ESP32 Powers Fresh Take On An IoT Geiger Counter

Over the years we’ve covered many projects aimed at detecting elevated radiation levels, and a fair number of them have been Internet connected in some way. But as they are often built around the Soviet-era SBM-20 Geiger–Müller tube, these devices have generally adhered to a fairly conservative design. With the current situation in Europe heightening concerns over potential radiation exposure, [g3gg0] thought it was a good a time as any to revisit the idea of an Internet-connected Geiger counter using more modern components.

Now to be clear, even this modernized approach still makes use of that same SBM-20 tube. There’s such an incredible wealth of information floating around out there about how to work with them that you’d almost put yourself at a disadvantage to chose something else to base your design on. Put simply, it’s hard to go wrong with a classic.

An unfortunate bug was discovered in the HV circuit.

That said, [g3gg0] decided early on that the design would use as many SMD components as possible, a considerable departure from many of the SBM-20 counters we’ve seen. That meant coming up with a new high-voltage power supply capable of providing the tube with the necessary 400 V, which from the sound of things, took a few attempts to complete. The final result is perhaps the smallest and cleanest looking board we’ve ever seen play host to this particular tube.

To run the show, [g3gg0] selected the ESP32-PICO-D4. You certainly don’t need such a powerful microcontroller to read the impulses from the SBM-20 tube and publish them via MQTT, but to be fair, the chip has a number of other duties. It’s handling the WS2812 RGB LEDs that go off in response to detected particles, running the (apparently optional) 2.9 inch WaveShare electronic paper display, and also pulling data from a BME280 environmental sensor as well as a CCS811 VOC sensor — so it’s keeping fairly busy.

As impressive as this build is, we do hate that it had to be built. From certain world leaders dropping casual comments about the strength of their nuclear arsenal to foolhardy attempts to capture the Chernobyl power station, having access to a reliable Geiger counter isn’t an unreasonable precaution right now. For everyone’s sake, let’s hope the fancy RGB LEDs on this particular build remain as dark as possible.

17 thoughts on “ESP32 Powers Fresh Take On An IoT Geiger Counter

  1. These Russian based “Tubes” are decades old and no one takes care to the degradation of the “quenching gas” inside, causing many double and false counting of radiation levels.

    All these based “Geiger Counters” are counterfit with no relation to real Rad´s.

    1. You wrong. I got hand made analyzer on same tube, made in 1980s. No false counting. It can jump for just a bit. But it’s not exactly super accurate and can be used only for domestic purposes. Anyway, you can filter all false with usual average from 5 measurements, in code. Extra simple.

  2. This is indeed a serious issue.
    I have ordered three Z1A tubes from
    Well, pollin isn’t particularily known for their high quality products here in germany, but it’s a russian glass geiger tube, what could go wrong?

    Ordered three of them. Two don’t detect any radiation and one of them is so insensitive, that it would’t even tell if you are dying from a deadly dose of gamma radiation.

    Both SBM-20 measure the same level of radiation and it maches the expected background radiation.
    At least from a tinkerer’s view they look fine.

  3. The unstable high voltage may be caused by R3. It looks like maybe an 0603 and is probably rated for about 150V. 400 volts accross a 150V resistor will likely cause problems.

    1. Yeah, wow, 400V across something that’s only 1.6mm long if you’re lucky (1.5 if you’re not) is a pretty hefty E field gradient. When you mess around with say 10 or 20kV you find that ten 1M resistors in series are a lot more reliable than a single 10M resistor – bc it stretches out the voltage drop and reduces the volts/meter. Same thing would apply here in miniature. Even just two resistors in series would mitigate the problem.

  4. Yeah thats also my experience.
    It drops in resistance and the voltage divider changes.
    I guess in that case it should have no further effect than a shifted divider.

    Thanks for the hint, I probably should look closer at the resistors too.

  5. Like with the bargain-basement radiation meters available from ebay (or Alibaba/Taobao): they are only ever as good as their calibration. If it’s not calibrated with a known source, it’s more of a novelty toy than a useful measurement of radiation.
    The actual raw count rates between “normal background” and “short/medium term hazard but not immediate seconds-of-exposure fatality risk”) are so small that even small offsets can swing you from a meter that tells you your bananas or brazil-nuts are about to kill you (but makes for good tiktok videos of wandering about the Fukushima evacuation zone and pretending it’s scary-scary as your cheap meter beeps away), or one that fails to note the proverbial truckload of Cobalt-contaminated rebar driving past.

    1. It doesn’t need calibrated, it’s fundamentally accurate in what it was designed to do. But note, it wasn’t designed to access radiation dose for personal safety, in orderfor that to occur you need to know the energy of each detection, it has no way to differentiate from say high powered x-rays and low powered x-rays since the energy of an emitted x-ray photon is dependent on how much energy was lost from electron deflection.

  6. Just some comments on the comments above.
    I’ve bought and sold 100’s of SBM-20 tubes over the years. I test them all and found very few (>~2%) that create multiple events due to the loss of quenching gas. However, sellers of “untested” tubes have a higher rate. Unfortunately, almost all
    SBM-20 tubes come from the Ukraine, and due to that terrible war they are now in very short supply and have become expensive.
    Regarding the need to calibrate in order to become useful, I personally disagree. 10X background is still 10X background and is certainly indicative of the lever of radiation. That said, if you feel you must make a judgment solely based on a measurement of exposure such as uSv, then some type of calibration (or comparison to another instrument) is worthwhile. However, the perfect is the enemy of the good in my opinion.
    I’ve built many models of DIY counters, some for IOT, and some for personal use using SMT components. I did not experience any leakage problems with SMT, but HV can be tricky!

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.