1
$\begingroup$

Recently, I got a Geiger counter (si3bg). I applied a 400V voltage and a 1m Ω resistance load to it, and connected a 47K Ω resistor in series. The voltage at both ends of the resistance was measured with an oscilloscope. Then I got the following output waveforms on the oscilloscope, some of which made me very confused. It can be seen that several tens of microseconds after the start of a discharge, the voltage begins to drop, some of which almost drop to 0V, but then there is a second rise. It seems that two separate discharges have occurred in succession, or two particles have been injected into the counter tube only tens of microseconds apart. This has happened many times. Considering the dead time of the Geiger counter and the low sensitivity of this type of Geiger counter, I don't think it is possible that two particles have been injected into the counter only tens of microseconds apart. But what is the reason? This is very important, because I am counting it with a single-chip microcomputer, I use the rising edge trigger, which will affect the judgment of the single-chip microcomputer, and two events with short interval will greatly affect the output value.

Among them, the first, fifth and sixth pictures are what I call the confusing curve, and the others are relatively normal.

enter image description here

enter image description here

enter image description here

enter image description here

enter image description here

enter image description here

enter image description here

enter image description here

Improve this question
$\endgroup$
1
  • $\begingroup$ The source is am241 in the smoke sensing detector $\endgroup$
    – gailulun
    Commented Nov 28, 2020 at 13:31

1 Answer 1

Reset to default
1
$\begingroup$

A Geiger tube is an “avalanche detector,” because one ionization event turns into enough conduction electrons for there to be a signal. This is a fast, highly nonlinear process in the vapor in the tube, and the duration of the tube’s response is hard to predict. None of your waveforms is surprising to me. (But if your oscilloscope is displaying at 5V per division, so that full scale is forty or fifty volts, you may need to twiddle with how your voltage divider is arranged so that you don’t kill your microcomputer; those are usually low-voltage devices.)

The way to ignore the messy structure of your detector’s raw signal is to use a “discriminator.” This is an edge-detecting circuit which generates a logical signal (“I saw it!”) whose duration you can adjust so that it is slightly longer than the fire-oscillate-recover period of your detector, here a few hundred microseconds. If there are multiple rising-edge triggers while your discriminator’s output is in the “true” state, that may re-set its internal timer and it may stay “true” for longer than the minimum duration that you set, but it won’t re-trigger until after it’s gone “false” again. Your counting and timing systems operate on the leading edge of discriminator pulses, not on your raw signal. The duration of the discriminator pulse, when you are ignoring re-triggers, is a factor in the “dead time” of your detector system, and tuning it is something that every nuclear physics grad student gets to figure out how to do: too short a dead time and you get a bunch of noise triggers; too long and you miss real “pile-on” events; hopefully the happy place in the middle is wide enough to find.

A sufficiently smart microcomputer could perhaps do this discrimination against retriggers internally; the “discriminator” is a dedicated piece of analog hardware for this purpose.

Improve this answer
$\endgroup$
5
  • $\begingroup$ I input the signal directly into Arduino and convert the voltage to a value between 0 and 1023. I think I should add another condition in the trigger logic, that is, the next trigger should be more than 30us (or longer) from the last time the voltage value drops below a certain threshold. $\endgroup$
    – gailulun
    Commented Nov 28, 2020 at 17:00
  • $\begingroup$ Thank you for letting me avoid losing a single chip microcomputer, I didn't notice this before.😁 $\endgroup$
    – gailulun
    Commented Nov 28, 2020 at 17:04
  • $\begingroup$ Note that if you're doing much processing on the chip, that may cause your voltage sampling to occur less frequently, so sampling alone could miss brief events. Another advantage to an analog discriminator is that you can ensure that your gate signal is long enough to reliably trigger your clock-driven digital electronics. I haven't played with an Arduino in a while, so I can't comment of the specifics of the timing, but I think you should not expect a fresh digital-to-analog reading to reliably arrive with every clock signal. $\endgroup$
    – rob ♦
    Commented Nov 28, 2020 at 20:27
  • $\begingroup$ Yes, I noticed that the analogread function takes at least 20us of time, plus the time taken by other operations, which is a big risk of missing a signal, maybe I should use the device you're talking about. $\endgroup$
    – gailulun
    Commented Nov 29, 2020 at 10:55
  • $\begingroup$ If you want to count the particles directly by the Arduino, you definitely need to use the "external interrupts". $\endgroup$
    – Dennis Bobrowsky
    Commented Feb 19, 2023 at 14:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.