Counter Configuration
Two settings on the Config page, both under the Advanced section of the Input tab, that often get confused: Debounce and Tube dead time. They are related but serve different jobs and getting them right matters at high count rates.
Debounce
Debounce is an input filter in the firmware. It drops any pulse that arrives within the configured window of the previous pulse.
The job of debounce is to reject electrical artifacts: tube ringing, capacitive pickup, amplifier overshoot. These are not real radiation events and should not be counted.
Debounce applies to interrupt-counting builds. ESP32 PCNT builds use the PCNT Filter instead.
| Setting | Key | Type | Default | Builds | Description |
|---|---|---|---|---|---|
| Debounce | counter.debounce | Int 0-10000 | 200 | pulse builds without PCNT | Debounce window in microseconds. |
If your front end is clean (well filtered amplifier, short cable runs), 100 µs is usually safe. If you see suspiciously high background CPM that drops sharply with shielding, your debounce may be too low and you are counting noise.
If you suspect real pulses are being dropped at high rates (you flash a hot source and the count flattens out at a too-low value), debounce may be too high.
Tube dead time
Tube dead time is a physical property of the GM tube itself. After each ionisation event the tube needs to recover before it can fire again. During that window real radiation events are missed by the tube hardware, not by the firmware.
ESPGeiger applies a statistical upward correction to the displayed CPS to compensate for what the tube missed. The correction kicks in only above 50 CPS (3000 CPM); at typical background rates it has no observable effect.
| Setting | Key | Type | Default | Builds | Description |
|---|---|---|---|---|---|
| Tube dead time | counter.dead_time_us | Int 0-1000 | 100 (0 on test builds) | pulse builds | Tube dead time in microseconds. 0 disables the correction. |
Set this to the value from your tube’s datasheet. Datasheet figures for some common tubes:
| Tube | Dead time | Source |
|---|---|---|
| LND 7317 | 40 µs | LND 7317 |
| LND 712 | 90 µs | LND 712 |
| SBM-20 | 190 µs (minimum, at 400 V) | SBM-20 datasheet |
| Robotron 70 013 | 100 µs or less | VACUTEC 70 013 datasheet |
The J305 and M4011 datasheets give no dead time, so leave the default for those.
Setting 0 disables the correction entirely.
When debounce and dead time interact
Debounce is a filter (drops pulses in the firmware), dead time is a correction (statistical adjustment to the displayed rate). They are not the same thing and should not be merged.
The recommended rule of thumb:
- Set dead time from the tube datasheet. This is fixed physics.
- Set debounce as low as your front end allows without false counts. This depends on your hardware.
A reasonable setup has debounce <= dead_time + a small margin. If your debounce is much larger than your dead time, you are filtering pulses the tube did produce and the dead time correction will not recover them because it assumes the only thing lost was tube recovery.
If you need a large debounce to reject hardware noise, that is a sign the front end could be improved (better shielding, lower impedance, an output buffer on the tube amplifier).
References
- Tube datasheets - search for “
datasheet" and look for "dead time", "recovery time", or "resolving time". - Wikipedia: Dead time