SU756331A1 - Method of measuring amplitude density distribution of signal stream - Google Patents

Description

The invention relates to the field of experimental nuclear physics, methods of measuring the density distribution of the amplitude of the signal flow, and, in particular, to methods of measuring the density distribution of the amplitudes of pulsed generators designed to measure differential nonlinearity '(DNL) instrumentation, nuclear instrumentation. ’

There is a method of measuring DNL instruments of nuclear instrumentation using a multichannel amplitude analyzer and a pulse generator-1 owl with a constant amplitude distribution density [1]. For implementing the method of a multichannel amplitude analyzer, two measurements are made, one of which reflects the total DNL generator, analyzer and the equipment under test, and the other is the total DNL of the generator and analyzer. Next, the difference between two measured differential nonlinearities (as a discrete function of the analyzer channel number) is calculated, which is taken to be equal to the DNL of the equipment under test.

2

The disadvantage of this method is the impossibility of its use for measuring the density distribution of the amplitude of the output signals of the generators and, in charter, of the generators intended for measuring the DNL equipment.

There is also known a method for measuring the density distribution of the amplitude of a signal stream by measuring the amplitude of individual signals by a multichannel amplitude analyzer with the subsequent calculation of the characteristics of the distribution density from samples in the analyzer channels [2).

A disadvantage of the known method ι is that the analyzer has a channel width inhomogeneity (differential non-linearity), which when measuring, for example, the uniform distribution of amplitudes of signal flow amplitudes leads to a distortion of the measurement results and an incorrect assessment of the measured (initial) density. Thus, the measure of the distortion introduced by the analyzer is its differential nonlinearity (DNL).

The aim of the invention is to improve the accuracy of measuring the density

3

756331

four

amplitude distribution of the signal flow. To achieve the specified goal, the measured signal is summed with a stable amplitude signal synchronized with the measured one, then the total amplitudes of the mentioned signals are measured successively in time, and each subsequent measurement is made at a new value of the stable amplitude signals, and then the group of samples is selected. in any measurement, and calculate the characteristics of the density distribution.

FIG. 1 shows a block diagram of the inclusion of instruments for measuring; in fig. Fig. 2a shows an exemplary image of the amplitude distribution density of signals on the oscilloscope screen of the analyzer for various values of signals of stable amplitude (Fig. 2.6). To simplify the graph, the measured distribution density, which is a discrete function of the analyzer channel number (axis X of Fig. 2), is depicted as a continuous function. The shaded areas correspond to the selected section of the analyzer scale (inside the section the channels are numbered as 1, 2, ..... K) in different

measurements (1, 2, ... ,,) corresponding to different amplitudes of the signals of the auxiliary generator

<p ,, and _a ....., and _e ).

The described method is as follows. In the analog and usually linear adder 5, the measured signal coming from source 1 is summed with a signal of stable amplitude generated by source 2 at times determined by synchronization circuit 4. The synchronization determines how many and which of the flow pulses will be measured by a multi-amplitude analyzer 3. The measurement of the density distribution of the amplitudes of the signal flow at the output of the adder 5 is carried out for some time (exposure time or set). During this time, the analyzer measures the amplitudes of the individual pulses of the stream, which is modified by the adder. The second and subsequent measurements are carried out with the amplitude values of the source 2 signal lying in the range of the minimum amplitude measured by the analog’3 and recorded in its storage device (memory). Since the capacity of the memory can be limited, after η measurements (usually η = 1), the information on the memory is output on a digital printout, another device or, much more convenient for subsequent calculations,

transmit to the memory of the computer.

Consider a special case of measuring the density distribution of the amplitudes of the pulses of the generator with the subsequent evaluation of the density non-uniformity (differential nonlinearity) and show that the proposed method allows to reduce the measurement error. . .

At the first measurement, we establish the amplitude of the signals of a stable amplitude, equal to zero. The measured amplitude distribution of the amplitudes of the signal flow of the generator (or, in abbreviated form) the measured density, distorted by the adder 5 and the analyzer 3, is recorded in a memory device. The measurement time should be sufficient for a set of several thousand samples in each channel of the analyzer. During the second and subsequent measurements, the exposure should not change. In the second measurement, the value of the signal of a stable amplitude is set to a value other than Zero, in the third one it is set to a value other than the second, and so on. The order of change in the amplitude of source 2 is arbitrary, as well as the number of measurements (after the second).

Next, select the area on the scale of the analyzer that contains the samples, in any measurement, and calculate the characteristics of the measured amplitude distribution density. The position of the plot on the analyzer scale depends on the ratio of the maximum output amplitude of source 1 and the maximum amplitude measured by the analyzer, 3. It is obvious that if the maximum amplitude of the source is greater than that measured by the analyzer, then at the input of the adder it is barely. blowing install the divider and achieve the most convenient case when the amplitude is of the same order. Then select the area close to the end of the scale (0.9 - 0.95).

To calculate such characteristics as the test generator DNL, it is advisable to use a computer, since with a selected section of the analyzer scale containing K channels, and with the full number of measurements, 2 is necessary. we calculate.

'-2 ~ -distance of the number of samples in each channel of the selected area in all measurements and compare these differences for finding the maximum. For example, when K = 100 channels and ζ. = 10 ς = 4500. If processing of one difference in the number of samples (calculation, comparison with others, etc.) takes about 300 µs, which corresponds to the speed of modern computers, then for the above example, the calculation of DNL will require about 1.5 seconds of machine of time. Obviously, when calculating manually, it is thousands of times more.

five

756331 '

6

When using a computer, you can offer the following algorithm for calculating

DAYS of the tested generator:

1. Averaged over the selected section of the analyzer scale and over all measurements, the number of samples in the channel.

- ι to I ³

(Νι ₄ is the number of samples in channel 1 in the 5th dimension);

2. The maximum difference in the number of ιθ accounts in the same channel (or group of channels) of the selected segment

in any two dimensions, gpach ^and so ^ y \ gpak)

X - 1, 2, ι. ·., K

. 'measurements of the DNL generator with a significantly larger dynamic range than the dynamic range of the analyzer used, since, by summing up with the signals of the auxiliary generator, the analyzer can measure the amplitude of the signals of the tested generator that is less than the sensitivity threshold ^ of the analyzer.

Claims (2)

Claim where: '. . Δ '<Γηαχ = τήαχ | Ν · _ι6 - ZD, ‘, C B 7 1 3. DNL of the tested generator ΛΗΝ ' ¹ А00%. To eliminate the effect ¹ 'even-odd' ¹ , inherent only to the analyzer, before the above calculations, it is advisable for each measurement to smooth the samples in the channels of the selected area, which is to replace the received samples with the values: 15 20 25 Where '4,1, .-, to ¹ thirty a ^ (w) weights determined by the type of smoothing function. The number of smoothing points (2m + 1) is chosen small (for example, 5%) in order not to distort the local generator DNL. The advantage of the described method is that the measurement result of DNL generators does not depend on the DNL analyzers, and, thus, the measurement error decreases sharply (to an error from the statistics of readings). Another advantage of the proposed method is the ability 35 40 45 A method for measuring the amplitude distribution of the signal flow by measuring the amplitude of individual signals with a multichannel amplitude analyzer and then calculating the characteristics of the density of the samples in the analyzer channels, characterized in that, in order to improve measurement accuracy, the measured signal is summed with a stable amplitude signal synchronized with the measured signal , then the total amplitudes of the above-mentioned signals are measured successively in time, each subsequent and The measurement is made at a new signal value of a stable amplitude, and then the corpse of the analyzer channels is selected, in which the samples are counted in any measurement and the characteristics of the density distribution are calculated.