I built a geiger counter! It started off as a bag of components and an empty circuit board. Here it is assembled on my desk detecting background radiation. Every beep/click corresponds to a particle of ionizing radiation.![[IMG_2021.mp4]] At the heart of this device is this neat component called a Geiger Muller tube. It's a tube filled with an inert gas. The inside of the tube is coated with a conductor (in this case steel) and a small wire runs down the center. When radiation strikes the outside of the tube, it knocks an electron loose. By itself, this single electron would be hard to detect, but since a large voltage is applied to the steel coating and the wire, it causes the electron to accelerate toward the wire and set off a chain reaction of downstream ionization events. [^1] ![[Pasted image 20220309214306.png]] This is called a Townsend avalanche and causes the gas to conduct electricity, which can be detected as a drop in voltage across the tube. That voltage drop is detected and the device emits a beep/click. Most of these hobbyist tubes are basically Soviet surplus from the Cold War. The tube in this kit is an SBM-20 which you can see written in Cyrillic characters on the side of the tube (СБM20). ![[IMG_2022.jpg]] Here are some data sheets from the internet on this particular tube. They're all written in Russian (I think)! [^2] ![[Pasted image 20220310081925.png]] One neat thing about this kit is that it needs to generate a large voltage across the tube. In the case 400V! But the device is only powered by two AAA batteries. This is the magic circuit that generates that voltage from 3V. It uses a 555 timer and an inductor. [^3] ![[Pasted image 20220309212601.png]] My understanding is that it uses the principle of inductive kickback to generate a high voltage. The 555 timer drives a current across the inductor (L1), and then suddenly shuts it off. The electric field collapses and produces a large voltage. The 555 timer repeats this process over and over. So how do I know this actually works and isn't just emitting random beeps? Well I have this glass bird that's made out of uranium glass. It's literally glass that contains a small amount of uranium which gives it the green/yellow color. I'm told you can find this stuff at basically any antique shop. I bought this off eBay. ![[Pasted image 20220309215017.png]] When you shine a UV light at it, it glows. You might think that it glows *because* it's radioactive, but I think its fluorescences and radioactivity are actually not related. Lots of minerals fluoresce and this form of uranium just happens to fluoresce green under UV light. ![[Pasted image 20220309215001.png]] Ok, so what happens when I place the bird near the geiger counter? It gets very excited. ![[IMG_2023.mp4]] Is this dangerous? Probably not, but let's ask the geiger counter. If I connect a serial cable and open a terminal, I get an estimate of μSieverts/hour, which I can then graph. ![[Pasted image 20220309221106.png]] ![[Image 3-9-22 at 10.09 PM 2.jpg]] So my normal background radiation is about 0.13 μSv/hr. This goes up to about 1.6 μSv/hr with the glass bird sitting on top of it. That's 12x the background level! To put this into perspective, if I strapped this to my forehead for an entire day, I'd be exposed to about 38.4 μSv. According to this handy xkcd chart, that's equivalent to a plane ride from NY to LA. Doing that for 104 days would equal one chest x-ray. [^4] ![[radiation.png]] This is not a calibrated device, so I wouldn't take the exact reading too seriously, but I'm not very concerned about displaying this glass bird on my bookshelf. The distance to it matters a lot too, as the radiation should drop off according to an inverse square law. So up above I mentioned that it's not just emitting random beeps, but actually it kind of is. A Poisson process is probably a good model for the arrival of radioactive particles. I haven't tried to model this, but I'd like to at some point. It also might make a decent source of entropy for a random number generator. Another future project I'd like to tackle is connecting this up to a wifi module for a DIY radiation monitoring station. Maybe it can tweet the current levels or something. [^1]: https://commons.wikimedia.org/wiki/File:Geiger_gamma_interaction.jpg#mw-jump-to-license [^2]: http://www.gstube.com/data/2398/ [^3]: https://mightyohm.com/blog/products/geiger-counter/design-files/ [^4]: https://xkcd.com/radiation/ # Kits - [MightyOhm](https://mightyohm.com/blog/products/geiger-counter/): This is the kit I got. You can also get it off [Spark Fun](https://www.sparkfun.com/products/17869). - [DIYGeiger](https://sites.google.com/site/diygeigercounter/home): These folks put out several kits that are a bit more advanced than the MightyOhm, and I think it is the source material for a lot of DIY geiger counters. If I had known about them first, I might have picked up one of these instead. # Radon Radon is a radioactive element that's gaseous at normal temperatures. It's well known for seeping into the basements of houses and causing lung cancer. This thread is about detecting it with my new shiny geiger counter. My geiger counter is not sensitive enough to detect the presence of radon in the air at the levels that are in my house. Is there some other way it can be detected? Let's think about what happens when radon undergoes decay. When radon decays, it emits a beta particle and turns into polonium. Polonium is also radioactive, and so when it decays it turns into radioactive lead, and so on down the chain. (Image credit: https://commons.wikimedia.org/wiki/File:Decay_chain(4n%2B2,_Uranium_series).svg) ![[Pasted image 20220311020522.png]] [This page](http://tuxgraphics.org/npa/radon-detection-geiger-counter-radioactive-hvac-filter/) points out that these radioactive decay elements of radon, such as polonium, can get stuck in HVAC filters at levels that are high enough to detect with a geiger counter. Let's try it! First some context: my HVAC filter is larger than the opening to the fan unit in my furnace. This means that the fan only draws air through about half the filter, and so only about half the filter actually gets dirty. I guess this means my filter is too large, but it's the one the builders installed and I haven't changed it. Anyway, in the following video, I place the geiger counter on the clean side first, and then the dirty side. There's a noticeable increase in the number of clicks on the dirty side! ![[Radioactive HVAC.mp4]] I think that's good evidence that there's something radioactive being caught in the filter on the dirty side. I can't say for sure if it's polonium or something else (I wish I had a fancier detector that could say something about the element that emitted it), but the theory that it's a decay product of radon seems plausible to me. Having to listen to the audio is a bit cumbersome. Below, I've turned the audio into a spectrogram, which is a visual representation of the frequencies in an audio sample. Each vertical streak is a click and time moves left to right. The vertical streaks are clearly more frequent on the dirty side! ![[Radioactive HVAC Spectrogram.jpg]] I think the video and spectrogram are pretty convincing, but something I've been wanting to try is apply a statistical test to see how likely it is that a sample of something is emitting above background radiation. My geiger counter gives me access to a Counts Per Second (CPS) metric, which is just the number of clicks in the last second. I can get this data for every second through the serial port of the geiger counter. This allows me to run the following experiment: Expose the geiger counter to the clean side of the filter and collect the CPS samples for 5 minutes. Then do the same with the geiger counter exposed to the dirty side of the filter. When I run this experiment I get the following average CPSs: | Condition | Average CPS Over 5 Minutes | | ------------ | -------------------------- | | Clean Filter | 0.406 | | Dirty Filter | 2.644 | That's 5 minutes of data so there's a total of 300 samples. We don't need a t-test to tell us that's significant, but just for fun, the p-value of the one tailed t-test is 6.98E-72 so I think we can safely say the dirty side is significantly different than the clean side. [Here's the spreadsheet](https://docs.google.com/spreadsheets/d/1ATSqVdqCXWhZNtvQwEzf63RkXGyFp3Wl6Zw2IQM7YHw/edit#gid=1751485462&range=B1) with the data if you're curious. The geiger counter estimates this is about 0.90 uSv/hr. Background is about 0.15 uSv/hr at my house. If my HVAC filter is full of radioactive elements, is it possible other bits of dust around my house are also? I tried to run this experiment again but with a ziploc full of dust from my vacuum cleaner. I could *not* detect radiation above the background rate. The p-value was only 0.145. | Condition | Average CPS Over 5 Minutes | | ---------- | -------------------------- | | Background | 0.397 | | Dust | 0.452 | Why the difference? I'm not sure, but it's plausible that my HVAC filter is much better at trapping tiny particles of polonium than my vacuum cleaner is. After all, trapping tiny particles in the air is what the filter is designed to do. By the way, these decay elements are part of why radon is dangerous. According to wikipedia: > Unlike the gaseous radon itself, radon daughters are solids and stick to surfaces, such as airborne dust particles, which can cause lung cancer if inhaled.