Geiger counter (RadiationD-v1.1)

The RadiationD-v1.1 Geiger counter board with its J305 tube: VIN pulses low for each count, at random, as radiation does.

Designed by CAJOE: the long glass tube in two clips, a 555 that makes about 400 V for it, a speaker and an LED that click on each count, and a three-pin header. Aliases: Geiger counter, radiation detector, radiation sensor, CAJOE, J305, GM tube.

Pins

Pin	What it does
GND	Ground.
5V	Supply, 5 V. Under 4.5 V the model drives nothing.
VIN	The pulse output, whatever the silkscreen says. High at rest, low for each count.

On the board's schematic the output is the collector of a 2N3904 with a 47 k resistor up to 5 V, so it sits high through 47 k and is pulled hard to ground for each count. Count the falling edges on an interrupt. It is a 5 V signal: a 3.3 V board needs a divider.

Properties

dose is the starting dose rate in uSv/h (0 to 100); the slider on the part runs on a log scale from 0.01 to 100. factor is the tube's conversion factor in uSv/h per count a minute, 0.00812037 for the J305 by default.

What the model gets right

Random counts. Decays are random, so the counts are a Poisson process: each gap to the next count is drawn from an exponential distribution, from a seeded generator, so a run repeats exactly. One minute's count differs from the next by about its own square root, which is why every Geiger sketch averages over a minute.

The rate. Counts a minute are the dose rate divided by factor: 0.00812037 uSv/h per CPM, or 123 CPM for 1 uSv/h, the J305 figure the board's guides use (the ESPHome Geiger counter page, from opengeiger.de's notes). Background, 0.1 to 0.2 uSv/h, is 12 to 25 CPM; the default 0.15 is 18.5.

Dead time. A count that would land while the output is still low from the last one is lost, as it is in a real counter at high rates.

No timer with no dose. At 0 uSv/h, or unpowered, the part schedules nothing.

What it does not model

The 400 V supply: the part makes counts, not high voltage, and the R100 trimmer that sets the voltage on the real board does nothing here. The pulse width is not on the schematic and no published measurement was found; the model holds the output low for 100 µs, an assumption on the order of a small tube's dead time. The conversion factor is approximate: no maker's datasheet for the J305 exists, other guides use 0.0057 (151 CPM per uSv/h, the M4011's figure) or 0.00332, and the true figure depends on the tube and on the energy of the radiation. The tube's own background counts beyond the dose you set, its plateau, its slower recovery after a count, the click and the LED are not modeled.

Common mistakes

Wiring VIN to 5 V. It is the output, not a supply input; the silkscreen is wrong.

Showing a count over a few seconds as CPM. At background a five-second count is zero or one; scale it to a minute and it jumps between 0 and 12. Keep a running minute.

Trusting the uSv/h to two decimals. The factor is approximate by tens of percent.

See it in action

Geiger counter counts on an interrupt and shows CPM and uSv/h on an I2C LCD. Open it at /templates.