Zero Inefficiency Anticoincidence CircuitM. Degallier Citation: Review of Scientific Instruments 21, 1025 (1950); doi: 10.1063/1.1745489 View online: http://dx.doi.org/10.1063/1.1745489 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/21/12?ver=pdfcov Published by the AIP Publishing Articles you may be interested in An Automatic Zero Suppression Circuit Rev. Sci. Instrum. 41, 1298 (1970); 10.1063/1.1684799 Anti-Coincidence Circuit for Use in the Millimicrosecond Region Rev. Sci. Instrum. 31, 664 (1960); 10.1063/1.1931290 A Lossless Anticoincidence Circuit Rev. Sci. Instrum. 23, 567 (1952); 10.1063/1.1746092 A OneShot Multivibrator Anticoincidence and Recording Circuit Rev. Sci. Instrum. 18, 31 (1947); 10.1063/1.1740813 Circuit for Anticoincidences with GeigerMüller Counters Rev. Sci. Instrum. 11, 84 (1940); 10.1063/1.1751656

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sun, 23 Nov 2014 17:11:46

LABORATORY AND SHOP NOTES 1025

A

B

KOVAR ==1="" 5.S.BELLOWS ~=='<;IIj"-FRAM E..

FIG. 2. Sintered glass-bellows valve.

thorough outgassing, it does eliminate the vacuum-pressure manipulating line and it is spill-proof.

* Now at RCA Laboratories. Princeton, New Jersey. 1 Taylor and Young, Ind. Eng. Chern. Anal. Ed. 17, 811 (1945).

Zero Inefficiency Anticoincidence Circuit M. DEGALLIER

Laboratoire de Recherches Nucleaires. Ecole Poly technique, Universite de Lausanne, Lausanne. France

September 12, 1950

L ET us consider a mixing circuit composed of two pentodes, one being normally conductive, receiving the negative

coincidence pulse (C) formed somewhere else, and the other, normally cut off, receiving the positive square pulses (5) from the anticoincidence counters. Any C-pulse not accompanied by an 5-pulse thus gives rise to a positive pulse which is termed anti­coincidence (A = C - 5).

A 100 percent efficiency has been claimed in the literature for such a circuit using a flip-flop to generate the 5-pulses. That is not exactly true, though the inefficiency may be small enough to be negligible in certain experiments; in others it can be very harmful and intolerable. It is due to the existence of the recovery time of the flip-flop. The expression giving the inefficiency has been calculated in the following way:

Let nc, n" n,a be respectively the rates of pulses C, 5, and spurious anticoincidences, and <p the dead time of the flip-flop, defined as the shortest interval between the end of a pulse and the front of the next one.

Thus every C-pulsefallingin that interval will give rise to a spuri­ous anticoincidence, the rate of which is therefore n,a=nc'n,' <p.

The resulting inefficiency is consequently n,' <p, and the only way to get rid of it is to design the circuit in order to have <p=0.

The proposed circuit for generating the anticoincidence pulses 5 (Fig. 1), similar to that of Putman, [Proc. Phys. Soc. 61, 312 (1948) ] satisfies this demand: its essential characteristic is that it can be re-excited at full pulse duration inside or outside its cycle. At each positive input pulse on the grid of T1, the condenser Cis fully discharged (this is true only if the input pulse is sufficiently long) and then recovers while generating the exponential timing wave which determines the duration of the 5 pulse.

This circuit, developed for very low inefficiency measurements of counters, has been tested by feeding the 5 input terminal from

'Olt-I IO~-' +ZSo

~:~ J ItAIJ7

...... , 13 F·

OI

T4 -- -- n '-+----i IM-l::

,0/(., '10k•,

-::- "";"

FIG. 1.

17 counters counting altogether at 170 c.p.s. while the C input terminal was fed only from the pulses of one of these counters (4 c.p.s.). The pulses from the single counter were, of course, coincidences and not a single anticoincidence has been registered for 1oo coincidences.

Varying Nuclear Emulsion Sensitivity with Altitude

ARTHUR BEISER

Department of Physics, New York University, University Heights. New York, New York September 12, 1950

I N reviewing cosmic-ray detection techniques, Korff! points out that a major deficiency in nuclear emulsion work is the

continuously sensitive property of emulsions, rendering uncertain the circumstances attendant upon the recording of a particular event. Methods of eradicating previous phenomena in a nuclear emulsion before use have been described2 and are frequently helpful, but a process for sensitizing and desensitizing plates at prescribed times would be of very much greater value.

The dependence of the sensitivity-reducing action of photo­graphic desensitizers on oxygen concen tra tion3 suggests the possi­bility of limiting the sensitivity time of nuclear emulsions by the use of such agents. The dyes used for this purpose apparently act as catalysts for the oxygen in its reaction with the emulsion, and are capable of preventing the formation of a latent image while not ordinarily sufficient to attack an existing one. Since the density of atmospheric oxygen varies with altitude, the use of desensitizers is of especial interest in connection with balloon and perhaps also airplane and rocket flights with nuclear plates.

To determine the degree of desensitization attainable and the magnitude of its reduction under the conditions of a balloon flight, preliminary experiments were performed with pinakryptol green (1: 5000), Eastman Kodak NTB emulsions and a polonium alpha­particle source. It was found that the desensitized plates, irradi­ated and processed with a standardized procedure, exhibited an optical density at the area of exposure approximately ten percent that of untreated control plates. The tracks themselves were very faint, from one-half to two-thirds shorter than the control tracks, and exhibited a grain structure along almost all of their length, while the controls showed no grain structure beyond the first few microns of track length.

The simulated balloon flight occupied one hour, with the pres­sure in a chamber containing the desensitized plate and the source being reduced to 2.5 cm Hg (corresponding to ~75,000 ft. or 23,000 m altitude), where it was held for five minutes. After processing, the plate had a density almost three times that of a control, exposed for the same length of time outside of the cham­ber. Reduced pressures had no influence on plates not desensitized. Examination of the tracks revealed two distinct kinds, with relatively few transitional ones: a background of weak tracks identical to the ones in the control, and a number of heavier tracks as long as those in the non-desensitized plates but with a slight grain structure for part of their length. The latter tracks were presumably recorded at the lower pressures and account for the increased density. This moderately complete separation of the tracks in two classes would seem to indicate the existence of a discontinuity in the desensitizer action. Such a discontinuity, or cut-off region, if its dependence on altitude and rate of ascent can be determined, would be extremely desirable, making possible both fairly reliable measurements of recorded phenomena and their correlation with altitude. .

A series of experiments is in progress to investigate the variation of the sensitivity of desensitized emulsions with the above factors and on the nature and concentration of the desensitizer and the emulsion composition and thickness. The ultimate utility of this method will depend largely on the magnitude of the sensitivity increase with altitude and on the existence and nature of a cut-off

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sun, 23 Nov 2014 17:11:46

1026 LABORATORY AND SHOP NOTES

region. It is hoped that some conclusions may be reached on these points.

1 S. A. Korff, Rev. Mod. Phys. 20, 327 (1948). 2 H. Vagoda. Radioactive Measurements with Nuclear Emulsions (John

Wiley and Sons, Inc., New York, 1949), p. 108; M. Wiener and H. Yagoda, Rev. Sci. Inst. 21, 39 (1950).

3 M. Blau and H. Wambacher, Nature 134, 538 (1934).

Cathode-Follower Fallacies* PAUL l. RICHARDS

Brookhaven National Laboratory, Upton, Long Island, New York October 19, 1950

I T is often statedl in the literature that the low output impe­dance of a cathode follower offers the best solution to the

problem of rapidly driving a large capacity or low impedance line. While the cathode follower is indeed an excellent driver for these loads and has certain definite advantages in many applications, the statement as it stands is somewhat oversimplified and has led to the use of cathode followers in circuits where they are unnecessary and some times even unsuited. Actually the ability to drive these loads rapidly is not a unique attribute of the cathode follower. Let us examine more carefully the relative properties of the cathode follower and conventional amplifier; various quantities of interest are listed in Table 1. Note that the output capacity of the tube can safely be omitted because it will either be small compared to a large load capacity or, across a low impedance line, will affect the rise time less than preceding or succeeding stages.

Considering first the problem of driving a low impedance line, it is often stated that the low output impedance of the cathode follower is required to match the sending end of the line. But it is not always important to match the sending end, for if the far end is terminated by the proper resistor, the reflections will in many applications not be serious. In such applications, then, the problem reduces to obtaining maximum signal voltage across the low impedance without unduly loading earlier stages. With a given tube and given load resistance Zo, greater gain is invariably ob­tained with the "amplifier" circuit (Table I). This gain, of course, will be unity or less with presently available tubes and coaxial lines, but it will still be greater than that obtained from cathode con­nection; moreover the low gains involved result in input capacities for the two circuits which differ, even for a triode, by less than might at first be expected. Therefore, in applications where maxi­mum gain is desired, the amplifier circuit may be preferred.2 The two circuits, of course, have opposite output polarity and this consideration alone will often save one (inverting) tube in such applications.

With a largely capacitative load, the situation is somewhat more

TABLE 1.*

Amplifier ~~~~~- --------

Output impedancE'

Gain =A

Rise time

Mid-band gain

rise time

Input capacity

gm/C L(21f")i

Triode: CgK+Cop (1 +A) Pentade: CuK+CgSc

Cathode follower

-~~~l/glll 1+"

1 +,,+rpYr,

CLCz,,)1

Triode: Cup +CgK(1-A) Pentocle. triode-connected CgSo+Cg KCI-A) Pentacle-connected (Cg S,+Cg K)(I-A)

* Notation: rp =piate resistance. J.L = amplification factor. gm = trans­conductance. Y L =GL +jwCL = load admittance. CL =load capacitance. CoK =grid to cathode capacitance. COP =grid to plate capacitance. CaSe =grid to screen capacitance. A =gain.

deceptive. It might appear that the low output impedance of the cathode follower is required to give a reasonably fast rise time. If, however, CL is given and CD may be varied (the usual situation), Table I shows that the gain band-width product is the same for the cathode follower as for the amplifier. Physically, the greater "inheren t" gain of the amplifier circuit allows one to use a lower resistor across CL thereby compensating for the higher output impedance and resulting in the same final rise time for the same gain. However, it will be noted that a given tube (that is, a given maximum plate current) cannot develop as large a maximum output voltage across the lower shunt resistance as it could with cathode connection. In other words, although the gain band-width products are the same for the two connections, the maximum­output-voltage-band-width products are not. Again the gains in­volved are low in any case, and previous remarks apply to input capacities. Thus, in those applications where large voltages are not required, the difference in output polarity of the two connections may lead to the choice of the amplifier.

Some general remarks on the preceding observations must be made. First, a pentode-connected cathode follower has a lower input capacity than any of the other circuits; if input capacity is the prime consideration the pent ode-connected cathode follower is definitely indicated. Secondly the cathode follower, being a feed­back circuit, has a slight edge as regards linearity and/or stability although the feed-back factor is not very large in those cases where RL""l/gm. Finally, as already noted above, the cathode connec­tion will drive a large capacity to a higher maximum voltage than any other connection with the same rise time and the same tube.

Thus the cathode follower has a definite superiority in the more exacting applications, but many less critical circuits can be unnecessarily complicated by its use.

As a simple example, suppose that a triggering pulse of 3 volts must be sent down a very long line of 100 ohms characteristic impedance and that the circuit to be triggered will not respond to pulses less than about! volt. If the line is terminated by a 100-ohm resistor the reflections will usually be very much less than! volt, so there will be no need to match the input of the line. Suppose further that the available trigger comes out of its generating circuit at a level of 4 volts. If half of a5687 tube (gm= 10,000, IL= 20, rp= 2000) is available to drive the line, the amplifier connection will give a gain of 0.95 or an output pulse of 3.8 volts which is quite sufficient; the input capacity will be 10.2 ILILf. Cathode-follower connection would give a gain of 0.5 or an output of only 2 volts, so that additional amplification would be required; the input capacity would be 5.1 ILILf. If, instead, an amplitude of 2 volts were sufficienl but the generator gave negative pulses while a positive trigger were required, the cathode follower would again require an extra tube to invert the signal whereas the amplifier alone would suffice.

As an example involving a large capacity, suppose that the layout of complicated equipment forces one to send a small signal (say 0.1 volt or less) through a length of cable (say 5 feet having a capacity of 60 ILlLf) to another chassis where the input capacity is 10 ILILf. Suppose further that a rise time no greater than 0.05 ILsec. is required and that the polarity must be inverted. It might seem that a cathode follower preceded by an inverting amplifier would be required. If, however, gain is not needed, a tube may be saved by using an amplifier to feed the cable directly. Thus a 6AKS pentode (gm =5000, rp=! meg) with a 200-ohm plate resistor will feed the 70-ILlLf load at a rise time of 0.04 ILsec. with a gain of unity and an input capacity of 4 ILIL.

Thus it appears that there are many non-critical applications where a low gain amplifier is actually preferable to a cathode follower. While none of the material here presented is especially new, it is hoped that the discussion will be of help particularly to those who are not electronics experts but who must often design their own circuits for experimental equipment.

* \Vork done at Hrookhaven ::--Jational Laboratory under the auspices of the AEC.

1 For example. F. E. Terman Radio Engineers Handbook (McGraw-Hill Book Company, Inc., New York, 1943), pp. 430-432.

2 Note that a large coupling condenser can be avoided by feeding the plate voltage through the terminating resistor.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sun, 23 Nov 2014 17:11:46