J.frtW WPPC-T13 
 
 y 
 
 MDDC - 773 
 LADC - 308 
 
 UNITED STATES ATOMIC ENERGY COMMISSION 
 
 THE MODEL 200 PULSE COUNTER 
 
 by 
 
 J. Gallagher 
 
 W. A. Higinbotham 
 
 M. Sands 
 
 Los Alamos Scientific Laboratory 
 
 This document consists of 13 pages. 
 Date of Manuscript: October 10, 1946 
 Date Declassified: March 9, 1947 
 
 This document is issued for official use. 
 Its issuance does not constitute authority 
 for declassification of copies or versions 
 of the same or similar content and title 
 and by the same author(s) . 
 
 Technical Information Division, Oak Ridge Directed Operations 
 Oak Ridge, Tennessee 
 
 12-2S8- c ove r 
 
Digitized by the Internet Archive 
 
 in 2012 with funding from 
 
 University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation 
 
 http://archive.org/details/model300pulsecou5240losa 
 
THE MODEL 200 PULSE COUNTER 
 By W. A. Higinbotham, James Gallagher, and Matthew Sands 
 
 ABSTRACT 
 
 A complete general purpose electronic pulse counter is described which consists of an amplitude 
 discriminator, scale-of-64, and register driver, and is suitable for use with pulse amplifiers in mak- 
 ing nuclear measurements. 
 
 INTRODUCTION 
 
 For the work of the Los Alamos Laboratory a large number of pulse counters were needed. It 
 was desirable, therefore, that a counter be available which could be used in many different applica- 
 tions, which would be quite reliable, and which could be manufactured easily. The counter to be 
 described was designed to fit these requirements. 
 
 The Model 200 Pulse Counter is used for counting (after amplification) the individual pulses from 
 an electrical detector of high-energy particles, i.e., an ionization chamber, proportional counter, or 
 Geiger-Mueller tube. 
 
 The information contained in the signal of an electrical ionization detector consists, at least in 
 part, of the distribution of amplitudes of the pulses under fixed experimental conditions, or of the 
 variation of the rate of occurrence of pulses with a given magnitude with changes in the experimental 
 setup. The counter is suitable for extracting such information from the amplified electrical signals 
 of a detector. 
 
 DISCUSSION > 
 
 The counter to be described consists of three principal parts: (1) an amplitude discriminator 
 which selects for counting only those pulses at its input whose amplitude is greater than some chosen 
 magnitude; (2) an electronic tallying circuit or scaler which provides one output pulse for each 64 
 pulses which it receives; (3) a circuit which operates an electromechanical counter, or register, each 
 time the scaler produces an output pulse. Each of these component parts will be discussed in detail. 
 
 The Amplitude Discriminator 
 
 The amplitude discriminator of the Model 200 Pulse Counter is a trigger circuit which produces 
 a pulse of standard size whenever it receives a pulse whose amplitude is greater than some pre- 
 determined value. The discriminator to be described has the following desirable properties: (1) it 
 can discriminate reliably between pulses which differ in amplitude by a small fraction of one volt; 
 (2) the amplitude at which discrimination occurs is stable to a similar extent; (3) the discrimination 
 voltage is easily adjustable from 10 to 100 volts; (4) it is capable of accepting pulses the duration of 
 which is a few tenths of a microsecond, or greater; (5) it will respond to two pulses separated by 
 intervals as small as 0.5 microsecond; (6) it presents a high impedance to the input signal source. 
 Circuits which in practice have been found to possess the above properties are modifications of the 
 Schmitt trigger circuit. 1 
 
 1 O. H. Schmitt, Jour. Sci. Inst. 15, 24 (1938). See O. S. Puckle also for a discussion of the >^asic 
 Schmitt circuit, Time Bases, John Wiley & Sons, New York, 1943. 
 
 MDDC - 773 f 1 
 
2] 
 
 MDDC - 773 
 
 The particular modification of the Schmitt trigger circuit used in the Model 200 Counter is shown 
 in Figure 1. It is based on two pentodes, since this allows increased speed and greater stability of 
 operation. The operation of the circuit is such that the steady potential at the output lead can take 
 either of two values, depending on the instantaneous potential at the grid of T-l. The components and 
 the supply potential have been chosen such that if the potential at the grid of T-l is less than about 
 + 105 volts with respect to ground, the tube T-2 will be conducting and the tube T-l will be cut off. 
 Under these conditions the output signal lead is at +265 volts. If the grid of T-l is above +105 volts, 
 tube T-l will be conducting, T-2 will be cut off, and the potential of the output signal lead will be 
 +300 volts. The transition from the first state to the second takes place rapidly and irreversibly when 
 the input grid potential reaches some well-defined "triggering" voltage in the vicinity of +105 volts. 
 The value can be made exactly 105 volts by adjustment of the resistor R e . The constancy to which 
 this value is maintained depends on the stability of the +300 volt supply, on the stability of the resist- 
 ances R3, R 4 , and Rg, and on the variations in the cathode emission and contact potentials in the tubes. 
 Because of the symmetry of the circuit these latter effects are minimized. For variations of t 10% 
 in the heater supply voltage and for normal tube variations, over a period of months the triggering 
 voltage seems to remain stable to about 0.2 volts. 
 
 When the input-grid potential is decreasing from values above +105 volts, triggering to the initial 
 state occurs at some value slightly below that at which the initial triggering took place. The difference 
 in the two triggering potentials is called the hysteresis of the discriminator. The hysteresis is 
 determined mainly by the value of the resistance in the plate circuit of T-l. For the components of 
 Figure 1, the hysteresis is about 3 volts, which seems to be a good compromise for a general purpose 
 instrument. If the hysteresis is made zero, there will be a tendency for the circuit to oscillate when 
 the grid is near the triggering potential because of the action of C,. Such a condition should be 
 avoided, since the discriminator may then produce several output pulses in response to one slow in- 
 put signal. 
 
 Operation of the discriminator for slow signals is illustrated in the cyclograms which are given 
 in Figure 2. In all the photographs the sweep voltage is a 2500-cycle per second sine wave of 20 volts 
 peak-to-peak amplitude (positive to the right), which is also fed to the input of the discriminator. 
 Figure 2-A shows the signal at the output of the discriminator. It can be determined from this figure 
 that this particular discriminator has a hysteresis of 4 volts and that a bias setting of 2 volts was 
 used. The vertical height of the oscillogram 2-A is 35 volts. In Figure 2-B and 2-C the vertical de- 
 flection sensitivity and horizontal sweep are the same as for Figure 2-A. In B the potential of the 
 plate of T-l is shown and in C the common cathode potential. The operation of the discriminator in 
 response to pulses may be deduced from the previous discussion. 
 
 As indicated in Figure 1, the signals are introduced into the discriminator by means of a block- 
 ing capacitor and a resistor network for biasing. If the bias potential of the input grid is set by means 
 of R 9 at, say +85 volts, any signal less than 20 volts in amplitude will fail to bring the input grid to 
 + 105 volts, and will not trigger the discriminator. If, however, a positive signal of amplitude greater 
 than 20 volts is introduced, the discriminator will be triggered to the other state and will remain there 
 as long as the instantaneous signal potential is greater than 20 volts. (The duration of the output 
 signal will depend on the amplitude and duration of the input signal.) If the bias potential of the input 
 grid is now set to some value greater or less than 85 volts, the amplitude of input signals which will 
 trigger the discriminator will be decreased or increased in that order. The bias setting of the dis- 
 criminator, or the signal amplitude which is just sufficient to effect the triggering action, is the dif- 
 ference between the triggering potential ( +105 volts) and the bias potentiao of the input grid. The bias 
 setting of the discriminator can be made any value between zero and 100 volts by means of R . If the 
 dial of R 9 is graduated from zero to 100 over 270°, it can be made to read directly the bias setting. 
 This is effected by setting the rheostat R, until a rotation of 270° of R 9 gives a potential change of 
 100 volts at the input grid and then setting the dial on the shaft of Rg so that the indicator is at zero 
 when the potential is +105 volts with respect to ground. Bias settings less than the hysteresis cannot 
 in general be used. For the sake of stability the resistors R 3 , R 4 , R^ , R 7 , R 8 and Rg are usually made 
 wire wound. Since R,, R 2 and R 5 do not directly determine the triggering potential, composition or 
 metallized resistors are used. If some sacrifice in long time stability can be made and if the direct 
 
 t 2-26M-p2- bu 
 
MDDC - 773 
 
 INPUT 
 
 OUTPUT 
 
 R4 
 
 R 5 J40K 
 
 10K ,*- Re 
 
 P20K 
 
 Figure 1. Circuit diagram of the amplitude discriminator. 
 
 Figure 2. Cyclograms of the discriminator operation 
 peak-to-peak amplitude of 20 volts. 
 
 A. Waveform appearing at the plate of tube T-l. 
 
 B. Waveform appearing at the plate of tube T-2. 
 
 C. Waveform appearing at the common cathode. 
 
 The time base is a 2500-cps sine wave with a 
 
 12-2»8-p3-»» 
 
4 I MDDC - 773 
 
 reading bias dial is not desired, an appreciable economy is effected by omitting R e and R 7 and by 
 using carbon resistors throughout. 
 
 For a large input pulse having a small rise time, the rise time of the output pulse depends only 
 on the resistance R 2 and the parasitic capacitance shunting it. For fast pulses that just trigger the 
 discriminator and for slow input signals, the time of rise of the output pulse is slightly greater. Some 
 decrease in the rise time of the output signal can be achieved by applying shunt compensation to the 
 plate circuits of both tubes. It should be noted that the output signal may never reach its full value if 
 the input signal does not remain above the bias voltage for a sufficient length of time. For this reason 
 it is desirable that narrow input pulses should either be rectangular in shape or at least have a suffi- 
 ciently long flat top. With the present circuit only pulses of duration greater than 0.1 microsecond 
 are counted. The capacitor Cj should ideally be chosen to obtain a frequency independent voltage 
 divider from the plate of T-l to the grid of T-2. The value given in Figure 1 is actually larger than 
 would be required according to the above criterion, but seems to be more suitable in a general pur- 
 pose instrument. 
 
 The maximum permissible duration of the input pulses is determined solely by the time constant 
 of the input coupling network. If slow pulses which are not large compared with noise are being 
 counted, it is possible for the discriminator to respond to each of several noise waves on the rise or 
 fall of the pulse, giving many counts for each pulse. This can be avoided either by making the hys- 
 teresis at least as large as the noise or by making the resolving time of the conjoining circuit slightly 
 greater than the duration of a single pulse. 
 
 Since it is often necessary to count small pulses in the presence of much larger ones, it is 
 important that the discriminator not overload easily. The grid of the tube T-l is normally well below 
 its cathode potential. Once the triggering voltage has been exceeded, however, T-l acts as a cathode 
 follower and continues to present a high impedance to the signal source. For the circuit values shown 
 in Figure 1, the tube T-l does not draw grid current on slow signals until its grid potential reaches 
 about +200 volts with respect to ground. Thus if a bias setting of a few volts is used, the circuit will 
 not overload for pulses of 100 volts amplitude. If the input pulses have a rise time of less than 0.1 
 microsecond and if the capacitance to ground of the cathode lead is unduly large, grid current will be 
 drawn for somewhat smaller signals. Grid current should be avoided, since it alters the charge on 
 the input capacitor and results in a temporary shift of the bias setting of the discriminator. Although 
 it might appear desirable to make the input time constant short, to allow a speedy recovery in such 
 cases, other considerations demand that the time constant be large. 2 The value 0.01 second seems to 
 be a reasonable compromise. 
 
 The Scaler 
 
 Since it is often either desirable or necessary to take data at a more or less rapid counting rate, 
 it is convenient to circumvent the limitations of mechanical registers by the use of scaling circuits. 
 The Model 200 Counter was developed primarily for counting pulses which are distributed randomly 
 in time, although it has been used for counting pulses the occurrence of which was neither regular nor 
 random, according to the usual definitions. 
 
 Although a quantitative discussion of the errors introduced by practical apparatus when used for 
 counting random events will be found in the literature, 3 a few remarks may be in order. Any counting 
 circuit possesses a dead time T (usually referred to as the resolving time of the counting circuit) 
 following each pulse it records, during which time it is incapable of recording another pulse. If the 
 arrival of pulses to the counter is perfectly random, and the pulses occur at an average rate of n per 
 second, then the fractional counting loss due to the dead time of the counting circuit is nT, so long as 
 nT«l. It is apparent that if one wishes to count random pulses occurring, for instance, at an average 
 rate of 1000 per second, with only one per cent counting loss, the resolving time of the counter must 
 
 2 This time constant is essentially one of the coupling elements of the preceding pulse amplifier. 
 
 3 
 
 W. B. Lewis, Electrical Counting, Camb. Univ. Press, 1S43. 
 
 12-26B-p4-bu 
 
MDDC - 773 
 
 [5 
 
 be 10 microseconds. A resolving time of 10 microseconds, however, does not necessarily require 
 that the counter be capable of handling 10 5 regularly spaced pulses per second. 
 
 It is true that a single resolving time completely characterizes a counting circuit only if the 
 counter comprises a long chain of identical scaling elements, or if the resolving time of each element 
 in the chain is just sufficient for it to pass along the pulses which it receives when pulses are enter- 
 ing the chain at the maximum regular rate of 1/T,, where T x is the resolving time of the first scaling 
 element. In a practical counting circuit, it is not convenient or economical to meet this requirement 
 in a strict sense. An electronic scaling circuit is terminated ordinarily by an electromechanical 
 register which can record 10* or more impulses but whose resolving time may lie in the range 0.1 
 to 0.01 second. It would be necessary to precede the register by a scale-of-10 4 , if a counter with an 
 ideal resolving time of 10 microseconds were to be achieved. Since it can be shown that when ran- 
 domly occurring pulses are being counted, a considerable "smoothing" effect exists after the count- 
 ing rate has been reduced by a large scaling factor, it is unnecessary to have the ideal resolving time. 
 A scale-of-64 seems to be sufficient for most cases, although a number of scales-of-256 and greater 
 have been used. 
 
 The Model 200 Counter contains a scale-of-64, consisting of six scaling stages, each of which 
 is a scale-of-2. The counter is usually used in connection with a register whose resolving time is 
 0.02 second. The resolving time of the first scaler is 5 microseconds, and the resolving time of the 
 successive stages is such as to meet the ideal requirement given above. The resolving time of the 
 register does not meet the ideal requirement, but, because of the smoothing effect mentioned, the 
 counter can be used to count random pulses at an average rate of 10 5 per minute with less than one 
 per cent counting loss. 
 
 The scale-of-2 employed in the counter seems to be superior to any yet suggested. It requires 
 no adjustment but operates with complete reliability. It is simple and economical to construct, and 
 the checking and servicing can be done by an average technician, since failures are usually due to 
 initially defective parts or wiring. The circuit was developed in 1943 by one of the authors, 
 W. A. Higinbotham. 
 
 Since more than a thousand scales-of-2 have been used at Los Alamos, it was found convenient 
 to construct each scale-of-2 as a plug-in unit. The type of construction employed is illustrated in 
 Figure 3. Since it is not necessary that all six units constituting the scale-of-64 be identical to main- 
 tain an ideal resolving time throughout the scaler, two different types of unit are used. The basic 
 circuit for both types is shown in Figure 4. Suitable values for the circuit components are indicated 
 in the caption for the figure. 4 The first two scale-of-2 units in the Model 200 Counter are 6SN7 units 
 and have a resolving time of 5 microseconds. 5 These are followed by four 6SL7 units, which have a 
 resolving time of 20 microseconds. Both units are constructed in much the same way and have all 
 necessary connections brought out to the pins of an octal plug. Since the 6SL7 units dissipate one- 
 half as much power and require one-fourth as much current from the positive supply source, it is 
 economical to use them in applications where the shorter resolving time of the 6SN7 units is not 
 needed. The operation of the two units is quite similar; only the 6SN7 unit will be discussed in detail. 
 
 The scale-of-2 circuit, shown in Figure 4, is based on the well-known Eccles-Jordan trigger 
 circuit 6 (called here a "flip-flop") which is here symmetrically coupled to a single source of trig- 
 gering by means of two diodes. The circuit is capable of being triggered alternately from one state 
 
 4 Recently germanium crystal diodes have been found to work quite satisfactorily as coupling ele- 
 ments in the scale-of-2. Two Sylvania -type 1N34 diodes in series are used to replace each half of 
 the 6H6. 
 
 5 The 6SN7 scaler built with the same components but not with plug-in type construction will usu- 
 ally having a resolving time of about two microseconds. 
 
 8 For operation and theory of the Eccles-Jordan trigger circuit, see H. J. Reich, Theory and 
 Applications of Electron Tubes, 2d ed, McGraw-Hill Book Company, New York, 1944. 
 
 1 2-258- pB-bu 
 
6] 
 
 MDDC - 773 
 
 Figure 3. Photograph of the plug-in scale-of-2. 
 
 OUTPUT 
 
 < 
 
 O 
 O 
 
 <n 
 
 o 
 
 25! 
 
 i- 
 o 
 
 UJ 
 
 o 
 o 
 
 Figure 4. Basic circuit of the scale-of-2. For 6SN7 unit: T-2 — 6H6; T-3 and T-4 — 6SN7; 
 R — 20k; Rj; — 5k; R s — 15k; R„ and R 5 — 200k; R 6 and R 8 — 100k; R 7 — 10k; Cjand C 2 — 50/n/Jf; 
 C s — 0.01. For 6SL7 unit: T-l and T-2 — 6H6; T-3 and T-4 — 6SL7; R— 100k; R 2 — 25k; 
 R 3 — 75k; R 4 and R 5 — 200k; R 6 and R 8 — 500k; R 7 — 40k; C x and C 2 — 40 m/*J C s — 0.01. 
 
 !2-2o8-pfl-bu 
 
MDDC - 773 
 
 [7 
 
 to the other by successive, identical, negative triggers. The essential properties of the circuit, in 
 addition to those inherent in the flip-flop, concern the switching mechanism by which the effect of the 
 trigger is different on the two halves of the circuit, depending upon which half is conducting, and the 
 choice of operating conditions which give a large measure of stability. The use of diodes to accom- 
 plish a unilateral switching action in the manner indicated in Figure 4 is believed to be new. 7 In 
 normal operation the "reset" lead is externally connected to ground, and the "interpolator" lead 
 goes to a high-impedance indicator. 
 
 If the input signal lead has a steady potential of +300 volts, neither of the diodes will be conduct- 
 ing, since the potential of the plates of both diodes is less than their common cathode potential. Let 
 us assume that the state of the flip-flop is such that tube T-4 is at first conducting. Its plate will then 
 be at +150 volts, its cathode at +80 volts, and its grid will be held at +82 volts by virtue of grid 
 current flow. The plate of T-3 will be at +280 volts and the grid of T-3 at +50 volts, thereby cutting 
 off the tube. If the input voltage is slowly decreased, the diode T-2 begins to conduct. Since its 
 resistance is small compared with the resistance Rj , the plate of T-3 follows the signal. The grid of 
 T-4 does not move in the manner which might at first be expected, but drops only slightly in potential 
 so long as grid current is being drawn. A decrease in the grid current takes up the decrease in volt- 
 age across R 4 until the input signal reaches 240 volts. It will be noticed that, when the input signal 
 is held fixed at either +300 or at +250 volts, the flip-flop can exist quite stably in either of its two 
 possible states. Also, as long as one of the tubes is drawing grid current, the circuit has no tendency 
 to be triggered by small fluctuations at the input signal lead. In tact, when the input signal lead was 
 at +250 volts, in the above discussion, a small variation at that point appears as a much smaller signal 
 at the output lead. This fact is of importance, since part of the simplicity of this circuit lies in the 
 possibility of connecting an indefinite number of such circuits directly together with no resulting 
 instability. 8 The self-biasing arrangement in the catnode circuit of T-3 and T-4 adds greatly to the 
 stability of the flip-flop. 
 
 Let us suppose now that the input signal voltage is brought suddenly from +300 volts to +250 volts. 
 The potential of the grid of the conducting tube T-4 immediately goes negatively by about 30 volts, an 
 amount which is determined by the ratio of Q to the static parasitic capacitance at the grid of T-4. 
 Immediately following the input signal, the plate potential of T-4 starts to rise rapidly, and the grid of 
 T-3 moves positively a proportionate amount determined by the ratio of Q to the parasitic capacitance 
 at the grid. The tube T-3 begins to conduct, and regeneration causes the circuit to proceed toward 
 its other equilibrium state. Before the state equivalent to the original state is reached, however, the 
 plate of T-4 is "caught" by the diode T-l and the circuit assumes a stable state with the plate of T-4 
 at +250 volts. If now the input voltage is returned either quickly or slowly to +300 volts, there is no 
 effect on the trigger circuit other than causing it to assume the stable state which is the exact mirror 
 image of the original state described. The application of a second trigger similar to the first causes 
 the circuit to pass through a similar cycle of operations in which the role of T-3 is now replaced by 
 that of T-4, and vice versa. 
 
 It should be noticed that the diodes provide a switching action, in that on application of the trig- 
 gering signal a potential change is transmitted to only one grid, and that no signal appears at the 
 opposite grid to counteract the effect of the trigger. Also, as soon as the triggering action has begun, 
 the negatively moving plate "disconnects" itself from the input lead and the flip-over process pro- 
 ceeds unhampered. Because of this, triggering can be effected by a short signal (as is done in the 
 first stage of the counter) or by a rectangular signal (as is done when a stage is triggered by the pre- 
 ceding one). The reliability of the circuit rests in the switching action of the diodes. The capacitive 
 portion of the couplings employed in the flip-flOR not only allows rapid operation of the circuit but is 
 
 7 Subsequent to the development of the circuit shown here, it was pointed out to the authors that 
 scales-of-2 using metallic rectifiers are described by W. B. Lewis (op. cit.). The mechanism of 
 operation, however, is somewhat different. 
 
 8 As many as twenty stages have been used in one continuous chain. 
 
 12-258-p7-bu 
 
8 ] MDDC - 773 
 
 an essential part of the triggering mechanism as described. The output connection of the scale-of-2 
 alternates in potential between +300 and +250 volts during the operation. This signal has the correct 
 d-c level and the correct amplitude for triggering the next scale-of-2 unit to which it is direct- 
 coupled (to pin 4). Evidently any number n of the scale-of-2 units can be direct- coupled in a chain 
 to form a scale-of-2 n . The use of diodes as automatic switching devices, as shown here, is also very 
 useful in many other types of circuit.. 9 
 
 Since a finite time is required for the flip-over action, and since an additional recovery time 
 must elapse after flip-over before static conditions again prevail, the scale-of-2 circuit will respond 
 to a second input signal only after a certain time has elapsed following a previous signal. The re- 
 solving times given above were achieved by striking a compromise between these two effects and the 
 required reliability of the circuit. Similar scales-of-2 with varying degrees of reliability have been 
 made, with resolving times down to 0.2 microsecond. 
 
 Pulse forms which appear in a chain of scaling units are shown in Figure 5. In all the photo- 
 graphs the vertical scale is the same (positive upwards) and in all but F the same linear horizontal 
 sweep was used. Double positive pulses with an amplitude of 50 volts are being fed to the input of the 
 discriminator. The pulses are 5 microseconds apart and are repeated 1000 times a second. In Figure 
 5-A the input pulses are shown. The signals which appear at the input to the first scaling stage are 
 shown in B, and the output of that stage (input to the second stage) is shown in C. Figure 5-D shows 
 the signal between stages two and three. The signal between stages three and four is shown in E. The 
 same signal is shown in F but here with a total horizontal sweep time of 11 milliseconds. 
 
 The interpolation indicator for the scale-of-2 is a 1/25-watt neon lamp connected to the plate of 
 T-3 through a 1-megohm resistor. As soon as T-3 is conducting and its plate potential has dropped 
 to. +150 volts, the lamp will fire and carry a current of 100 microamperes. A neon lamp so operated 
 gives sufficient light to serve as an effective indicator for normal visual purposes and has a negligi- 
 ble effect on the operation of the scale-of-2. The interpolation lamps are, of course, not necessary 
 when a large number of counts occur in each observation, but they are useful at slow counting rates 
 and serve as a visual indication of the proper operation of the circuit. 
 
 The Register Driving Circuit 
 
 Since economy dictates the use of an electromechanical impulse register to terminate a scaling 
 circuit, the final function of a counter circuit is the production of signals for operating a register. 
 Most registers consist of a ratchet-operated mechanism which involves a dial or a wheel counter, 
 indicating one digit for each electrical impulse received. The limiting speed at which a register will 
 record impulses faithfully is dependent on the mechanical inertia of the mechanism and on the amount 
 of power which can be supplied to effect the necessary mechanical motion. Energy is supplied to a 
 register by means of current through a coil which has a large inductance. Evidently the rate at which 
 a register can operate depends in part on the driving voltage, the maximum value of which is limited 
 by the breakdown voltage of the insulation used in the register windings. Also, if the mechanical iner- 
 tia of the register mechanism is to be the limiting factor in the speed of operation, it is necessary 
 that the resistance of the driving source be high enough that the time constant L/R of the register 
 associated circuit will be short compared with the time for mechanical motion. For this reason the 
 register is best operated in the plate circuit of a pentode and with a plate supply having the highest 
 voltage conveniently available. 
 
 Registers are commercially available which have been designed specifically for use with elect- 
 ronic counters. Registers of this type made by the Cyclotron Specialties Company of Moraga, Califor- 
 nia, are used with the Model 200 Counter. They have a minimum operating current of 10 ma, an 
 inductance of 3.3 henrys and a d-c resistance of 1000 ohms. A register of this type can be operated 
 at continuous counting rates of 50 impulses per second with the circuit shown in Figure 6. 
 
 9 A diode switching arrangement has been used to convert a scale-of-16 into a decade (scale-of-10) 
 scaler, which operates completely satisfactorily. The scale-of-10 circuit will be described at a 
 later date. 
 
MDDC - 773 
 
 [9 
 
 Figure 5. Pulse forms which appear in the pulse counter with a linear time base. 
 
 A. Input pulses of 50-volts amplitude separated by 5 microseconds. 
 
 B. Input to the first scale-of-2. 
 
 C. Output of the first scale-of-2. 
 
 D. Output of the second scale-of-2. 
 
 E. Output of the third scale-of-2. 
 
 F. Output of the third scale-of-2 with the linear time base 11 microseconds in length. 
 
 REGISTER 
 TERMINALS 
 
 
 
 iaf 
 
 Figure 6. Diagram of the register driving circuit. 
 
 !2-2B8-p9-t>u 
 
10 ] MDDC - 773 
 
 The plate voltage for the output stage of the driver circuit is secured from the unstabilized part 
 of the power supply. This arrangement prevents transients, which arise when the register is actuated, 
 from interacting with other parts of the counter and causing spurious counts. The higher supply 
 voltage is also advantageous. The driver stage shown does not load the final scaler stage to such an 
 extent that the scaling operation becomes unreliable. The impedance inserted across the register 
 winding is sufficient to prevent the production of voltage surges large enough to damage the insulation. 
 The circuit operation consists essentially of placing a sufficiently high voltage across the register 
 until it closes, and then of removing this voltage to enable the register to recover in preparation for 
 the next count. If the square voltage pulse is too long, the maximum permissible counting rate is 
 reduced, whereas if it is too short, the register may occasionally miss. In the present circuit the 
 signal which is applied to the grid of the driver tube is a very nearly rectangular, positive voltage 
 pulse of 10 milliseconds duration. This seems to be sufficient for reliable operation of the registers. 
 
 The tube T-l of Figure 6 and its associated components serve as a device for producing the 
 requisite pulse shape. The negative step pulse which appears at the output terminal of the last scaling 
 stage is converted into a triangular pulse at the grid of T-l by the resistance -capacitance network in 
 the grid and the mechanism of grid current. This triangular wave cuts off the plate current in T-l 
 for 10 milliseconds. The plate of T-l would rise to the positive supply potential during this 10 milli- 
 seconds were it not that the grid of T-2 will always be driven at least to the cathode potential and held 
 there for the 10 millisecond period of operation of the register. The circuit shown also insures that 
 operation of the register is reliable at rates from the slowest to the maximum permissible (about 50 
 per second). The register can be put directly in the plate circuit of T-2, but for the safety of person- 
 nel, as well as the register insulation, one side of the register is grounded and the operating current 
 is supplied through the blocking capacitor C s . 
 
 This driver circuit can be used for any register which does not require more than 40 milli- 
 amperes or 300 volts. For the optimum speed of operation of any register, the time constant in the 
 grid circuit of T-l should be suitably chosen. The time constant in the grid of T-2 is not critical and 
 need only be sufficiently large. 
 
 The Complete Counter 
 
 The circuit diagram of the complete Model 200 Counter is given in Figure 7. In addition to the 
 circuit elements already described, it contains a coupling stage between the discriminator and the 
 scaler, a regulated power supply, and an output stage to an external counting rate meter. A photo- 
 graph of a completed counter containing plug-in units is shown in Figure 8. 
 
 In Figure 7, tubes T-l and T-2 comprise the discriminator. The additional components in the 
 grid circuit permit the input to be connected and disconnected at the beginning and end of a measure- 
 ment without introducing spurious switching counts at the same time. 
 
 The positive signals from the discriminator are converted into sharp pulses, inverted, and ampli- 
 fied by the tube T-3A and its associated components. The 65-volt negative signal thus produced then 
 goes to the first scale-of-2. The wiring diagram to the sockets for plug-in units is shown in the 
 figure. The numbers which appear correspond to those shown in Figure 4. The first two plug-in units 
 are 6SN7 units, the last four, 6SL7 units. The reset leads are connected together and to ground 
 through a normally closed push-button switch. When this switch is opened momentarily, the potential 
 of the reset line rises about 50 volts and returns all the flip-flops to the starting state. The resistor 
 across the switch is necessary to prevent the reset line from going too far positive when the connec- 
 tion to ground is broken. If the reset line goes much more than 50 volts positive, the cathode potential 
 of the flip-flops changes so much that the circuits are momentarily unstable when the switch is closed 
 again, and resetting is not accomplished. The interpolation indicator circuit (neon lamps and 1-meg- 
 ohm resistors) is shown. The lamps are extinguished in the starting position. They are labelled 1, 
 2, 4, 8, 16, and 32. The sum of the numbers corresponding to the lamps which are lighted shows the 
 number of pulses received after the last multiple of 64. The tube T-3b is a cathode follower which 
 couples the output of the first scaling stage to an external counting rate meter. The signal which 
 
 12-208-plO-bu 
 
MDDC - 773 
 
 [11 
 
 
 O 
 
 o 
 
 <D 
 W 
 
 r— I 
 
 a 
 
 <u 
 
 ■m 
 
 •— i 
 
 s 
 
 o 
 o 
 
 0) 
 
 5 
 
 a 
 SP 
 
 73 
 
 3 
 O 
 
 -r-f 
 
 u 
 
 u 
 
 bD 
 
 1 2-2B8-P 11-bu 
 
12] 
 
 MDDC - 773 
 
 0! 
 O 
 00 
 
 act 
 
 3 
 
 c 
 
 X! 
 
 co 
 u 
 
 V 
 
 I 
 
 u 
 
 "8 
 
 0) 
 
 a 
 
 S 
 o 
 u 
 
 0) 
 
 -C 
 
 •a 
 
 5 
 
 a, 
 
 12-2SB-P12-- 
 
MDDC - 773 [ 13 
 
 appears at this point has a shape suitable for operating a very simple meter, which can be used for 
 checking bias curves or operating conditions with given discriminator settings. The output of the 
 sixth scaling stage is coupled directly to the register circuit described above. 
 
 Since a constant plate supply potential is needed for the discriminator, it is nearly as economical 
 and somewhat more reliable to operate the complete counter from a stabilized plate supply. The 
 supply shown in Figure 7 is of the degenerative type. This supply provides 55 milliamperes at 300 
 volts and an additional 5 milliamperes at 450 volts. The 300 -volt supply is stable for input line volt- 
 ages from 95 to 135 volts. It has a stabilization factor 10 of 700, an output resistance of 2 ohms, and 
 has less than 3 millivolts of 120 cycle per second ripple at its output. 
 
 Several Model 200 Counters or their equivalent have been made and are in use at the Los Alamos 
 Laboratory. Satisfactory operation has always been obtained. Failures which have occurred are 
 usually due to either: (1) poor mechanical construction in the scaler units, i.e., bad solder joints, etc; 
 (2) resistors which were more than 20 per cent outside of the rated values, particularly in the sym- 
 metric parts of the circuit (no trouble has been experienced when 5 per cent tolerance or matched 
 resistors were used); (3) failures or extreme unbalance in tubes. All tubes are operated within rated 
 dissipations and voltages, but, since such equipment is usually kept operating 24 hours a day, weak 
 tubes soon show up. Many counters, however, have been operating for more than a year with no 
 servicing. 
 
 ACKNOWLEDGEMENTS 
 
 The authors are greatly indebted to the many individuals contact with whom has made the develop- 
 ment of the Model 200 Counter possible. One of us owes much to the ideas exchanged with those of 
 the Radiation Laboratory of the Massachusetts Institute of Technology. Also, much credit is due to 
 those members of the Los Alamos Laboratory who gave valuable suggestions during the developmental 
 work and aided in preparing this paper — in particular to D. K. Froman, W. C. Elmore, and R. J. Watts. 
 
 10 Ratio of fractional change in input voltage to fractional change in output voltage. 
 
 12-2b8-pl3-noou-fl 
 
lB N i V .£i£ l I V OF FLORKM 
 
 3 1262 08910 5240 
 
 I ! i