I recently have built smiple Geiger counter based on cheap STS-5 tube, high voltage power supply from China and ATmega168 with LCD. Currently, I'm facing a problem with counting number of pulses from tube. I connected the output of simple Darlington amplifier to ext_int pin of ATmega168 and it sometimes does not count properly. I get 3-4 pulses instead of 1. I do not own scope or anything so I cannot measure output.
I also tried to collect pulses from tube cathode and adding additional filtering capacitor to HV supply output with same results. I do not suspect code to be a cause of incorrect counting, I checked it with basic TTL pulse generator and it works perfectly fine. Can anybody help me with this circuit? Also, my HV power supply is this module: Link
Electronic – Geiger counter pulse detector – too many pulses from amplifier
amplifiercountermicrocontroller
Related Solutions
Yes, lots of people I suspect.
However, using a microcontroller would be a simpler way. In a micro you can count pulses in a known period, or measure time between pulses. Counting pulses within a period is simpler, but can take a while to get a meaningful reading or you have low resolution. Most of the time you measure time between pulses and then invert to get speed.
Many micros have a capture capability where the value of a free running counter is captured in hardware when a pulse is received. By subtracting the previous capture value each pulse, you get a high resolution measurement of the period between pulses. I'd low pass filter that a bit, then invert when needed to update the user interface to show speed.
The 7 segment display encoding is also something the micro is well suited to do. It can then perform the multiplexing on the multiple digits, or you can get a micro with lots of I/O lines and hook up each segement individually.
You can solve this problem with just a couple of parts. TI makes logic parts in their 74AC logic family that operate correctly down to 1.5V. Use a 4 bit counter as shown below to repeatedly count from 8 to 15. Use the RCO output, which is a one clock wide pulse to reload the counter and to also drive an R/C circuit to produce the narrow pulse that you require.
If you desire to produce the pulse for every 16 input cycles then just connect the D input to GND instead of being pulled up to the VCC rail.
Best Answer
I'd clamp the base of Q1 to +5 and to GND. And bias the Darlington to almost-on, using the 4.7Meg and 1Meg suggested by the editor.
To clamp, simply hang diodes (installed to be reverse-biased) from the base to +5 (thus the diode "bar" to +5v), and from the base to Ground (thus the diode "bar" to the base.
Try it with considerable high resistors. E.g. 4.7 MΩ to +5V and 1 MΩ to GND.