForFM, all sideband pairs higherthan first order are ignored.Markers are used to measurethe amplitude and phase of thesidebands, and the programuses this information to performthe necessary demodulation.Normally, if a modulationfrequency of 1 MHz is used, thefirst order sidebands will beseparated by 2 MHz. With thezoom FFT mode, the programsets the sampling frequency ofthe HP 71500A such that thesideband separation is reducedby about twice the zoom factor.

Carrier

1st order sideband pair 2nd order

sideband pair

3rd order sideband pair

4th order sideband pair

Area expanded by zoom FFT

Figure 6: Screen display of an FM modulated RF signal as it would appearin the IF using zoom FFT mode, but without the zoom FFT applied.

8

When the program is running,the top trace of the displayshows the first order sidebandpair, and the bottom trace showsthe baseband signal, but thistime the signals are centered onthe screen (figure 7).

With the zoom FFT mode, bothamplitude and phase versusfrequency can be measured,with no additional measurementtime.

Marker Options (zoom FFT on)

When the zoom FFT mode isselected, the group delayprogram defaults to placing themarker on signals using thehighest peak function. Thisallows the measurement to bedone even if the LO in the DUTdrifts somewhat, causing thesidebands to move around, or ifthe modulation frequency is notknown exactly (this is often thecase when using an externalsignal generator to generate themodulation signal).

If there are contaminatingsignals present, it is possiblethat the marker will find anincorrect signal which may bebigger than a desired signal.This can happen, for example,when the DUT producessignificant carrier harmonicsthat are not filtered away. Toavoid this, an option is availableto use direct marker placementand the local peak function. Inthis mode, the marker is set towhere the signal is expected,and a local peak is done toensure that the peak of thesignal is found. However, if the

LO frequency of the DUT hasdrifted enough, or hassignificant amounts of residualFM, or the modulationfrequency cannot be setaccurately enough, the markermay be placed too far from thesignal for local peak to workproperly. This will result in abad measurement. Directmarker placement with localpeak should only be used in anenvironment with very stableand accurately known signalfrequencies.

When using a manualmodulation source, it may benecessary to fine-tune themodulation frequency to matchthe value used in the program,to ensure correct markerplacement. If it appears thatneither marker mode is findingthe correct signals, then thedirect demodulation modeshould be used. The user canchoose between highest or localpeak, by using the markerssoftkey under the Optionsmenu.

Figure 7: Screen display using zoom FFT mode. The upper trace shows thefirst sideband pair of the modulated RF carrier. The lower trace shows themodulation signal. Both signals are in the frequency domain.

9

points, so set the start, stop,and step frequenciesaccordingly.

Care should be taken whenmeasuring frequency-invertingdevices, where the IF frequencysweeps downward in frequencyas the RF is swept upward. Inthis case, the reference and testsweeps as measured at the HP71500A input should both be inthe same direction (i.e., eitherboth upwards or bothdownwards in frequency). Thismeans that the RF sweepdirection will reverse from thereference to the test sweep.

What frequency span to use?

When performing anormalization on a frequencytranslating device, there aretwo choices for the frequencyrange to use for the referencesweep: either the frequencyspan corresponding to theDUT’s input, or to its output. Inboth cases, either the inputcable or the output cable will bemeasured during the referencesweep at a different frequencythan during the test sweep.This will not be a problem ifhigh quality interconnect cablesof the appropriate type for thefrequency range of themeasurement are used. Cablessuch as these will have flatgroup delay over the full RF andIF frequency range. Theircontribution to absolute delaywill then be removed by thenormalization process.

When using good cables, thechoice of which frequency spanto use for the reference trace issimple. For relativemeasurements, use thefrequency span corresponding tothe DUT output for establishingthe reference trace. This willremove any measurementirregularities due to the HP71500A (these are moresignificant at lower frequencies,especially below 300 MHz).

For absolute measurements, thefirst choice for establishing thereference trace is to use thespan corresponding to the DUTinput. This ensures that thedelay contribution of themodulator will be removed. Ifthe modulator delay isindependent of carrierfrequency, then thenormalization could be done atthe span corresponding to theDUT output, which has theadvantage of removing anyirregularities due to the HP71500A. The delay of some FMmodulators is also a function of the modulation frequency(this is true for the FMmodulator in an HP 8360 seriessynthesizer). If the modulationfrequency is changed after anormalization has been done, a new reference sweep isrequired.

NormalizationTrace normalization is desirablefor relative measurements, andrequired for absolute delaymeasurements. For obtainingthe best relative accuracy,performing a normalizationremoves the group delaycontributions of theinterconnect cables and the testequipment. An absolute groupdelay measurement can only bedone by first establishing areference, which is the groupdelay of the test setup withoutthe DUT in place. In general,normalization improvesmeasurement accuracyconsiderably, but is not as goodas what can be achieved usingcalibration standards and 12-term error correction as done instandard vector networkanalysis.

How to do normalization

Normalization is done by doinga measurement without theDUT, and then pressing theNORMALZ key under theProcess menu to store thereference trace. Themeasurement configurationmust be set up for a non-frequency translatingmeasurement, by setting the LOfrequency to 0 Hz and the RFand LO multipliers (N and M) toone, under the Config menu.Next, the DUT is inserted andanother measurement is done,this time setting the correct LOfrequency and multipliers thatcorrespond to the frequencytranslation of the DUT (see step3 on page 14 about the Configmenu for a more detaileddescription of setting up ameasurement). The referenceand test measurements musthave the same number of

10

For the zoom FFT mode, theunambiguous range is furtherreduced by a factor of two to0±(.25/fmod). If the absolutedelay is beyond these limits, itwill be displayed incorrectly.For example, if the absolutedelay of a DUT and test setupwas 273 ns, and the zoom FFTmode was used with amodulation frequency of 1 MHz,the program would measure–227 ns (273 ns - .5/1 MHz). Ifthe direct demodulation modewas used (zoom FFT off), theprogram would measure theabsolute delay correctly, sincethe maximum delay range inthis case would be ±500 ns.

The maximum measurabledelay range must be consideredwhen measuring the absolutedelay of a DUT if a normalizedmeasurement is done withsignificant reference delay. Thereference delay and the delay ofthe DUT must be less than theunambiguous range. Forexample, if the unambiguousrange is ±500 ns, and thereference sweep had 373 ns ofdelay, then the DUT delay mustbe less than 127 ns. Beyondthis, the delay value will benegative (until it wraps arounda second time).

The only way to increase theunambiguous measurementrange is to lower the modulationfrequency. However, thisreduces the aperture as well,

which increases measurementnoise. If the measured absolutedelay of your device changessignificantly with decreasingmodulation frequency(assuming normalization isdone), then the electrical lengthof the device is too long for avalid measurement at thatmodulation frequency. If youknow the approximate electricallength of your device and cantherefore infer how many phasewraps occurred, then an invalidabsolute group delaymeasurement can be corrected.If direct demodulation (zoomFFT off) was used, add N ×(1/fmod); if the zoom FFT wasused, add N × (.5/fmod), whereN is an integer which accountsfor the number of times themeasurement wrapped around.

Negative delay

Negative values of group delaycan occur when themeasurement wraps around asdescribed above, or when thereference path length is longerthan the measurement path.The program has a flag whichcontrols whether negative delaynumbers are corrected or not.The default is set to shownegative numbers (nocorrection). This allowsnormalization to work correctlywhen the reference path islonger than the measurementpath. For example, if thereference measurement had -4ns of absolute delay and theDUT measured 56 ns, then thenormalized measurement wouldcorrectly show 60 ns of delay (56ns - (–4 ns)).

When the group delay flatnessof the cables is not a significantcontribution to measurementerror, the normalization processshould be done with both theinput and output cablespresent. If cable contribution islikely to be a significant factorover one or both of thefrequency ranges used in themeasurement, then thenormalization can be done withonly one of the cables. The othercable can then be measuredseparately over its intendedfrequency range, using thedirect phase measurement(either with a vector networkanalyzer or the HP 71500A),and its delay contributionmanually removed from thenormalized test data. The groupdelay program currently doesnot incorporate a two-partnormalization such as this.Measuring the second cable inthis way is necessary for anabsolute measurement, but maynot be needed for relativemeasurements if itscontribution would not be asignificant factor in themeasurement.

Unambiguous delay range

The nature of the modulationtechnique for measuring groupdelay is such that absolutegroup delay values are validonly for 0±(.5/fmod), where fmodis the modulation frequencyused for the measurement. Thiscorresponds to 0±180° of phaseshift. After +180°, themeasurement will wrap aroundto –180°, which changes thesign of the group delay frompositive to negative. This isequivalent to subtracting1/fmod worth of delay.

11

Measurement Time vs. Trace Length

1 4 8 16 32 64 1280

3

6

9

12

15

18

21

24

27

Trace Length

min

Zoom=128

Zoom=64

Zoom=32

Zoom=Off

Zoom=16

Zoom=8

Figure 8: Measurement times versus trace length for various zoom factors.For N averages, the measurement times will increase by a factor of N.

To change the negative delayflag, the user must edit theprogram. The variable can befound in the Measure_groupsubroutine. With the USER keysactive, type:

EDIT Measure_group

Set Neg_delay_flag to 0 tocorrect negative delay values;set to 1 to display negativenumbers.

Measurement TimesThe time required to perform agroup delay measurementvaries considerably, dependingon the accuracy and resolutionneeded. Figure 8 shows typicalmeasurement times versustrace length for various zoomvalues. These measurementtimes will increase by a factorof N when N averages are used.

If the flag is set such thatnegative delay values arecorrected, then 1/fmod worth ofdelay is added when the directdemodulation mode is used, and.5/fmod worth of delay is addedwhen the FFT zoom mode isused. This makes any negativevalues positive. This operatingmode can correct ameasurement when one phasewrap has occurred, but canmake a normalizedmeasurement invalid.Following the previous example(using the zoom FFT and 1 MHzmodulation), the normalizedmeasurement would show -440ns, which is clearly not correct(56 ns - (-4 ns + 500 ns)).

Using thePersonalityObtaining the ProgramThe HP 71500A group delaypersonality is distributed free ofcharge on an “as is” basis,without a formal HP productnumber. It is neither a fullysupported nor specified product.The user is free to modify theIBASIC code to customize it fortheir particular measurementapplication. The program can beobtained by contacting yourlocal HP sales or applicationengineer.

Firmware UpgradesWhat if I have an olderinstrument?

This personality needs firmwarerevision 1.1 or greater. If youhave an HP 71500A or an HP70820A module with firmwarerevision 1.0 or 1.01, then afirmware upgrade kit must beordered (product number HP70863A). This upgrade sells for$2000, and adds several newfeatures besides the capabilityto run IBASIC programs.Additional features includeincreased trace memory forsingle-shot measurements, limitlines, and enhanced trace mathoperations.

Downloading theProgramThe group delay personalitymust be downloaded into thememory of the instrument eachtime the instrument is poweredup (program memory is volatile).There are two basic ways toaccomplish this. One method

12

UNIX® is a registered trademark of UNIX Systems Laboratories Inc. in the U.S.A. and other countries.

address and volume number.The following examples are foran HP 9122 or HP 9133 floppydisk drive set to HP-IB address 0:

GET“DOWNLOAD:CS80,700,1” orGET “DOWNLOAD:,700,1”

If you are not getting theprogram from a floppy disk,skip the “:,700,1” portion of thefilename, but include the fullpath name for the file. The MSIcommand could be used to setthe HP BASIC environment tothe correct directory.

After loading the program, pressthe RUN softkey or type RUN,followed by the return key. The program will prompt youfor the file name to download. Inkeeping with the above exampleusing a floppy disk, you wouldtype:

G_DEL_IB:,700,1

You should see the programlines displayed on the screen asthe program is beingdownloaded to the HP 71500A.This process will take a fewminutes to complete. At the endof the download process, theprogram will ask you if you wantto store the program to themainframe’s default drive. If youhave a formatted 128K RAMmemory card available, and thedefault drive is set to MEMORYCARD (select under the DISPLAYkeys: Mass Storage; msi),then enter “Y” to store theprogram. An alternate way tostore the program to a memorycard is described in the next

section, along with details ofhow to format a memory card.

If you are working in a UNIX®

environment, it is possible todownload the program directlyfrom the hard disk drive using ascript program, without havingto run HP BASIC or use afloppy disk. This method ismuch faster than using the HPBASIC download program.

Storing to a memory card

Once the group delay programhas been downloaded to theinstrument, it can then be storedto a memory card. A 128K RAMmemory card (HP 82215A)should be used.

Note: the memory card uses aCR2016 lithium battery. It isrecommended that the batterybe changed annually. To retainthe data stored on the card, thebattery should be replacedwhile the card is installed in themainframe with the power on.

The memory card must beformatted before it can be used.To format a new card, press theMENU key below the display,followed by mass storage(more 2 of 2) under theStates menu (page 2 of 2).Next, press more followed byFORMAT. You can select eitherDOS or LIF format. After theformatting is compete, press theUSER key to activate theIBASIC keys. To save a programon the card, simply type:

SAVE “G_DEL_IB”

involves using an externalcomputer, and the other usesthe mass storage capability ofthe HP 70004A colordisplay/mainframe. The lattermethod can load the programeither from a floppy disk or amemory card. Since theprogram is likely to be suppliedon a floppy disk, it is desirableto copy it to a memory card assoon as possible. Subsequentdownloadings can then be donefrom the memory card, whichalleviates the need to have anexternal computer or floppydisk drive.

Loading from an externalcomputer

The most common way ofdownloading the program withan external computer is to use afloppy disk. Alternately, theprogram can be downloadeddirectly from the computer’shard disk drive. This is useful ifthe program was obtained via e-mail or ftp. Several choices for computers exist, such as an HP 9000 series 300 running HP BASIC/WS, an HP 9000series 300/400/700 running HP-UX (with HP BASIC/UXrunning as an application), or apersonal computer (PC) runningHP BASIC/DOS. The computermust have an IEEE-488 (HP-IB)port to communicate with theHP 70004A mainframe.

From the HP BASIC operatingenvironment, load and run theASCII program DOWNLOAD.To load an ASCII file using HPBASIC, use the GET command.You may need to specify themass storage logical name,

13

Either an HP-HIL keyboard canbe used for entering thecommand (plugged into thefront of the HP 70004Adisplay/mainframe), or the“typing aids” key under theUtility menu of the USER keys.As the program is stored on thecard, the LED to the left of thecard will flash intermittently. If there is a problem with themass storage definition, refer tochapter 3 of the HP 71500AInstrument BASIC User’sGuide (70820-90055). After youhave successfully stored theprogram onto the memory card,slide the write-protect switch tothe outside position. This willprevent accidental erasure.

Loading from a memorycard or disk drive

The second method ofdownloading the program is touse the mass storage features ofthe HP 70004A mainframe.Press the DISPLAY key underthe display, followed by MassStorage on the left, followed bymsi on the right. SelectMEMORY CARD if the programhas already been stored to amemory card (see above), or HP-IB disk if loading from anexternal disk drive (connectedto the HP-IB connector on theback of the mainframe). Using an HP 9122 or HP 9133disk drive as before, make surethat the HP-IB ADDRESS is setto 0, the UNIT NUMBER to 1, andthe VOLUME NUMBER to 0.

The next step is to actually loadthe program. The easiest way is to use the MENU keys of the HP 71500A. Press the MENU keybelow the display, followed byStates (page 2 of 2) on theleft, followed by mass storage(more 2 of 2) on the right.

Next, use the up/down stepkeysto highlight the file to be loaded.Then press LOAD FILE on theright. After the file has loaded,it will automatically run.

It is possible to have the programautomatically load and runwhen you power up the analyzer,by having an “auto-start” filestored in addition to the groupdelay program. The HP 71500Awill look for the auto-start fileon whatever mass storagedevice was last used. If you areusing a memory card to storethe group delay program, usethe procedure describedpreviously to download and copythe program AUTOST from thefloppy disk to the memory card.See page 3-8 of the HP 71500AInstrument BASIC User’sGuide (70820-90055) for moreinformation on auto-start files.

Menu OverviewThe group delay personalityprovides the user with a custommenu structure, tailored to thegroup delay measurementapplication. These menus areaccessed under the USERsoftkeys. Note that when thegroup delay program is running,successive presses of the USERkey will toggle between thepersonality softkeys and theIBASIC softkeys. The personalitysoftkeys are structuredsimilarly to the standard MENUkeys in that the left-hand keysare fixed, while the right-handkeys change according to whichleft-hand key is pressed. Thepersonality has six left-handmenus to choose from:

Sweep: the start, stop, and stepfrequencies are set here, as wellas the number of averages touse during the measurement.The VIEW SPECTRM key allowsa preview of the measurementto confirm that everything is setup correctly. The MEASURE keystarts the actual measurement.A HELP key is available.

Config: this menu is used toenter the parameters whichdefine the hardware setup, suchas mixing numbers and LOfrequency, and the modulationtype and frequency.

Options: this menu allows youto choose whether to turn thefollowing functions on or off:single shot, FFT zoom, phaseand amplitude traces, andinvert phase mode. Also choosewhether markers use highest orlocal peak mode. A HELP key isalso available here.

Process: the normalizationsoftkeys are located here, aswell as an autoscale key. TheFIND LO key invokes a signalidentification routine, which isused to determine the exact LOfrequency of the DUT.Sources: set RF, LO, andmodulation source type andaddresses here.

Power: set the power levels ofthe various sources used. Alsoset the modulation depth (forAM) or modulation deviation(for FM) when using a sourcewith internal modulation.

14

Setting Up and Doing aMeasurementThe order of keystrokes forsetting up a group delaymeasurement is not crucial. TheHELP key under the Sweepmenu describes one possiblesequence:

1. Configure the sources (Sources menu): the personality allows for the HP-IB control for up to three sources (RF, LO, and modulation). At a minimum, you must have control of the RF source to perform a measurement. When enteringan HP-IB address, use the last two digits only (for example, enter 19 rather than 819). If you are using anHP 8360 series synthesizer with the internal modulation generator option (Opt. 002), the mod address should be the same as the RF address, and both RF src: and Mod src: should be set to HP836XX.

When using an HP 70340Asynthesizer, set RF src: toHP 836XX.

2. Set the power level of the sources (Power menu):usually, only the RF power is set. If a non-HP-IB modulationgenerator or the internal modulation option of the HP 8360 is used, you need not set the MOD POWER. If the internal modulation option is available, set the AM DEPTHor FM DEVIATion, as appropriate. For best results with either internal or external modulation, use an

RF frequency. Use the following guide for setting Mand N:

Upconverters:M = 1, N = 1(IF = RF + LO)

Downconverters:For LO > RF,M = -1, N = 1(IF = LO - RF)*For LO < RF,M = 1, N = -1(IF = RF - LO)

* frequency inversion occurs

If the DUT is a multiplier, setM to the correct harmonic number, and set N and LO to zero.

The softkey labeled RF is used for setting the synthesizer frequency after VIEW SPECTRM is pressed. The value for RF is normally the center of the passband of interest. During the actual measurement, this value is not used, as the synthesizer is programmed according to the start, stop, and step frequencies.

If the LO frequency is not known exactly, use the FIND LO routine under the Processmenu. This routine also programs the synthesizer to the value set by the RFsoftkey. After the LOfrequency is computed, it automatically updates the value for the LO softkey. This routine requires that M and N are set correctly.

AM depth of 90%, or an FM deviation of twice the modulation frequency (ß=2). This will maximize the amplitude of the modulation sidebands of interest.

Note: that when measuring amultiplier with FM, thedeviation must be corrected bythe multiplication factor. For example, if using 1 MHz ofFM to test a x3 multiplier, thedeviation should be set to 2 MHz/3 or 667 KHz.

3. Configure the system (Configmenu): the first step is to select the modulation type and frequency with the top two softkeys. Remember, the measurement aperture is equal to twice the modulationfrequency, and is independentof the frequency step.

Next, enter M, N and the LOfrequency which correspond to the hardware setup. The following equation is used to compute the IF frequency at the output of the DUT:

IF = M*RF + N*LO

The signs of M and N are determined from whether theDUT is an up or downconverter, and whether it uses high or low-side LO mixing. The IF frequency must always be positive, but it can sweep either up or down in frequency. This flexibility allows measurementof frequency-inverting devices.Frequency inversion only occurs when measuring a downconverter where the LO frequency is higher than the

15

The circuit ON/OFFsoftkey is used to display or disable the block diagram drawing. Although the drawing shows a connection of the 10 MHz reference between the sources and the HP 71500A, it is usually not necessary to do so.

4. Choose the options (Optionsmenu): the sglshot ON/OFFsoftkey determines whether the measurement is performedin the single-shot or repetitivemode of the HP 71500A. This product note describes the operation when the single-shot mode is used, where the modulation frequency must be less than one half of the single-shot sample rate, which is usually between 10 and 20 MHz. The repetitive mode can be used for higher modulation frequencies, but all of the frequencies involved(including the LO in the DUT)must be known much more accurately than in the single-shot case. In general, it is recommended that everythingbe locked together for the repetitive mode. The zoom FFT mode is not available when sglshot is set to OFF.

The zoom ON/OFF softkey determines whether the zoomFFT mode (zoom ON) or the direct demodulation mode (zoom OFF) is used. When using FFT zoom, the ZOOM FACTOR softkey is used to enter the amount of zoom desired (see discussion of Measuring in the Frequency Domain on page 6). This zoom factor is only used during the actual measurement, and not when VIEW SPECTRM is pressed.The markers softkey is used

to select between the two marker modes, LOCAL or HIGHEST (see discussion of Marker options on page 8).

The phase ON/OFF softkey determines whether a phase trace is displayed below the group delay trace. When using the zoom FFT mode, anamplitude trace is also displayed when phase is set to ON.

When set to ON, the invert ON/OFF softkey adds 180°of phase shift to each measurement. This is to account for some modulators that perform an inversion of the modulation signal. This has the effect of adding .5/fmod worth of delayto all of the group delay values.

5. Set the sweep parameters (Sweep menu): enter the start, stop and step frequencies of the RF source with the top three softkeys. If the source is to be swept down in frequency, make the start frequency higher than the stop frequency, and enter a negative value for the frequency step.

Ideally, the frequency step should be equal to twice the modulation frequency for uniform sampling across the measurement band. This provides full group delay characterization with the minimum number of data points. Setting the frequency step smaller than this fills in the trace with more points, but doesn’t provide any additional group delay resolution (over-sampled case). If the frequency step is larger than twice the

modulation frequency, a warning message is displayedon the screen to alert the user that group delay variations may be missed (under-sampled case). Care should be taken to make the measurement aperture and frequency step small enough to resolve all of the group delay variations which occur across the measurement band for the particular DUT being tested.

The AVERAGE softkey is usedto enter the number ofaverages used during the measurement.

Before doing an actual measurement, it is a good idea to confirm that everything is set up correctly.One way to do this is to pressthe VIEW SPECTRM softkey. An FFT of the modulated carrier is displayed, centered at the DUT output frequency.If no signal appears, check that everything is correctly set in the Config menu. Thisspectrum display is very useful for setting the right amount of modulation when a manual modulation generator is used.

If the frequency span of the spectrum is too wide to clearly see the modulation sidebands, then the zoom transform can be used from the standard HP 71500A menus to expand the display (increase the frequency resolution). To access the zoom transform, press MENU, then Traces, then trnsfrm control. Set zoom to ONafter setting the correct CENTER and SPAN values.

16

6. Start the measurement (Sweep menu): to invoke a measurement, press the MEASURE key. At any time during the acquisition of data, the ABORT key can be pressed to stop the measurement. It may take a few seconds before anything happens because the abort is processed after the current trace acquisition is completed.If the instrument appears to be frozen after pressing ABORT, press INSTR PRESETfollowed by the USER key to return to the personality.

If the top trace does not showa signal during ameasurement, check that the stop, start, and step frequencies, and N and M are correctly set. As the program steps along, the IF frequency and delta phase of the point being measured is shown on the display.

7. Normalize and scale the results(Process menu): if the measured trace is to be used as a reference for normalization, then press theNORMALZ softkey (see the previous discussion about normalization on page 9). The normal ON/OFF key is used to turn normalization onor off. The trace definitions will change according to the status of this softkey (e.g., MEM1 for an unnormalized trace, and MEM1-MEM3 for a normalized trace).

The AUTO-SCALE softkey is used to automatically scale the traces to fit the screen. If a phase trace is displayed, then the electrical length is automatically subtracted to

show deviation from linear phase. This auto-delay routine uses the entire phase trace to do a least-squares fit to a straight line. If it is desired to flatten out a smaller portion of the phase trace, such as the center of the passband of a filter, then the auto-delay routine must be done in conjunction with markers. The markers shouldbe placed on trace 2 to define the frequency band of interest.The auto-delay routine can then be invoked again by pressing AUTO-SCALE. This routine can also be done via the MENU keys by pressing AUTO DELAY under the Scale menu (more 2 of 2). All of the traces can be manually scaled or offset if desired by using the appropriate keys under the Scale menu (more 1 of 2).

Printing/Plotting theResultsAn HP-IB printer or plotter isneeded to directly obtain ahardcopy of the traces. It shouldbe connected to the HP-IBconnector of the HP 70004Adisplay/mainframe, and NOT tothe HP-IB connector on the HP 70820A module. Make surethat the HP-IB switch on the rearpanel of the display/mainframeis set to the on (up) position.Non-HP-IB printers such asLaserJets and DeskJets can beused with an HP-IB toCentronics converter, such asthe HP 92203J/K.

The PRINT or PLOT keys on thedisplay/mainframe are used todump the traces to the printeror plotter. Hardcopy parameterssuch as printer type, and theaddress of the print or plotdevice are set under the Hard Copy menu of theDISPLAY keys. When using anHP-IB to Centronics converter,select either the THINK JET orthe PAINTJT BLACK printerdriver. Do not press any keys onthe display until the printout iscomplete — otherwise, theprintout will be aborted.

If the program was loaded intothe memory of the HP 71500Avia a memory card or externaldisk, and no computer isconnected to thedisplay/mainframe HP-IBconnector, then the PRINT andPLOT keys on thedisplay/mainframe workproperly. If an externalcomputer was used to downloadthe program to the HP 71500Aand it is still connected to theHP-IB bus, then it either mustbe physically removed from thebus, or it must relinquish buscontrol to allow thedisplay/mainframe to controlthe printer or plotter. To removethe computer as a buscontroller, type the followingBASIC commands in thecomputer:

LOCAL 7SEND 7; UNT UNL DATA

Now press the appropriate keyon the display/mainframe toinitiate a print or plot.

Group Delay vs. Zoom Size

RF=1 GHz RF=20 GHz

Zoom = off +/- 1 ns +/- 3 ns

Zoom = 8 +/- 300 ps +/- 1.2 ns

Zoom = 16 +/- 275 ps +/- 1.1 ns

Zoom = 32 +/- 250 ps +/- 1.0 ns

Zoom = 64 +/- 200 ps +/- 750 ps

Zoom = 128 +/- 150 ps +/- 500 ps

Table 1: Typical group delayaccuracy versus zoom size with noaveraging. Excludes mismatch error.

Phase and Amplitude vs.Zoom Size

RF = 1 GHz

Phase Ampl.

Zoom = off +/- 1° –

Zoom = 8 +/- .5° +/- .1 dB

Zoom = 16 +/- .4° +/- .1 dB

Zoom = 32 +/- .3° +/- .05 dB

Zoom = 64 +/- .2° +/- .05 dB

Zoom = 128 +/- .2° +/- .05 dB

Table 2: Typical phase andamplitude accuracy versus zoomsize with no averaging, at 1 GHz. Excludes mismatch error.

Phase and Amplitude vs.Zoom Size

RF = 20 GHz

Phase Ampl.

Zoom = off +/- 5° –

Zoom = 8 +/- 2° +/- .2 dB

Zoom = 16 +/- 1° +/- .15 dB

Zoom = 32 +/- .8° +/- .15 dB

Zoom = 64 +/- .7° +/- .1 dB

Zoom = 128 +/- .5° +/- .1 dB

Table 3: Typical phase andamplitude accuracy versus zoomsize with no averaging, at 20 GHz. Excludes mismatch error.

17

AccuracyConsiderationsTypical AccuracyTables 1, 2, and 3 show typicalmeasurementrepeatability/accuracy versuszoom size at 1 and 20 GHz.These numbers do not take intoaccount mismatch errors, orfrequency normalization error,so they represent best caseconditions (mismatch isdiscussed in more detail on page18). Each data point in the tablewas derived from two non-frequency translatingmeasurements of a cable, eachwith a 100 MHz span centeredat the carrier frequency. Thefirst measurement was used asthe reference trace for anormalized measurement. The measurement was thenrepeated with normalization on,to determine how much tracenoise was present. These valuesare typical of what would beachieved if normalizing to theDUT output frequency, whichoptimizes for relative groupdelay measurements. The RFcarrier power was set to 0 dBm(the maximum allowable input),and no averaging was used.

Noise ReductionAveraging

The results in Tables 1, 2 and 3can be improved by averaging,at the expense of measurementtime. N averages increase theaccuracy and measurementspeed by a factor of N. Theaccuracy improvementexpressed in dB is 10*log(N). A 3 dB improvement requiresdoubling the number of averages.

Smoothing

Smoothing may also be appliedto the trace to reduce theexcursions caused by noise. Thesmoothing function isimplemented as a variable-length median smoothing filter.The number of trace points usedto compute the median is set bythe smoothing value. Smoothingtends to work best when thenumber of trace points is large.

The smoothing function isaccessed via the MENU keys,under the Traces menu. Pressavg,hld, followed by smoothON|OFF until ON is underlined.Enter the smoothing valuedesired. The results of thesmoothing can be compared tothe original trace by togglingbetween on and off. Whilesmoothing can be a useful toolto reduce displayed trace noise,too much smoothing affects thedesired data, by reducing non-noise variations caused bythe DUT itself.

18

Signal-to-Noise RatioThe accuracy of the measurementis directly proportional to thesignal-to-noise ratio (SNR) ofboth signals. For bestmeasurement results, thesignals at both inputs to the HP 71500A should be as close to(but not above) 0 dBm aspossible. Tables 4, 5, and 6 showthe accuracy degradation due to sub-optimum RF input signallevels.

When using the FFT zoommode, choosing the modulationindex to maximize themeasured sideband levels isalso important (see step 2 onpage 14).

When measuring a DUT with afrequency responsecharacteristic of a filter, thegroup delay accuracy will degradewhen measuring the filter skirt,due to a smaller output signallevel.

The bandwidth of the DUT willalso affect the measurement.The narrower the bandwidth,the more accurate themeasurement will be due to lessnoise. Similarly, a DUT withexcess noise will degrade themeasurement accuracy. DUTnoise may be in the form of LOphase noise or broadbandamplified noise.

Mismatch ErrorImpedance mismatch in thesystem between cables, theDUT, and test equipment causesgroup delay, phase andamplitude ripples versusfrequency. Error caused bymismatch can be difficult toeliminate since normalizationonly removes mismatch errorpresent without the DUT inplace. If the DUT has poor portmatches, the mismatch errorintroduced when the DUT isadded will be measured as partof the DUT’s response.Judicious use of attenuatorsand isolators will improve portmatches, insuring that themeasurement contributions ofthe test setup are minimized.

Mismatch error can occurwithin the DUT as well. Forthis reason, a DUT consisting ofa filter and a mixer may havesignificantly worse group delayperformance as compared to thefilter alone, due to theinteractions between the filter,mixer and internal connections.

LO StabilityThis measurement applicationassumes that both the internalLO and the RF source used asthe stimulus are of high quality(i.e., typical synthesizerstability and phase noise).While the LO needs to be stable,its exact frequency need not beknown initially. The Find LOroutine can be used to determinethe LO frequency with sufficientaccuracy to perform a groupdelay measurement.

Group Delay vs. RF InputPower

RF=1 GHz RF=20 GHz

0 dBm +/- 275 ps +/- 1.1 ns

-10 dBm +/- 300 ps +/- 1.5 ns

-20 dBm +/- 350 ps +/- 2.0 ns

-30 dBm +/- 500 ps +/- 2.5 ns

-40 dBm +/- 1.2 ns +/- 3 ns

-50 dBm +/- 4 ns +/- 7 ns

Table 4: Typical group delayaccuracy versus RF input powerwith no averaging. Zoom size = 16. Excludes mismatch error.

Phase and Amplitude vs. RF Power

RF = 1 GHz

Phase Ampl.

0 dBm +/- .4° +/- .1 dB

-10 dBm +/- .5° +/- .1 dB

-20 dBm +/- .7° +/- .1 dB

-30 dBm +/- 1.0° +/- .1 dB

-40 dBm +/- 1.5° +/- .1 dB

-50 dBm +/- 10° +/- .25 dB

Table 5: Typical phase andamplitude accuracy versus RF inputpower with no averaging. Zoom size = 16. Excludes mismatch error.

Phase and Amplitude vs. RF Power

RF = 20 GHz

Phase Ampl.

0 dBm +/- 1.0° +/- .15 dB

-10 dBm +/- 1.5° +/- .2 dB

-20 dBm +/- 2.0° +/- .2 dB

-30 dBm +/- 2.5° +/- .2 dB

-40 dBm +/- 4° +/- .3 dB

-50 dBm +/- 10° +/- .5 dB

Table 6: Typical phase andamplitude accuracy versus RF inputpower with no averaging. Zoom size = 16. Excludes mismatch error.

19

In addition to the abovemanuals, the command previewand key-logging capability ofIBASIC allow you to see on thescreen or insert in the programthe remote commandequivalents of front-panel keypresses. This greatly reducesthe need for the instrumentprogramming manual. Thesefeatures are accessed under theIBASIC Miscel menu,followed by key log. SelectPREVIEW for screen display, orON for inserting lines into theprogram. Press OFF toterminate the key-logging mode.

Changing DefaultsThe program can be modified sothat it comes up with the all ofthe variables preset to yourmeasurement needs, with theexception of the source HP-IBaddresses and types. To changethe defaults, make sure theprogram is not running and theUSER keys are selected, andtype:

edit “default_values”

After you have changed theappropriate values, be sure tore-save the program onto amemory card or floppy disk. Ifyou get an error indicating thatthe card or disk does not haveenough room to hold theprogram, it may be necessary todelete the old file first, and thensave the current program withthe same name as before.

Modifying theProgramThe group delay personality forthe HP 71500A may be modifiedas necessary to tailor it to yourparticular measurement needs.

ManualsThe following manuals will behelpful for modifying the code:HP 71500A IBASIC User’s Guide:70820-90055

HP 71500A ProgrammingManual:70820-90053

HP IBASIC User’s Handbook:E2083-90005

HP BASIC CondensedReference:98616-90005

All of the commands in thecondensed reference which arecompatible with IBASIC areindicated by IN. An asterisk *following the IN indicates thatsome restrictions or extensionsmay apply.

When using the FFT zoommode, the highest peak functionof the markers can be used totrack an LO that is driftingslightly during the measurement.Larger zoom factors result inless tolerance to LO drift duringthe measurement, since thefrequency span becomessmaller. If the sidebands driftout of the displayed frequencyspan, then obviously themeasurement will be invalid.The direct demodulation modehas the most tolerance to LOdrift and residual-FM. Even ifthe LO drift is small enough toallow the measurement to bedone, it can cause degradationof the group delay accuracy.

Adding SourcesTo add support for a differentsynthesizer, the followingroutines in the program mustbe modified:

Easy_readReset_source

Set_freq

Set_mod_freq

Set_power

Set_output

Set_modulator

Set_mod_pct

Set_mod_dev

Check_assign

Source_menu

Where needed, add theappropriate output commandsthat are consistent with theprogramming syntax for thenew synthesizer.

20

Appendix A

Comparison to OtherInstrumentsThis discussion is meant toclarify where the HP 71500Afits in compared to other groupdelay measurement solutions.

Vector Network Analyzers

The most common tool formeasuring group delay of lineardevices is the vector networkanalyzer (VNA). VNAs utilizethe direct phase measurementtechnique. They are very fastand accurate. A VNA by itselfcan only measure non-frequencytranslating devices. Forfrequency translatingcomponents, an external mixeror extra DUT must be used sothat the frequency at thereference and test channels isthe same. However, in order tomake a group delaymeasurement, the LO for theexternal mixer must be phaselocked to the LO used withinthe DUT. This can beaccomplished by either providingor having access to the internalLO (or its frequency reference).This prevents VNAs frommeasuring components wherethe LO is embedded with noexternal access.

The VNA cannot directlymeasure absolute group delayeither. Absolute delay must beinferred from severalmeasurements involvingswapping of mixers, or anassumption that the mixercontributes little delay. The HP71500A can measure absolutedelay directly, with a relativelysimple normalization routine.

Trade-offs

Although the HP 71500A canmake these absolute andrelative group delaymeasurements, there is a trade-off in performance compared tousing a vector network analyzer— reduced speed and accuracy.Measurement speed is slowerfor two reasons. The first is thata stepped CW approach isrequired, which is slower thanusing a continuous sweep. Thesecond reason is that the noisereduction techniques needed forhigh accuracy measurementstake a fair amount of time (FFTzoom, trace averaging, or both).An example of speed for atypical measurement using FFTzoom and no averaging, isabout 2 minutes for a 32 pointsweep (using a zoom factor of16). For absolute delaymeasurements, there may be anoverall measurement speedimprovement because of asimpler calibration routine —no external mixer calibrationsare needed.

The second disadvantage is thatmeasurement accuracy is likelyto be less than what can beachieved using a VNA due to lackof full vector error correction,and linearity limitations withinthe HP 71500A. An HP 8510Cnetwork analyzer can achieverelative group delay accuraciesbelow 100 ps for linear devices,or for measurements relative to a golden standard.

Scalar Network Analyzers

Some scalar network analyzers(SNAs) can be configured tomeasure AM group delay,providing fast sweeps andmedium accuracy for relativelylow cost. Their primary target isfor high volume manufacturingtest, where throughput is moreimportant than accuracy. SNAsuse the AM envelope technique,using broadband diode detectorsto strip the AM from the RFcarrier. As such, they canmeasure both linear andfrequency translating deviceswith internal LOs. The lack ofcapability to provide FM groupdelay precludes their use forsome applications where AMcan’t be used. Although slowerand more expensive, the MTA’smain advantages are providingFM group delay and betteraccuracy.

Microwave Link Analyzers

The microwave link analyzer(MLA) provides fast andaccurate FM group delaymeasurements, and also makesother important measurementson analog microwave radio links.For loop-back testing (testing alink over a short distance), oneMLA is used. For end-to-endlink measurements, two MLAsare used. MLAs are used directlyonly for IF-IF measurements(non-frequency translatinglinks). For measuring frequencytranslating links, (RF-IF or IF-RF measurements), the system must provide thenecessary converters and accessto IF ports. For measuringsubassemblies like frequencyconverters, extra componentssuch as mixers, filters, and LOswould be needed, thereby greatly

21

increasing the system cost andcomplexity. The HP 71500Asolution is much simpler,because it directly provides RF-IF or IF-RF measurementswithout additional components.

HP 11770A LinkMeasurement Personality

This personality provides theHP 8590 E-series spectrumanalyzers with group delaymeasurement capability verysimilar to an MLA. This solutioncan measure devices from 300KHz to 2.9 GHz; it is not limitedto 70/140 MHz IFs as with a conventional MLA. Using twospectrum analyzers with this personality, RF frequencytranslating devices can bemeasured relatively inexpensively.The link analyzer personalityhas one potentially seriouslimitation — it cannot measuredevices that perform afrequency inversion (i.e., the IFsweeps downward in frequencyas the RF is swept upwards).This can occur when measuringdownconverters such as directbroadcast satellite (DBS) tunersthat use high-side LO mixing(LO > RF). The MTA does nothave this limitation as it istuned independently from the source.

Summary

For microwave measurementsof frequency converters withinternal LOs, the HP 71500A isthe highest performancesolution, especially when FMgroup delay is required. For RFmeasurements, the choices aregreater. For many applications,lower cost solutions will suffice,such as an SNA, or two HP 8590E-series spectrum analyzerswith the link analyzerpersonality. The HP 71500Asolution can fill certain nichessuch as when FM group delay ofa frequency inverting device isneeded, or where the instrumentcan provide additionalmeasurements such as timedomain pulsed-RF profiling, ordynamic AM-to-PM conversionmeasurements.

For more information, callyour local HP sales office listed in your telephone directory or an HP regional office listed below for the location of your nearest sales office.

United States of America:1 800 452 4844

Canada:(416) 206 4725

Japan:(81) 426 48 0722

Latin America:Miami, Florida(305) 267 4245/4220

Australia/New Zealand:(008) 13 1347Melbourne Caller272 2555

Asia Pacific: Hong Kong(852) 599 7070

Korea(2) 769 0800

Taiwan(2) 717 9524

Singapore(65) 291 8554

India(11) 690 355

PRC(1) 505-3888

In Europe, Africa, and Middle Eastplease call your local HP salesoffice or representative:

Austria/East Central Europe:(1) 25000-0

Belgium and Luxembourg:(02) 778 31 11

Denmark:45 99 10 00

Finland:(90) 88 721

France:(1) 69.82.65.00

Germany:(06172) 16 0

Greece:(01) 68 96 411

Ireland:(01) 284 4633

Israel:(03) 5380 333

Italy:(02) 92 122 241

Netherlands:(020) 547 6669

Norway:(22) 73 56 00

Portugal:(11) 301 73 30

South Africa:(011) 806 1000

Spain:900 123 123

Sweden:(08) 750 20 00

Switzerland:(01) 735 7111

Turkey:(312) 425 83 13

United Kingdom:(0344) 366 666

For countries not listed, contact Hewlett-Packard,International Sales Europe, Geneva, SwitzerlandTel: +41-22-780-4111Fax: +41-22-780-4770

Data Subject to ChangeCopyright © 1994Hewlett-Packard CompanyPrinted in U.S.A. 4/945091-8634E