1959/60, No. 7 201

A TRANSISTORIZED RADIATION MONITOR

by M. van TOL and F. BREGMAN.

This article adds the first measuring instrument to the long list of transistor-equipped devicesthat have been described in this Review in recent years. The emphasis here lies on the specialrequirements the circuit of a measuring instrument has to satisfy.

Radiation monitors are simple portable instru-ments for the detection and measurement of ioniz-ing radiations. They are used in laboratories andhospitals whose staffs work with radioactive sub-stances or with apparatus generating ionizing radia-tions (X-ray equipment, for example). Anotherfield of application is the tracing of radioactive ores.These simple radiation meters are also suitable forspecial purposes such as in portable instruments formeasuring the level of liquefied gases in cylinders 1).Batteries are the obvious form of power supply

for portable radiation monitors, and since it isimportant that their dimensions and weight hesmall, the designer will he concerned to keep currentconsumption as small as possible. A radiation moni-tor which contained only one electronic tube apartfrom its Geiger-Müller tube, and which in conse-quence was very economical in operation, was de-scribed in this Review in 19532). It neverthelessrequired three batteries - a 40 volt anode battery,a 1.5 volt heater battery and a battery for givingthe control grid a negative bias of 14. volts. Thelatter was rendered necessary by the special circuitdeveloped to operate with only one tube.Small portable electronic devices represent an

extremely worthwhile field of application for tran-sistors. In such devices the fullest advantage istaken of the characteristic properties of transistors,namely the small bulk, the low power consumption(and hence low heat dissipation) that results fromthe lack of heater currents, and the modest voltagerequirements that can he met by small batteries,i.e. batteries comprising few cells. A prototype two-transistor radiation monitor, supplied from a singlethree-volt battery, was' in fact developed in theEindhoven Research Laboratories as early as 1954.Transistors have less convenient aspects too,

these being the spread in characteristics displayedby different individuals of the same type and the

1) See for example J. J. Arlman and H. N. L. Hoevenaar,Een niveaudetector voor de praktijk, De Ingenieur 71,Ch. 8 - Ch. 9, 1959 (No. 19) (in Dutch).

2) G. Hepp, A battery-operated Geiger-Müller counter,Philips tech. Rev. 14, 369-376, 1952/53.

621.317.794:621.375.4

sensitiveness of their characteristics to temperaturechanges. Forming as they do an obstacle to preci-sion, these drawbacks are ordinarily more seriousin a measuring instrument than in an audio-ampli-fier, for example, where no great harm is done if theamplification changes slightly. However, a radia-tion monitor is a measuring instrument in which thedemerits of transistors are not felt so keenly. Forone thing, the main requirements are reliabilityand sensitivity rather than high accuracy; what ismore, the measurement in question is based onpulse-counting, and for this purpose one can employcircuits in which the transistor characteristics havelittle or no influence on the actual reading. It isunderstandable that reliability should be given firstplace, for the purpose of the monitor is to measureradiation that is hazardous to human beings. Greatsensitivity is called for because in certain cases itmay have to detect radiation levels well below thehuman tolerance level. These requirements cancertainly be satisfied just as well with transistorsas with tubes.Two commercial versions, types PW 4014 and

PW 4012, have finally been evolved from theprototype. Equipped with four and five transistorsrespectively, the commercial models (illustrated infig. 1) do not differ in any essential respect from theprototype, merely being improvements thereon.

We shall now discuss the circuit of the prototype,given in fig. 2. It can he divided into the followingparts:(a) H.T. supply circuit.(b) Geiger-Müller tube supplied from this circuit,

and delivering electrical pulses when exposedto ionizing radiation.

(c) Pulse-shaper that standardizes the height ofthe pulses.

(d) Diode pump-circuit which, supplied with arandom sequence of pulses of uniform height,transforms them into a direct current whosevalue is proportional to the average number ofpulses per second. This current is measured withthe aid of meter M.

The H.T. supply is generated by" a simple oscil-

202 PI-IILIPS TECHNICAL REVIEW VOLUME 21

a b

Fig. 1. Two commercial versions of a transistorized monitor for hard y and f3 radiation whichhave been evolved from the prototype developed in the Philips Research Laboratories in1954 and discnssed in the present article.a) Type PW 4014, pocket radiation monitor.b) Type PW 4012, pistol model.

lator operating at a frequency of about 10 kcjs,using a transistor Tl' type oe 71. The voltage deliv-ered by the oscillator is stepped up in the trans-former and applied to a cascade circuit of threestages, which converts the alternating voltageinto a direct voltage of treble its amplitude. Thecascade circuit is equipped with selenium diodes.A direct voltage of the same value could be ob-

tained from a cascade circuit having some othernumber of stages and a transformer with a differentoutput voltage. Losses in the cascade circuit in-crease with the number of stages; those in the trans-former increase with the transformation ratio. Theoverall loss curve exhibits a shallow minimum cor-responding to the use of three stages, as in the cir-

I

'0, III

b c da

Fig. 2. Basic circnit of the prototype two-transistor radiationmonitor. a: high-tension supply circuit for Geiger-Müller tube.b: G.M. tube with series resistor. c: pulse-shaper. d: diode pump-circuit with milliammeter M. The circuits of the commercialversions appearing in fig. 1 do not differ in essentials from thatshown here.

1058

cuit under discussion; however, the particular trans-formation ratio and number of cascade stagesdecided upon make hardly any differences to thebnlk and weight of the circuit.

The G.M. tube is of type 18 503 or 18 504; inboth types the quenching gas is a halogen. Onaccount of their robustness and almost unlimitedlife, these types of tube are particularly suitable fornse in an instrument whose prime requirement isdependability. Moreover, their characteristics ex-hibit a long "plateau" with a particularly gentleslope (jig. 3). The high tension across the tube cantherefore vary within wide limits (370 to 650 V)without greatly affecting the rate at which thetube delivers pulses when exposed to radiationof constant intensity 3). For this reason it is possibleto use a high-tension supply circuit of such simpledesign; for the same reason, there is no need to payovermuch attention to the characteristics of tran-sistor Tl' Provided the oscillator oscillates - andthat is not asking a great deal - a usable high-tension supply will be available.

The voltage supplied is about 500 V, and theaverage current taken by' the G.M. tube does notexceed 30 fJ.A. The maximum power required istherefore 15 mW. A three-volt battery supplies theoscillator (and the other circuits). The high-tensioncircuit has an efficiency of about 60% and thereforetakes less than 0.01 A from the battery - a verylow current for a rod battery.

3) K. van Duuren, A. J. M. Jaspers and J. Hermsen, G. M.counters, Nucleonics 17, No. 6, 86-94, 1959.

1959/60, No. 7 TRANSISTORIZED RADIATION MONITOR 203

, As already stated, the value of the high-tensionvoltage has little effect on the rate at which the G.M.tube delivers pulses. It does, however, have an effecton the height of the pulses. Since there is no stabili-zation of the H.T. supply, the average currentthrough the G.M. tube cannot be used directlyas a measure of the count rate. For this reason thepulses are fed to the pulse-shaping circuit c (see

, :fig.2), which consists of a type OC 71 transistor T2

and a transformer Tr. Circuit c is a blocking oscil-lator 4). T2 acts as a switch that closes each time theG.M. tube delivers a pulse. When this happens, theprimary winding of Tr carries the full batteryvoltage (3 V). The secondary winding of Tr thushas induced in it a voltage pulse whose amplitudeis related to the battery voltage by the ratio of thenumbers of turns. The time that T2 remains in theconductive state, and hence also the width of thevoltage pulse in the secondary winding of Tr, depend

N

t

500

BJOO 600 700V

-V 99253

Fig. 3. Count-rate characteristic of a type 18 504 (halogen-quenched) Geiger-Müller tube. N, the number of pulses thetube delivers per second when exposed to a constant radiationintensity, is plotted as a function of the voltage V across thetube. The characteristic has a particularly low plateau slope(the "plateau" is the part of the curve between A and B).

amongst other things on the characteristics of thetransistor. These also have an influence on the slopeof the pulse edges, but the height of the pulsesdepends on the battery voltage alone.A diode pump-circuit d measures the rate at

which the voltage pulses, now of standard height,are produced. The circuit owes its name to the anal-ogy with a' piston force pump, in particular as usedfor compressing a gas. Diodes Dl and D2 representthe suction and delivery valves respectively; stor-age capacitor Cl may be compared with the barrelof the pump, and the much larger buffer capacitorC2 with the tank into which gas (i.e. electrons) isbeing pumped. However, charge continually drainsaway from the latter via resistor R and ammeter M.

4) For the functioning of a blocking oscillator see, for example,B. Chance et al., Waveforms, No. 19 of the RadiationLaboratory Series, McGraw-Hill, New York 1949.

A characteristic property of the circuit is that thecurrent through M is dependent on the number andthe height of the pulses entering it but not - pro-vided they are sufficiently wide, allowing Cl to befully charged - on their duration or the slope oftheir edges. (Compare with the piston force pump,which displaces a quantity of gas that is dependenton the number of strokes per sec and on the pres-sure of the gas entering the pump, but not on themanner in which the barrel fills up, provided onlythat it fills with gas up to the full inlet pressure.)The diode pump-circuit is therefore very suitable forhandling the pulses of uniform height that are de-livered by the pulse-shaper c. It is possible, by à. cor-rect choice of values for Cl' C2 and R, to make thecurrent through M proportional, or nearly so, to thepulse rate. A recent article in this Review 5) may bereferred to for a more detailed discussion of the wayin which a diode pump-circuit functions.We have seen above that the constituent parts

of the monitor have been adapted to one anotherwith the object of getting a meter reading that isindependent of transistor characteristics. We mayrecapitulate as follows.The H.T. voltage yielded by circuit a (fig. 2)

depends on Tl' It is applied to G.M. tube b, whichdelivers pulses at a rate which is independent of .the voltage applied and hence of Tl' The height ofthe pulses is dependent on Tl' Pulse-shaper c givesthese pulses a constant height. Their shape (i.e.their duration and the slope of their edges) depends,however, on the characteristics of T2; the influenceof Tl has been eliminated. Diode pump-circuit dconverts these pulses of standard height into acurrent that is independent of their shape. Thusthe influence of T2 has also been eliminated.Even so, the current through the measuring in-

strument is still dependent on the battery voltage,for the height of the pulses delivered by circuit c isproportional to that voltage (see above). Facilitiesfor checking the battery voltage have therefore beenprovided in the commercial versions in fig. 1, theindication being given on the meter (M).The ranges measured by pocket radiation monitor

PW 4014 are 0 to 3 mrjh (milliröntgen per hour)and 0 to 30 mrjh 6). The pistol model, PW 4012,has the ranges 0 to 1 mrjh, 0 to 10 mrjh and 0 to100 mrjh. The design of both models allows for the

5) J. J. van Zolingen, Philips tech. Rev. 21, 134-144, 1959/60(No. 4/5), in particular p. 138 et seq.

6) Röntgen, röntgen per hour and other units relating toionizing radiation were discussed in N. Warmoltz andP. P. M. Schampers, A pocket dosemeter with built-incharger, for X-radiation and gamma radiation, Philipstech. Rev. 16, 134-139, 1954{55.

204 PHILlPS TECHNICAL REVIEW ,VOLUME 21

connection of a type 18 509 counter tube, whichextends the PW 4014 ranges to 0-60 and 0-600 mr/h, 'and the PW 4,012 ranges to 0-20 and 0-200 mr/h.To give some idea of the significanee of these figures,we would add that 5 röntgens per year is regardedas an acceptable dose for persons exposed to radia-tion by reason oftheir occupation. In a working yearof 2000 hours this amounts to 2.5 mr/h. The toler-ance dose for other classes of the population islower.

A circuit giving a new kind of meter scale

A new circuit has meanwhile been developedin the laboratory in which the scale of the meteris progressively compressed (though not logarith-mically) at higher pulse rates. This means that, asfor a logarithmic scale, more economical use is madeof the available scale length than when a linear scaleis used. The deflections obtained with this circuitdiffer from those of a logarithmic scale in that zerodeflection corresponds to absence of radiation andfull-scale deflection to an infinitely high radiationlevel. Consequently the needle never goes beyondthe extremes of the scale - a very convenient fea-ture in a radiation monitor: the user always hasimmediate indication of the radiation level withoutrange-switching.

The principle of the new circuit is as follows.Fig. 4 shows a flip-flop circuit with two transistorsT3 and T4• Negative pulses from the G.M. tube arriveat terminal A at an average rate of N per second.After such a negative pulse on A the circuit is leftin the stable state in which Ta conducts; suppose

1'1pulsftS/sf>C(average)

11111 11111N:, pulsl>S/ SfJC

I I I I I

Fig. 4. Flip-flop circuit in which terminal A goes to theG.M. tube, and terminal B to a pulse generator that deliverspulses at a constant rate of No per sec. The mean currentthrough transistor Ta is a measure for N, the average numberof pulses delivered by the G.M. tube per sec.

that the current i3 then flowing through T3 is la. Apulse generator supplying negative pulses of con-stant repetition frequency No is connected to ter-minal B. After each of these pulses the circuit is

left in the state in which T3 is non-conducting. Themean current through Ta is now a measure for N;this can be seen with the help of figs. Sa and b.In each diagram the pulses arriving at A, those

____ .- -r __~~--~--~ANPu7ses/sllC

, (average),,18I:' Nopu7ses/sllC

tIi ,, ,

j, ,

I;tlrrn

""

nPII

H-t

--rrTrr".-~~~~~~~~--A!! III II 11I ! Ill! I I I ! I I Npu7ses/sec: : ' ' I I (average)I I : I II' , I I I

r-~17'~~r.-,-~,~~,_~i~~~~~'~_8.1 '11 jJ I' I "1, I I:13 i:: :: :: I " IJ" I I ' , Nopulses/sec

tj_'" " " "'HH1'nl': I : II3nHNh ULJ_ H dd Hh

1483b -t

Fig. 5. Diagram to explain the functioning of the circuit infig. 4.

arriving at B and the current through Ta are plottedone below the other as functions of time; the currentia through Ta can only have one of two values, zeroor la. When Ta has been rendered conducting by apulse from the G.M. tube, a current la flows untilstopped by the next pulse of the regular train fromthe generator. It will be seen from fig. 5a that if Nis much less than No, as in that diagram, then themean current through Ta will have a low value.Indeed, if the G.M. tube stops delivering pulsesaltogether, the mean current will be zero. Fig. 5bshows the situation when N greatly exceeds No.Here Ta is conducting almost all the time. At veryhigh radiation levels the mean current through thetransistor has a value approaching la.

The two extreme cases N ~ No and N~ Nolend themselves to a simple quantitative treatment,as follows.

For the case N ~ No, almost everyone of the Npulses delivered by the G.M. tube per sec willreach Ta when it is in the cut-off state. At anyinstant the average time that will elapse before thenext pulse arrives from the generator is HI/No} sec.Roughly, then, Ta conducts for a proportion of thetime given by tNINo, and the mean current throughit is i(N/No}Ia. The lower end of the scale thereforehas a linear character.

The case N~ No can be dealt with analogously.Of the No pulses delivered by the generator eachsecond, almost everyone reaches Ta when it is in

--~------_.~-~--~ - -.

1959/60. No. 7 TRANSISTORIZED RADIATION MONITOR 205

the conducting state. At any instant the averagetime required for the next pulse to arrive from theG.M. tube is I/N sec. (The factor t, appropriateto pulses arriving in a regular train, is absent hereon account of the statistical distribution of thepulses delivered by the G.M. tube.) Roughly, then,Ta is cut off for a proportion of the time givenby No/ N, and the mean current through it is(I-No/N)Ia. At the upper end of the scale, there-fore, the deflection of the needle approaches thefull-scale value (la) hyperbolically.

_.!:!., 1 { 1 ,~o! j J I! ,i 4. 5 10 aJ5D",

, I, I 111'b" I10 0.5, , , lp , , ,

,,,,I

100

a

Fig. 6. a) Scale obtained with the new circuit; it providesreadings of N/No• the ratio between the average number ofpulses delivered by the G.M. tube per sec and the constantnumber of pulses delivered by the pulse generator per sec.b) The quantity F indicates the percentage measuring errorarising from a (constant) deflection error of 1% of am.". thefull-scale angular deflection. F is plotted logarithmically as afunction of a/am." for a linear scale (curve 1) and for thenew scale (curve2).

The general formula for the mean current iathrough Ta IS:

_ ( 1_ e-N/N,)ia=Ia 1- N/N

o.

A scale in terms of the ratio N/No, based on theabove formula, is drawn in fig. 6a. If No is madeequal to the G.M. pulse rate corresponding to 1mr/h (this rate depends on the properties of thetube employed), then we obtain a scale whose unitdivision corresponds to 1 mr/h.

Comparison of the new scale with a linear scalemakes clear how economical the former is in itsuse of the available scale length. The most obviousbasis for comparing two meter scales, both of whichare used for measuring a quantity x, is the relativeerror of measurement, (LJx)/x. This error in xproceeds from the error LJa in the deflection a

undergone by the needle, and is given by

LJx 1 dx-=--LJa.x x da

Hence the relative error in x per unit deflectionerror is equal to (l/x)(dx/da); this expression,plotted as a function of a, is therefore a suitableyardstick for comparing the merits of differentscales. In fig. 6b the dimensionless quantity

F= amax dxx da

is plotted as a function of another dimensionlessquantity a/amax to give curves that are independentof the unit in which a is expressed. amax standsfor the full-scale deflection. F denotes the percent-age measuring error proceeding from a deflectionerror LJa that is 1% of amax. If it is stipulated thatF shall nowhere exceed 10, then it will be clearfrom fig. 6b that the linear scale has a usable rangeextending from a/amax = 0.1 to a/amax = 1, whilethat of the new scale extends from P to Q, i.e.from a/amax R:i 0.11 to a/amax R:i 0.9. The usableportion of a linear scale covers a measuring rangewhose upper limit occurs at a value 10 times greaterthan the value read at its lo~er limit. On the newscale in fig. 6a points Pand Q occur at values ofapproximately 0.233 and 10 respectively; the upperlimitthus occurs at a value 10/0.233 times or a good40 times greater than the value read at the lowerlimit. Fig. 7a shows two linear scales covering ad-joining ranges; together they cover the range from0.3 to 30 mr/h with a percentage measuring error Fnowhere exceeding 10. Two scales of the new typecovering adjoining ranges, so chosen that the lowerusefullimit of the top scale again lies at 0.3 mr/h,are reproduced in fig. 7b; with the same maximumF value of 10, these have an overall range going upto 500 mr/h. A reasonable estimate can moreoverbe made up to 2000 mr/h. '

Calibration of the new scale is particularly easy.The flip-flop circuit is held first in one and then

mr/h, I , ,o 0,.1

Q I ,,'.1o 3

ft20

J,....iJ

30

IIla

QI

4 5 10: .;pSOoo

I 111'1' 1'1 " I IIIJ IJÎ0,2 0,51200

99257

Fig, 7. a) Two linear scales which together cover the rangefrom 0.3 to 30 mr/h with a percentage measuring error F ~ 10.b) Two complementary scales of the new type. Here the totalrange covered, over which F ~ 10, extends from 0.3 mr/h(as on the linear scale in a) to 500 mr/h (0.5 r/h).

206 PHILlPS TECHNICAL REVIEW VOLUME 21

in the' other of its stable states; the circuit is adjustedso that the meter reads zero in the s~ate in which T3

is cut off and adjusted to give full-scale deflectionin the state in which T3 conducts. All that is nownecessary is to give No the desired value. The nor-mal practice 'will he to place the monitor in theradiation field of a standard radioactive prepara-tion and then to adjust the generator to give a pulserepetition frequency such that the correct meterreading is obtained. Thus only one point has to becalibrated on each measuring range, as in the case.of a linear scale. The new scale is much more con-venient to calibrate than a Iogarithmic scale. Owingto the lack of a zero, a logarithmic scale has to be

calibrated at two points, and adjustment at onepoint entails readjustment at the other.

Summary. Portable radiation monitors are an obvious case inwhich transistors can be used to advantage. As measuringinstruments, however, the circuit employed must be so de-signed that the readings are insensitive to the transistorcharacteristics (whichmay shift considerably with changes oftemperature). A circuit of this kind is described. The authorsalso discuss a method of arrivin.g at a (non-Iinear) scale thatcovers the whole range between zero radiation (zero deflection)and an infinitely high radiation level (full-scale deflection).Readings are sufficiently accurate (less than 10% error for adeflection error equivalent to 1% of the scale length) over apart of the scale extending from 0.1 to 0.9 of the scale length.The radiation-level reading at the upper limit of this range is agood 40 times higher than that at the lower limit. The instru-ment can easily be adjusted to bring any given radiation levelwithin the usable range. The calibration procedure is partic-ularly straightforward.