DE1472112A1 - Method for stabilizing the amplification of pulse spectrometers, in particular scintillation spectrometers - Google Patents

Description

PATENT Attorney E.

DR. ING. E. LIEBAU Au ^ sburj-Göggingen, H.0.I909

DIPL. ING. G. LI E BAU

• M2 AUGSBUR & GÖGGINGEN 1472112

vBCHeNooiw-STYLE ·.! ».» «" β A ^IHUIU

Dr Lb/R Au G 57o2

Society for the Promotion of Research

at the Swiss Federal Institute of Technology. Zurich, Switzerland)

Method of stabilizing the amplification of pulse spectrometers, especially scintillation spectrometry ^ n

The present invention relates to a method of stabilization the amplification of pulse spectrometers, especially scintillation spectrometers, for radiation of nuclear physics Origin.

Several methods of this type are already known} they can be divided into two large groups, namely:

a) according to the type of reference,

b) the location where the correction signal used for stabilization is obtained.

It is clear that for the stabilization of the entire chain of equipment a scintillation spectrometer is included Scintillator only one reference source can be used whose rays are of the same type as those to be examined. Furthermore is the place where the correction signal is obtained the output of the pulse analyzer. Some well-known Methods partially meet these requirements to a high degree.

All methods that use analogue circuits, i.e. apparatus that convert one type of signal into another

900851/0497

undertake, e.g. number of pulses in voltage, and which compare the converted signal with a normal, suffer from the fact, that all conversion methods are dependent on environmental conditions. There is therefore only one way in which only the same Large numbers can be compared with one another, i.e. pulse numbers with pulse numbers, voltages with voltages, etc. Also this path has been walked a long way, but not to the end.

The present method has been developed for the stabilization of pulse spectrometers, especially for γ-spectroscopy.

As with already known methods, two pulse numbers are compared with one another. It is based on a zero adjustment, so that the accuracy is essentially due to that of the reference standard, the sensitivity due to the amplification of the comparison system is given. The method according to the invention consists essentially in the fact that a reference source radiation is periodically generated by means of a chopper device During certain time intervals the source radiation is superimposed on that originating from both radiations Pulses are analyzed and registered to obtain a first pulse number value that is practically the same length during that time Time intervals that are shifted compared to the first, so that the reference source radiation is blocked in them, the impulses originating from the measurement source radiation are analyzed and registered in the same way to obtain a second pulse number value, and that the second pulse number value is periodically subtracted from the first pulse number value to obtain a difference variable, the temporal mean value of which over the entire period is a controlled variable for compensation of gain variations.

The two numerical values to be subtracted from each other can be obtained, for example, as follows:

a) The first numerical value is taken from the electrical output pulses a photomultiplier scintillation

009851/0497

3 H72112

System won, its scintillator (NaI) permanent with the measurement source radiation I and during the former Time intervals additionally with a reference

1 Ύ1
renzpräparat (e.g. Cs ^ ') is irradiated. The number of pulses that _{exceed a voltage threshold U Q} form the first numerical value that is dependent on the gain / U. _{The voltage threshold U Q} is preferably selected to be equal to the pulse height Ep at which the photopeak of the reference source radiation lies in the pulse spectrum of the non-stabilized system.

b) The second numerical value is the sum of the number of Hessquelle radiation in the manner described derived pulses alone and the number of gain-independent standard pulses generated during of the second-mentioned time intervals are supplied by a crystal-stabilized oscillator.

The reference pulse source is a γ source that is periodically interrupted by a mechanical chopper with a rotor diaphragm, for example. The rotor diaphragm has an opening ratio = ■ £ -. Synchronous, i.e. in fixed phase relation to the chopper function, an optical system controlled via the rotor axis generates electrical Control signals.

With the period T, a group of standard pulses is created in each case during the time intervals ΔΤ and a group of reference pulses during the time intervals ΔΤ *, which, however, are around

■ κ are shifted compared to the former. In addition to the pulses discussed, there are also the X-pulses originating from the source of the γ-rays to be examined, which are superimposed on both the standard pulses and the reference pulses for equally long time intervals ΔΤ.

All pulses are read into a memory, whereby the pulses decrease in the successive time intervals

909851 / 0A97

are counted alternately positive and negative. Thus, on the one hand, the reference _{pulses plus the I-pulses exceeding the threshold U Q} and, on the other hand, the standard pulses plus the! -Pulses exceeding _{the threshold U 0, the standard pulses plus the!} If the intensity of the! Preparation does not change noticeably within £, the influence of the measurement source intensity is eliminated by the subtraction process. The time average of the memory content is therefore a hatred for a gain drift.

Both reference pulse groups and the standard pulse groups run through one and the same chain of equipment, which is why no part of this chain has high requirements with regard to gain constants must be asked.

IAn away from the spectrum of the! Source that of the reference source to hold, the counting is suppressed by a no gate during the time intervals AT.

Further details of the example described above and further examples of the method according to the invention are explained in the following description with reference to the drawings.

Fig. 1 gives a schematic overview of a possible overall arrangement of the apparatus of a single-channel Y-spectrometer.

FIGS. 2a to 2g show the sequence of events synchronously various functions as well as the control signals, the measurement and reference pulses

Fig. 3a shows the pulse spectrum of the reference source and Fig. 3b the pulse spectrum of the Hess source Z.

FIG. 4 explains the time sequence of method steps of an exemplary embodiment of the method according to the invention in use on multichannel analyzers.

5 is a block diagram of a multichannel spectrometer gain-stabilized according to this method.

909851/0497

The well-known single-channel spectrometer arrangement with which the rays of the source Z are to be analyzed «ground, shows, as in Flg. 1, a scintillation crystal 2 lands a photomultiplier 3 with its high voltage source

4 on. The output pulses of the photomultiplier are in a subsequent amplifier 5 enlarged, shaped and sorted by a measuring discriminator 6 according to their size. The sorted out pulses are brought to a uniform size in the same apparatus and by a pulse counter 8 registered.

In the case of the gain-stabilized spectrometer, this is Keep the basic arrangement except for a no gate 7 »which is switched on between the Kessdlkriminator 6 and the pulse counter 8 is.

Further essential additions are the chopper having the rotor diaphragm 9, the reference pulse source 10, and a chain of apparatuses 11 to 15, a yes gate 16 with the standard oscillator 17, one coupled to the chopper rotor 9 by a shaft 18 Drive motor 19 and the optical control system 20 to 22. The apparatus chain 11 to 15 is at the output of the pulse amplifier

5 connected and has a reference discriminator 11 Yes gate 12, a diode pump 13 »one by electrical signals controllable, electromechanical pole changer 14 and an integrator 15 with servo amplifier. The optical control system 20 to 22 consists of a rotary shutter 20 coupled to the shaft 18, two light sources 21 and two associated photocells 22nd

All boxes shown in bold in Fig. 1 form the electronic part of the actual stabilization device combined in one device, while the others Box represent commercially available components.

In Fig. 2a the time sequence and size of the electrical Pulses shown at the output of the photomultiplier 3, as indicated by the scintillations of the X source rays in the crystal 2 can be effected.

909851/0497

The chopper rotor 9 t has its opening characteristics in Flg. 2b, periodically causes the rays of the reference source 10 to act on the crystal 2, so that the! Pulses within the chopper opening times ΔΤ * reference pulses are superimposed. Fig. 2c shows the resulting pulse train.

The optical control system 20 to 22 generates the control signals I and II, the timing of which is shown in FIGS 2e are shown. The control signals follow one another with the period

T and each have the duration ΔΤ. The control signals II are opposed

τ
over the control signals I shifted by «.

Activate the control signals I of duration ΔΤ, within which a group of reference pulses appears:

- the yes gate 12 to the reference and the! pulses, which exceed threshold U of reference discriminator 11, input into the integrator 15;

- The electromagnetic pole changer 14 to the same on Switch to the addition position at the end of time interval I;

- the no gate 7 to suppress counting of the pulses in the pulse counter 8.

T The control signals II of the same duration ΔΤ, but by -s

▼ pushed, press:

- the yes gates 16 and 12 to a group of standard pulses and the threshold U _Q in the reference discriminator 11

. to enter the exceeding! pulses into the integrator 15;

the no gate 7 to suppress counting of the pulses in the pulse counter 8;

- The electromagnetic pole changer 14 to the same on Switch to the subtraction position at the end of time interval II.

The discriminator 11 emits uniform pulses that through the yes gate 12 in groups according to FIG. 2f to the entrance

• 06861/0407

reach the diode pump 13, which in turn supplies positive or negative charges to the integrator 15, depending on whether the pole of the 14 in position is addition or subtraction (FIG. 2g). The integrator 15 forms an analog signal which is proportional to the arithmetic sum of the charged charges, ie also proportional to N ₁ -N ₁₁ , where Nj is the number of pulses during the time interval I of the duration ΔΤ and Njj is the number of pulses during of the subsequent time interval II of the same duration ΔΤ. The pulse numbers Nj and Njj can be expressed as follows:

Nj = ν (/ u). ΔΤ + V _x . ΔΤ () ΔΤ,

where υ » (/ u) is the frequency of the reference pulses that is dependent on the gain / U, Vj is the frequency of the Z source pulses _{that is also dependent on the% gain and v ^ T is} the frequency of the standard pulses that is independent of the gain. If the operation Nj - Njj is carried out, the result is

). ΔΤ »- v £ _T. ΔΤ.

It can be seen that the mean influence of the Z source pulses is eliminated is, provided the intensity of the! source and thus the frequency v4 does not change noticeably in the time interval between the reference pulse groups and the subsequent standard pulses The direct voltage signal formed by the integrator is used as a corrective feedback signal for the sake of simplicity fed to the cathode of the photomultiplier. The so educated The feedback circuit is illustrated in Fig. 1 by the dash-dotted line. Thus, there will be any gain variations proportional within the spectrometer

Λ Τ '

to the pulse number difference Nj - Nj ₁ s & ^ "" V (/ ^u ) ~ ^ s * reduced.

Contrary to known circuits in which two controlled diode pumps and a preceding one are in front of the integrator

909851/0497

H72112

Phase reversal stage are used, there is the one described Arrangement only a single diode pump 13 bits of the following electromagnetic pole changer 14 »which the group wise occurring current pulse according to the control signals Z and II feeds the integrator 15 with a positive or negative sign. This circuit allows the reference as well as to feed the standard pulses to the integrator / without exception via the same chain of apparatuses 5 and 11-14, with all of them Subject to parameter drifts of some elements of the feedback and are thereby corrected.

As a further embodiment, it should be mentioned how the standard frequency can be obtained from the reference source 10 itself. The greatest amplification sensitivity of the reference pulse frequency ν (/ u) occurs when the voltage threshold U _{0 of} the reference discriminator 11 is selected to be equal to E _p , ie the value of the photo peak. This fact has already been used in the first example. However, the lowest gain sensitivity of the reference pulse frequency ν (/ u) occurs when U = Ey, where Ey is understood to be the pulse amplitude value of the valley between the Photo and Compton peak. This last circumstance, not used in the first example, allows the reference source 10 to also be used as a standard pulse source when the discriminator voltage threshold U _{Q is} at Ey. Periodic switching of the discriminator voltage threshold U _Q from E _p to Ey thus allows the reference pulse frequency ν (/ u) to be obtained for U _Q = E _p and the standard pulse frequency V £ _{T for U Q} β Ey.

While in the example with a standard oscillator, a comparison cycle comprises two time intervals ΔΤ within which the one NaI while ΔΤ 'the reference source spectrum and the other Hai while ΔΤ superimposes the standard pulse spectrum on the measurement spectrum a comparison cycle of this second example comprises four time intervals ΔΤ in which the four pulse

909861/0497

H72112

counting

N ₁ «V (, u) T '

ΔΤ ¹

Y ^ p. ΔΤ ν ^ ρ. ΔΤ " ⁷ IV · ΔΤ YV ₇ . ΔΤ

be registered. V ^ p means the frequency of the _{! Pulses discriminated at Fhotopeak at U 0} »Ep, V ^ ₇ the frequency of the! Pulses discriminated in the valley between FhotopeaJc and the Compton peak at U ₀ = E ₇ and V ₇ the frequency of the reference pulses also discriminated in the valley mentioned

The pulse frequency Y ₇ is now the almost gain-independent standard pulse frequency.

If the scintillation crystal has a reasonably good resolution, Y ₇ »s 2 ν (/ ^u _o )» venn / U _{0 means} the gain value at which the discriminator threshold U halves the photo peak.

By carrying out the arithmetic operation 2 (Νγ-Njj) - (N ₁₁₁ -N ₁₇ in each cycle, numbers are obtained whose mean value over time represents a hatred for a strengthening variation.

The two examples presented are particularly suitable for spectrometers with single-channel analyzers. To the first case It should also be mentioned that it can also be used for pulse spectra that do not show a photopeak. The second However, example makes explicit use of the presence of a photopeak.

If a spectrometer is equipped with a Tielkanalanalysator and the radiation detector responds with at least partially uniform signal amplitude to monoenergetic radiation (photopeak), the stabilization method is advantageous modified as is explained in the third example described below and with reference to FIGS. 3, 4 and 5.

While the first two examples placed a certain emphasis on the fact that the gain-dependent re-

909851/0497

. Jp Η72ΊΊ2

f.erent pulse frequency ν- (λχ) was compared with a completely or almost gain-independent _{pulse frequency Vg T} or vy "the presence of a photopeak is now used in a way as is done in various commercial spectrum stabilizers. This method described in an article by H. de Yaard in the journal "Nucleonics", Volume 13 (July 1955), page 36, is shown in FIG. 3a on the basis of the pulse spectrum of the reference source. The photopeaJc is halved by two windows 1 and 2 into two spectral ranges, namely from U _Q - Δ17 to U _Q and from U _Q to U ₀ + AU. The two oppositely gain-dependent pulse frequencies from these areas are called v ^ and V ^ ₂ . The task of the stabilizer is now _{to keep the pulse frequency difference V ^ 1} - V ^ ₂ in the mean zero. A device that works in this way stabilizes perfectly; however, the difficulties arise as soon as this so-called spectrum stabilizer is still to be used for measuring unknown spectra of the measurement source Z, as FIG. 3b shows one. Only if the measurement spectrum runs horizontally in the spectral range of stabilization from U ₀ - AU to U ₀ + AU is a set gain value independent of the addition of the measurement preparation Z. However, this requirement can hardly be met, for example, in the activation analysis.

The special feature of this third example is the way in which the influence of the measurement source on the stabilization is switched off i.e. how the known spectrum stabilizer is made into a gain stabilizer.

Analogous to the first two examples, the interference of the measurement and reference spectrum is switched off by a pulse number subtraction process, but now the fact that the two spectral ranges from U ₀ - AU to U _Q and from U _Q to U _Q + AU of the window are used 1 and can now be viewed at the same time.

909Θ51 / 0497

ti

In principle, during a first time interval of duration ΔΤ, the registration of the Anmahl + (N _R1 + Nj ₁ ) (Nj ₁₂ + Mj ₂ ) pulses from the windows 1 and 2 originating from the reference and the Z source and, by if later, takes place , with the reference source closed during a second time interval of the same duration ΔΤ the sole registration of the number of pulses from the Z source with the opposite sign, di - Nj ₁ + Nj ₂ * The index R relates to the reference source, the index Z to the measurement source, index 1 on window 1 and index 2 on window 2.

However, since the intensity of the measurement source Z varies over time, it is advantageous not to use the one just described To go method path, but rather the one shown in FIG. A comparison cycle comprises four of the same length Time intervals I, II, III and IV that repeat with the period T. The time intervals I and IV of the duration ΔΤ, within of which during ΔΤ * of the measurement source radiation Z the references source radiation is superimposed, lie apart by £. the Time intervals II and III of the same duration ΔΤ, within which only the measurement source radiation Z is registered, are phase-shifted with respect to the intervals I and IV, namely the interval II negative compared to the interval I, and interval III is positive compared to interval IV. The pulse numbers registered in time intervals I to IV are then:

interval Window 1 Window 2 I. H _R1 ♦ N ₁₁ N _R2 + N ₁₂ II ^N I1 ^N Z2 III ^N Z1 N ₁₂ IV N _R1 + N ₁₁ N _R2 ♦ N ₁₂

The pulses falling in the window 1 are corresponding to the sequence of the time intervals I to IV with the + + sign

909851/0497

Λ U72112

and the impulses falling in window 2 are registered with the goal oaks - + + -. With this gate string for registering the pulse numbers, it is achieved that both the X source pulses of the number N ₁₁ falling in window 1 and the X source pulses of the number N ₂₂ falling in window 2 are compensated for, even if the Nessource intensity and thus the pulse numbers Nj ₁ and Nj ₂ vary linearly over a period of time. The mean value of the registration thus only corresponds to the difference N _R1 -N _R2 , ie the number of reference pulses in window 1 minus the number of reference pulses in window 2. This difference can be positive or negative in the case of strength variations and is an exact measure for the strength variation. If a correction signal proportional to this is obtained from the difference mentioned and counteracts a change in gain, the gain is stabilized until the difference between the pulse numbers N _R1 and Np _{2 is} zero.

The method just described can be carried out with the apparatus arrangement shown schematically in FIG.

The usual arrangement of a non-stabilized γ-spectrometer, which is to analyze the radiation of the Z-source 1, includes the detector, consisting of the scintillation crystal 2 and the photomultiplier 3. The signals from the detector are fed to a Tielkanalanalysator whose subunits are ι the high voltage source 4, the pulse amplifier 5 $ an analog to digital converter (ADC) 6a followed by an address scaler 6b, the output signals of which are fed to the memory 8 via the no gate 7 required for the stabilization method.

The mechanical part of the stabilizer, the chopper, is designed similarly to the first example and includes the Rotor diaphragm 9 »the reference source 10 and the motor 19, the the optical rotary screen rigidly connected to the rotor screen 9

909851/0497

20 drives. The latter is generated together with the light source 21 and the photocell 22, the gate control signals necessary for the functioning of the stabilizer according to the time intervals I to IY as also shown in FIG.

The electronic part of the stabilizer is essentially designed as follows:

A control unit 23 receives the gate control signals corresponding to the time intervals I to 17 from the photocell 22 and derives from it the unit 23 shown below in FIG Control signal sequence from which the no-gate 7 of the analyzer s, a yes gate 28 and a command inverter 29 of the stabilizer, which come from the ADC 6a of the Anaitysators Pulse trains are provided by an address scaler 24 and two numeric Units 25 and 26, which represent windows 1 and 2, analyzed and via a delay element 27 through the yes gate 28 opened during each interval I to IV Computer 30 supplied. This counts the pulses from unit 25 (window 1) positive, from unit 26 (window 2) negative during the time intervals I and IV and, as a result of the command inverter 29, with a correspondingly reversed sign during intervals II and III. The computer 30 is followed by a digital-to-analog converter (DAC) 31, the output signal of which is im present example of the cathode of the photomultiplier 3 is supplied and the desired stabilization of the gain of the spectrometer.

909851/0497

Claims (12)

U72112 claims 1. Procedure for stabilizing the amplification of pulse spectrometers, in particular scintillation spectrometers, characterized in that by means of a chopper device a reference source radiation is periodically superimposed on the measurement source radiation during certain time intervals that the impulses originating from both radiations are analyzed and registered in order to obtain a first one Pulse number value that during practically equally long time intervals that are shifted compared to the first, so that the reference source radiation is blocked in the same, the pulses originating from the measurement source radiation can be analyzed and registered in the same way to obtain a second pulse number value, and that the second Pulse number value is penodically subtracted from the first pulse number value to obtain a difference value whose temporal mean value over the entire period is a controlled variable to compensate for gain variations. 2. The method according to claim 1, characterized in that the occurrence of the second-mentioned time intervals is ^{controlled by the ü} hopper device. 3. The method according to claims 1 and 2, characterized in that in the second-mentioned time intervals of the impulses originating from the measuring source radiation are superimposed on gain-independent standard impulses which together deliver the second pulse number value with the pulses of the measuring source radiation. 909851/0497 Document «ι v * L. * ***. * wr. l sentence 3 d «s Ämtorungsoes. w. 4.9.1967) H72112 4. The method according to claim 3t, characterized in that the standard pulses are obtained from an oscillator and controlled by the chopper device Yes gate can be superimposed on the impulses originating from the measurement source radiation. 5. The method according to claims 1 to 4 _t characterized in that the deriving pulses from the reference source radiation and the standard pulses of a pulse counter of the spectrometer through a no-gate which is controlled in dependence on the chopper device are kept. 6. The method according to claims 1 to 5 »characterized in that that the difference size by alternately positive and negative counting in the time intervals occurring pulse number values is obtained in a counter. 7. The method according to claims 1 and 2, characterized in that that during the first-mentioned time intervals the reference source radiation except for the generation of reference source pulses is also used to generate standard pulses obtained by discriminating the reference source pulses at a point with little gain dependency and that the first pulse number value from both the standard pulses, the reference source pulses and the measurement source pulses is obtained. 8. The method according to claim 7t, characterized in that the pulses are obtained by means of a radiation detector, the response to monoenergetic radiation with at least partially uniform signal amplitude due to the appearance of a peak responds in the spectrum, and that in the discrimination of the reference source radiation the entire peak is recorded. 909851/0497 H72112 9. The method according to claim 1, characterized in that the pulses are obtained by means of a radiation detector that responds to monoenergetic radiation with at least partially uniform signal amplitude by the appearance of a peak in the spectrum that responds by discrimination of the Pulses by means of two preferably adjoining windows, the peak is divided in its width, whereby during every time interval a pulse number value is obtained from each window, of which pulse number values half of the measurement source radiation and the other half of the measurement and reference source radiation are supplied together, that by subtracting the second pulse number value coming from one window from that from the same Window originating first pulse number value, a first difference variable is obtained, while by subtraction of the second pulse number value coming from the other window from the first one coming from the same window Pulse number value a second difference variable is obtained, and that by subtracting the one difference variable from the on the other hand, a variable is formed, the temporal mean value of which over the entire penode duration is the controlled variable for compensation of gain variations. 10. The method according to claim 9, characterized in that that in a cycle of four consecutive time intervals (I to IV) for the first pair of intervals (I, II) the shift between the interval with the reference source radiation and the one without reference source radiation is made negative and the other pair of intervals (III, IV) is made positive. 11. The method according to claim 10, characterized in that the negative and the positive shift of the time intervals (II and III) has the same amount (v). 90985 1/0/197 - 12. The method according to claims 9 to 11, characterized in that during the time intervals (I and IV) ₁ in which the reference source radiation is superimposed on the Hessource radiation, the pulses are kept away from a memory (8) by a no gate (7) . 909851 / CH97