Safecast Geiger Counter Reference Design

This past weekend marked the anniversary of the Tohoku-Oki earthquake that devastated Japan. I had not felt my blood so cold since I watched the twin towers fall almost a decade earlier. I still vividly remember the twisting knots I felt in my stomach as I watched the footage of a tsunami wiping out huge swathes of Japanese countryside. In a matter of hours, entire cities were washed off the map, leaving an eerie post-apocalyptic landscape of a few survivors weeping amongst twisted wreckage. Then, in the ensuing days, Fukushima Daiichi melted down, leaving in its wake one of the worst on-going radiation contamination crisis since Chernobyl.

I have good friends in Japan, and I visit often. I wanted to do something to help, but I didn’t know what I could do. I was connected by Joi Ito to Safecast, and I joined the effort to build an open sensor network that could aggregate trustable, source-neutral radiation monitoring data. Safecast itself has many talented and hard working volunteers who have done a remarkable job of achieving their goals by instrumenting Japan with radiation monitors and aggregating data through cleverly designed and rapidly deployable mobile monitoring capabilities.

I decided my tiny contribution to the effort would be to design a radiation monitor suitable for everyday civilian use. This is a preventative/preparedness measure, addressing the long-term issue of empowering a civilian population with few available options for power generation to self-monitor their environment. The problem with the current crop of radiation monitors is that they are basically laboratory instruments: accurate & reliable, but bulky, expensive, and difficult to use, requiring a degree in nuclear physics to understand exactly what the readings meant. Another problem with crises like these is that while radiation monitoring is important, it’s something that is typically neglected by the civilian population until it is too late.

Therefore, the challenge set out before me was to design a new Geiger counter that was not only more intuitive and easier to use than the current crop, but was also sufficiently stylish so that civilians would feel natural carrying it around on a daily basis. Furthermore, it had to provide extensive logging capabilities, as radiation monitors are typically not turned “on” until after the fact. It also had to operate effectively in catastrophic conditions, i.e. in scenarios where internet and power have been cut for days. Finally, the data collected by the instrument had to pass any scrutiny thrown its way, and the collected data had to be traceable to a given instrument so that if its calibration is incorrect, its data can be selectively excluded without poisoning the entire dataset. Radiation monitoring is a politically sensitive subject, and certain parties have interests to manipulate the data one way or the other to promote their views with the public. Ad-hoc data collection networks suffer from the possibility that their efforts can be discredited by institutions with big budgets who find that the readings represent an inconvenient truth.

Radiation sensing primer

Radiation measurements are subtle, partly because radiation comes in many flavors. Many Geiger counters can only efficiently detect the most energetic kind of radiation, gamma radiation. This includes the Geiger counters frequently favored by government and regulatory agencies. However, there are weaker forms of radiation (alpha and beta) which often go overlooked that can also pose a human health risk, particularly if they are ingested or inhaled. These weaker forms of radiation are also by-products of a nuclear meltdown, and because they come from different isotopes they have different patterns of distribution and absorption in the environment.

Because of the diversity of radiation sources and their varying biological impact, it is very hard to determine if an environment is safe in the face of an elevated Geiger counter reading. However, improved historical and spatial distribution records of background radiation measurements can help identify when there is a spike in radiation, which is a clearer cause for concern.

In the interest of creating a complete solution for public health needs, a core design requirement of the new Geiger counter is to incorporate a sensor that could detect all three forms of radiation. This type of sensor is a “pancake” style Geiger tube, which has a large mica window that enables sensitivity to all three kinds of radiation. The ultimate selection of the LND7317 pancake tube plus iRover HV radiation sensing core influenced every aspect of the industrial design (ID) and internal electronics.

There and Back Again: a Hacker’s Tale

I thought it would be interesting to share not only the final design, but also the intermediate designs that were scrapped en route to achieving a final design. Design is an iterative process, where one has to make difficult choices about what to include and more significantly what to leave out. It’s extremely rare to see what got left on the cutting room floor, but I saved my notes along the way so I could share them with you now.

Initial Design Sketch

Above is a rendering of the first design sketch, made back when Safecast had the name of “RDTN”. I do all my industrial design using Solidworks, a survival skill I picked up during my tenure at chumby designing consumer electronics. I came up with this in the first couple of weeks after the disaster. This design incorporated a low-sensitivity tube from Sparkfun, because at that time I did not understand the importance of using a pancake tube.

The biggest problem I wanted to solve with this design is user abandonment. Radiation leaks are thankfully rare events. However, this also means that when an event happens, there is typically a lack of pre-crisis background data against which to compare the post-crisis readings. Therefore, I wanted to build a device that people would be compelled to carry around every day and use for years at a time.

My thought is that the average consumer would have a hard time justifying carrying around yet another gadget in their pockets or purses for the sole purpose of measuring radiation. Within weeks or even days of getting nothing but “safe” readings, the Geiger counter would be forgotten and left at home to languish until after the next crisis.

One way to compel users to carry around a Geiger counter is to put it into something they already carry all the time. While it would be tough to convince a smartphone vendor to incorporate a very expensive and bulky Geiger tube, many smartphone users also carry around a spare battery pack, which they use almost every day. So, I thought it might be a good idea to trojan horse a Geiger tube into such a battery pack.

The sketch above demonstrates such an incarnation. This design is basically a battery pack that can charge a smartphone, but also incorporates a Geiger tube, an LED flashlight (handy in an emergency when there is no power), and some logging circuitry. The Geiger counter would upload its log data to the Safecast network whenever a user plugged in to charge the phone.

The design itself is minimalist, with a shape inspired by the steam cooling towers frequently used to iconify nuclear power in western media. The shape was chosen to remind us that sometimes we have no choice but to harvest the power of the atom, and a well-equipped and informed civilian data collection network is a key factor in trusting the safety of our power sources.

A second iteration

The first sketch had to be abandoned, primarily because the sensor it was designed around was too small to effectively measure alpha and beta radiation. After Safecast settled on the LND7317 Geiger tube as the standard reference sensor, I started re-designing the sensor around the new tube.

The problem with the larger, more sensitive sensor is that it was big – over a half-inch thick, and a couple inches in diameter. Below is a sketch of a design study aimed at creating the smallest possible Geiger counter that could also incorporate the large pancake tube. It’s about the size of a hockey puck, but a bit thicker. In order to keep the size and weight of the Geiger counter reasonable, I had to abandon any notion of a dual-purpose as a battery pack. Instead, I had to rely on “sex appeal” alone to compel users to carry the device around. I wanted to make the Geiger counter something unique and aesthetically pleasing, something you would enjoy carrying around frequently. I started from a minimalist design – the puck – and endeavored to design-out any outward indicators or displays. Hence in this sketch, the radiation measurement is provided by a set of super-high efficiency 7-segment LEDs that could shine the numbers through a seemingly opaque white shell. The design’s shape and feel was meant to be somewhere in between “Eve” from WALL-E and an egg.

Unfortunately, this design, too, had to be abandoned because at the time when I was drawing up the sketches, I didn’t have detailed mechanical drawings of the LND7317 tube. When I was finally given a sample of the tube and drawings for it, I discovered there was not only the puck-like body, but also a nearly 1″ long protrusion for the cathode and anode. This completely destroyed any notion of building a puck-like sensor.

Closing in on the final design

Below is a rendering of an attempt to accommodate the accurate CAD model for the LND7317 into an ID that stayed faithful to the Eve/egg design inspirations. The puck was elongated to the minimum dimensions required to house all the internal components. Again, the hidden 7-segment LED display motif was employed.

The final design

After much discussion and review with the Safecast team, we decided that a key component of the user experience should be a graphic display, instead of a 7-segment LED readout. Therefore, a 128×128 pixel OLED panel was incorporated into the design. The OLED panel would be mounted behind a continuous outer shell, so there would be no seams or outward design features resulting from the introduction of the OLED. However, as the OLED is not bright enough to shine through an opaque white plastic exterior shell, a clear window had to be provided for the OLED. As a result, the naturally black color of the OLED caused the preferred color scheme of the exterior case to go from light colors to dark colors. User interaction would occur through a captouch button array hidden behind the same shell, with perhaps silkscreen outlines to provide hints as to where the buttons were underneath the shell. I had originally resisted the idea of using the OLED because it’s very expensive, but once I saw how much an LND7317 tube would cost in volume, I realized that it would be silly to not add a premium feature like an OLED. Due to the sensor alone, the retail price of the device would be in the hundreds of dollars; so adding an OLED display would help make the device “feel” a lot more valuable than a 7-segment LED display, even though the OLED’s presence is largely irrelevant to the core function of the apparatus.

The design also lacks any integrated radio connection. A popular request for the design was the incorporation of a bluetooth or zigbee style radio; however, a combination of a very stringent battery life goal (several months of standby time) and low manufacturing volumes meant that it was impractical to incorporate a radio into the device. It’s a slippery slope to start adding features like GPS and bluetooth – to add those features, you’d need to upgrade the microcontroller, at which point you’re basically building a very expensive, heavy and large cell phone with a geiger counter in it. Furthermore, the entire development effort was being done by an unpaid volunteer operating on a shoestring budget – Safecast isn’t Apple. So, rather than build a buggy cell phone that can sense radiation, I’d rather build an outstanding Geiger counter; hence the decision to focus efforts and resources on core functionality, with the sole allowance being the inclusion of the OLED + captouch array for improved UI. This is a controversial design decision and I fully expect to be chastised for it.

The Prototypes

Once the design was finished, the next step was to build prototypes. This is the really fun part, where you turn your ideas into something you can touch and hold.

The prototypes are made out of CNC-machined ABS (even the clear part!). The cosmetic moldings that go over the connectors were also built and they do fit, but because of their expense and fragility (CNC milled ABS lacks the robustness of injection-molded ABS), I try not to install them, even for glamor shots. To wit, the whole thing was done on a shoestring budget, as Safecast is a non-profit; two full prototypes were built, including PCB fab, assembly, and CNC milling for one and a half revisions of the cases, for a bit under $3k total.

An important point readers should note about this design is that I’m not manufacturing this Geiger counter reference design. My contribution is limited to design IP only. Practically speaking, I’d make a terrible Geiger counter supplier, because I don’t have the credibility or history in the industry. Instead, the design has been donated to the community, thereby enabling International Medcom, a business that has spent decades specialized in producing high-quality Geiger counters, to bring this to the market. If you’re interested in getting one of these, keep an eye on their website.

The final design features include:

  • LND7317 pancake tube + iRover HV board
  • STM32-based microcontroller; sufficient CPU power to digitally sign logs with a unique private key as a non-repudiation/anti-tamper measure
  • 450 mAh Li-poly battery
  • 3-axis accelerometer so sensor orientation can be recorded
  • 128×128 color OLED display
  • 6-button captouch array
  • “hold” button on the back to lock the captouch array and prevent false triggering of the power-hungry UI elements
  • lanyard attachment (important for the Japanese market)
  • microUSB port for charging and data upload interface, featuring an FTDI-based serial chipset capable of loading firmware into the microcontroller
  • 3.5mm jack capable of bidirectional audio
  • embedded hall-effect sensor (to detect attachments, e.g. for occluding alpha or beta radiation)
  • audible event notification via piezo buzzer
  • low-power visual event notification via conventional LED
  • real-time clock
  • a high-quality entropy source ;-)
  • I am a proponent of open source hardware; so here’s the source files for my design! All of the following source files are licensed under CC3.0-BY-SA with my XL1.0 automatic patent cross-license rider (CC doesn’t address patents, so I invented my own rider that piggybacks on CC to ensure that any patents that may arise from this or its derivatives are automatically cross-licensed to the community).

  • Altium design source / schematics / gerbers / BOM for the mainboard electronics
  • Altium design source / schematics / gerbers / BOM for the buttonboard electronics
  • Solidworks design source / IGES / STEP for the industrial design
  • For those who don’t have 3D design tools, you can install Solidwork’s free e-drawings viewer and look at the easm file, or if you run windows you can download this executable and just run it
  • Of course, a hardware prototype is only the beginning – there’s a huge amount of effort remaining on the software. To bootstrap things, xobs and I have coded up a core demonstration system based on Leaf Lab’s libmaple. You can peruse the code and/or download it at github. Basically, this demo system provides an architecture to easily register drivers and facilitate power management. The validation demo shown running on the prototype photos above indicate that all of the hardware features work. But, the software has yet to have a layer of polish and shine added in terms of the UI and power management optimization.

    A key design goal electronics’ system design was to enable community participation. As such, I eschewed the use of JTAG adapters during development. Instead, hooks were provided in the hardware to enable the integrated FTDI USB-serial controller to flash the microcontroller’s firmware via a “bitbang” interface. As a result, anyone who has an interest in developing for this Geiger counter can simply plug it into their laptop’s USB port and start coding without any need for proprietary JTAG adapters or proprietary software to purchase, as the entire developer’s toolchain is available in source form. We were able to code up and test the entire functionality demo (including sleep/stop/standby power management) using nothing more than the USB-serial capability built into the design. As I write this, I realize I had neglected to upload the firmware loader to github, so here’s a tarball for it; currently, the loader only runs under Linux and OSX.

    I think there’s some fun things the community could do with the UI on a Geiger counter. At the very least, the microcontroller has sufficient power to play Tetris. Another whimsical thought was to build a subsystem that would play music out the audio port based upon the current radiation level — calm, ambient music in low-radiation environments escalating to death metal and the sound track of “Run Lola Run” at dangerous levels.

    So that’s it! I hope that the design ultimately helps the people of Japan – or people anywhere in the world where radiation contamination may be a concern – to feel more empowered and in control of their situation.

    57 Responses to “Safecast Geiger Counter Reference Design”

    1. […] Safecast Geiger Counter Reference Design @ bunnie’s blog. Amazing. Filed under: open source hardware — by adafruit, posted March 14, 2012 at 5:06 pm Comments (0) hr{display:none} .in_the_store{background-image: url(""); background-repeat: repeat-x;width:634px;border: 1px solid #CCCCCC; float:left;} In the Adafruit Store: Adafruit Proto Cape Kit for Beagle Bone Arcade Button – 30mm Translucent Clear SMT Cool White 5050 LED – 10 pack […]

    2. Bunnie, this looks amazing! Great work! Can’t wait to see how this develops more!

    3. Andy Lutomirski says:

      This looks great, but I feel obligated to point out a pet peeve of mine. You say:

      However, there are weaker forms of radiation (alpha and beta) which often go overlooked that can also pose a human health risk, particularly if they are ingested or inhaled. These weaker forms of radiation are also by-products of a nuclear meltdown, and because they come from different isotopes they have different patterns of distribution and absorption in the environment.

      There’s a big difference between radiation and radioactive things. Radiation consists of fast-moving, highly energetic particles — mostly alphas, betas, and gammas. They are very short lived and are not environmental contaminants [1], just like “light pollution” is not a persistent pollutant. You can’t ingest an alpha particle or a beta particle in a dangerous way.

      Radioactive material is stuff that emits radiation. This is what contaminates the environment. Some radioactive materials emit alpha particles; these materials are mostly harmless unless inhaled or ingested.

      This distinction matters — if a nuclear plant leaks *radiation* due to some kind of error, then you really don’t want to be around when the leak is happening, but there’s no harm in visiting the plant afterwards. If the plant releases radioactive material into the atmosphere or water, then there could be a long-term problem.

      • bunnie says:

        My understanding is that alpha and beta emitters are dangerous specifically because they can be ingested and incorporated into human tissue, where they can cause damage.

        Isotopes of Strontium and Caesium are both decay via beta radiation, which is relatively harmless because beta can be stopped by the layer of dead tissue surrounding you (your skin).

        However, these isotopes can be incorporated by your metabolism into your body. My understanding is that Caesium in particular tends to be mistaken as calcium by your body and incorporated into bone material, which can lead to cancer in the surrounding tissues as the beta particles are emitted.

        While you’re right that environmental Caesium is no harm, when you live near a reactor that has generated a large amount of Caesium, that Caesium can be enriched through the food chain and can eventually end up on your plate.

        Perhaps your opinion is more correct than mine, but rather than simply assert to residents around a reactor that they have nothing to worry about because low-grade emitter isotopes cannot be incorporated into their bodies, I’d rather give them the option to detect them, and if they are worried, seek their own professional opinions about what to do next.

        • Christian Vogel says:

          Hi Bunnie,

          I think Andy mostly refered to your mixup of “radiation” and “contamination” as visible in you stating “radiation … if ingested or inhaled.”, and while he’s for sure right, it’s unfortunately a much too common mistake.

          But keep in mind that you will most likely not be able to detect α, those particles already are shielded by a piece of paper. And β will be somewhat attenuated by the case -> usual commercial counters will have only a very thin (capton-)foil window (which is very delicate) for that matter.

          From the standpoint of public safety, that’s not really important though, because contamination will almost certainly not be isotope specific, and γs are abundant wherever you have “dirty” material anyway.

          If you are interested in the actual isotope composition, you would have to design a energy-resolving (spectroscopy) detector, by definition the geiger-counter is not. Spectroscopy detectors commonly are high-volume (e.g. the size of a small battery) PIN diodes, made of Germanium and cooled to liquid nitrogen temperature. And those things are hideously expensive…


        • Mike B says:

          Hi, Bunnie.
          Thank you for this article. I found the description of the design process quite interesting.
          Andy is correct that it is important to keep in mind the distinction between radiation and radioactive material, both to prevent confusion in the general public and maintain credibility in the rad community.
          I am not sure it is worth trying to make an instrument of the size you want that can detect alpha and beta, for several reasons.
          First, in order to detect alpha and beta, you need a very thin window on your detector. The window can be broken easily, and will never survive being carried in a pocket/purse/backpack/whatever.
          Second, alpha and beta in almost any media are subject to “self shielding”. This means that if you have something like lettuce leaves, you might be able to detect contamination on the outside of the leaf, but you probably wouldn’t be able to detect alpha and beta emitters that were taken up into the plant from the soil. Finding those requires a lab and wet chemistry, usually. Even a thin film of moisture will shield alpha/beta.
          Third, after a fairly short time after an event, most of the short half-life isotopes will have decayed away, leaving, in most cases, tritium (H3) and strontium-90 (Sr90) as the two beta emitters of interest. Unfortunately, both of these isotopes have such low energy that they can’t be detected with any handheld instrument that I am aware of.
          It should be noted that all alpha and most beta emitters also emit a gamma ray, though sometimes the gamma has such a low energy that it is hard to measure.
          All this adds up to it probably being more practical to design a robust unit that only detects gamma, rather than a fragile unit that can detect gamma, and would be able to detect alpha/beta if it weren’t for other factors.

    4. Frank Oppenheimer says:

      Is there a good alternative to the OLED technology? Inorganic LEDs, perhaps?

      While I might not mind replacing a cell phone when it’s OLEDS degrade, I certainly don’t want to replace a counter. OLEDs blue LED degrades after only a couple of years, based on continuous use.

      • bunnie says:

        A conventional LCD isn’t bad either, but they are always thicker than OLEDs due to the need for a backlight diffuser. Thinness was an objective here, and thus an OLED was chosen.

        An option for the design was to use inorganic LED alphanumeric displays, similar to those made by Avago (HP) back in the day that were found on the original motorola startac phones. However, they are extremely expensive as they are not manufactured as an integrated sheet, but as an array of wire-bonded LEDs.

        OLEDs will have a lifetime of thousands, if not tens of thousands, of hours of continuous operation. Therefore I don’t think that wear-out will be a huge issue for a Geiger counter, particularly as I would recommend the final software dim the display after a few seconds of non-use, simply for battery life reasons. The original Sony OLED panel did have a problem with bluing, but that was when the technology was quite new. They are continuously improving it and extending the lifetime of the panels.

        • Juhele says:

          I am working for some institute in the area of radiation protection and detection and I must say, that modern color displays are a big problem – especially on direct sunlight.

          We have some detectors with classic BW lcd displays and this sort of display seems to be the best we can have – good contrast, no problem with light, low power consumption.

          It is a pity that you cannot use such display as you say.

          • Kazriko says:

            I wonder if an ePaper display like the Pebble watches would be superior to LCD in both power consumption and thickness?

    5. Luke Weston says:

      If you could provide some more description, details, schematics or links etc. regarding the HV power supply circuit or the “iRover HV radiation sensing core” I would appreciate that.

      • bunnie says:

        The iRover HV board will be made available by International Medcom. You can check with them for more information on the board.

        However, I can share with you that the HV board interface is just a three-pin header with power, ground, and a pulse that gets kicked out whenever there’s a radiation event. There are a number of really good articles and designs in circulation that can give you guidance on how to roll your own HV supply if you don’t want to buy one. Here’s a link to one.

        I opted not to design my own HV supply in part because while it’s relatively easy to get one up and running and detecting a few pips of radiation, getting one that’s low power, can work over a broad range of temperatures and conditions, not kill you if you accidentally touch it, and can properly handle the dead-time and quench issues of a Geiger tube is difficult. Credibility of sensor readings is absolutely critical for the product to have a social impact; otherwise it’s just a toy. Therefore, I decided to use a vetted and tested core consisting of a matched tube and HV supply pair that essentially does a radiation-to-digital conversion at a high fidelity. It’s a bit like the distinction of either building your own R2R ADC or buying a high-end Burr-Brown ADC for your audio equipment, or rolling your own wifi module versus buying a pre-built wifi PCB that’s already shipping in volume equipment. Intellectually, it’s neat to make your own, but as far as the market is concerned, it’s safer to go with the pre-built module.

    6. I’d love to use this in my movie. I was getting ready to order an old russian geiger counter. Do you have a prototype you’d like to loan out?

      • bunnie says:

        There’s only two in existence at the moment in the world, so probably no loaners today. But I know that International Medcom is working on building more. You should try to reach out to them, they may have prototypes they can lend you:

    7. A nice concept, but if you’re using a tube sensitive to alpha and beta particles, it makes no sense to entirely enclose it in a plastic case; that’ll block all of the alphas and many of the betas. (They’re not strictly speaking weaker — they can pack high energy — just less penetrating.) The thin mica window is fragile, so you do need to enclose it, but you also need a door that opens to reveal it. I can’t tell if the design currently incorporates such a door, but it doesn’t look like it, and you don’t mention one.

      Another use of the door is to discriminate between types of radiation: cover it with plastic, and you cut out the alphas; cover it with a couple of millimeters of aluminum, and you cut out the betas too. So two doors, one inside the other, might be desirable.

      Unfortunately with a Geiger tube there’s no possibility of radiation spectroscopy, as per this amateur video:

      • Hmm, playing some more with the Solidworks viewer, I notice that the component covering the tube is labeled “mesh”, so evidently you’re not unaware of this issue. Finding a “mesh” that’s resistant enough to abuse to live in a pocket (where pocket lint and dirt might find its way behind the mesh, and keys might scar it), yet transparent enough to radiation, seems a bit difficult, but maybe…

        • bunnie says:

          Yes, indeed, the mesh is a critical component. I didn’t develop that technology, I’m taking it straight from the pros. This is part of the reason why International Medcom is doing the manufacturing, they have decades of experience in this area and they know how to do it right. My understanding is that the final mesh is made out of a thin alloy of copper-beryllium, that is perforated to allow alpha and beta through. Certainly, it doesn’t provide super robust protection but it seems to be good enough for a large range of high-end, field-tested geiger counters.

          Here’s a link to an image of the bottom of the counter, showing the mesh shield:

          The shield itself is inset from the plastic body by a couple of mm to prevent it from interacting too much with the environment.

          • Rinky says:

            I hope that the tube is easily replaceable — those mica windows pop easily (even with the mesh in place). Typically, a pancake GM will have a plastic cover to protect the tube when one is not using it (designs like this one will typically have a sliding cover).
            Perhaps the recess will allow the placement of a protective cover..

        • Nicholas says:

          This “mesh” is probably a removable cover for the tube’s mica window, which will be easily damaged by loose change, keys, placement on rough concrete, etc.. Detecting installation of the cover is probably the purpose of the occlusion-detecting hall effect sensor Bunnie mentioned. It’s important for the Safecast to make it clear when the alpha/beta environment cannot be evaluated due to window occlusion.

          I am curious how this particular tube was chosen and would like to hear more about the use-cases that drove the product development. Promising (and cheap) experimental detectors like 4H-SiC or polyethylene naphthalate seem like a great fit for the radical consumer-orientation of this design, but there are a lot of tradeoffs and (I suppose) IP issues involved. Along those same lines: can the Safecast be operated as a proportional counter, or is it just for should-I-stay-or-should-I-go decisions?

          • bunnie says:

            If I were to choose a theme for product development around the radiation sensing core, it would be “no risk”.

            Radiation sensing is a politically charged topic, and we didn’t want to leave an avenue for someone to come in and toss our entire corpus of data. There’s a lot of great new sensing technology out there — scintillators, cameras, etc. — but they are relatively unproven and difficult to compare to existing bodies of data.

            The move to use an a/b/g sensitive sensor was controversial in itself. The issue being that at the hopefully low levels of elevated radiation you’d be finding, it’s controversial as to what has a biological impact on your health. Most of the corpus of data relating to studies linking exposure to health uses readings taken by Geiger tubes. Since each sensor technology has its particular biases, we decided to stick with a Geiger tube so that our data can be as comparable as possible to existing data. Of course, the introduction of a/b sensitivity throws a skew in that, but the a/b sensitivity can be dulled by simply adding a shield to the detector to block the two forms.

            Calibrated shields that can be slid on to the geiger counter similar to the silicone rubber smartphone “condoms” is an accessory that’s on the drawing board. The inclusion of a magnetic sensor in the design allows there to be a magnet inside the shield so that the counter itself can note if the data is a/b sensitive or not.

            In the end, I agree with the notion of including a/b sensitivity, because I don’t like the idea that we shouldn’t care about a particular hazard because we never cared about it before, particularly when there’s emerging evidence that we should. And, as I mention above, the tube can be easily retrofitted to do pure-g sensing and thus be more comparable to the energy compensated gamma tubes often used in research settings.

            The design is ultimately about empowering individuals to sense their environment, and doing it in a way that their readings can be taken credibly and understood by local authorities.

            • Nicholas says:

              Great reply! Thanks for the information. This project sounds like it was a lot of fun to work on, and making instrument-grade radiation measurement available to consumers is really exciting. I look forward to seeing what’s possible with the data analysis once these devices are available for sale.

            • What’s controversial about damage from alpha and beta particles? Sure, alpha particles don’t penetrate the skin, so having an alpha emitter on your palm is not a big deal. But if alpha particles are floating around such that they’ve gotten to your detector, they’ve gotten into your lungs, too, and nobody with even an eighth of a clue denies that that’s a major hazard. (Well, the hazard level depends on dose, as always; but for a given energy dose delivered inside the body, the standard estimate is that alphas are twenty times as damaging as gammas.) It was the alpha emitter Polonium-210 that killed Alexander Litvinenko.

              The only way I can see this being controversial is if people are arguing that probably a radiological disaster will have enough gamma emitters floating around that measuring betas and alphas doesn’t give much more information — at least not the sort of information that tells you where to run from, and how fast to run.

              As for the protective mesh’s prior use in high-end devices, high-end devices usually have high-end cases for storage, and well-trained operators. Oh, and organizations ready to pay for replacement tubes and meshes when those operators screw up, training not being all that it’s cracked up to be. A device that’s to live in the average person’s pocket, and to be paid for by him personally, is a different world. I mean, this sort of feature is catnip to me — I know what it’s worth, and know how to take care of delicate devices — but there aren’t enough of me to make a market.

    8. emeb says:

      I’ve been working on a few projects based on the STM32 family and find them fairly easy to use, both from software and hardware standpoints. Any particular reason you chose to use FTDI/serial for your USB rather than using one of the STM32 parts that has USB built-in?

      • Colin D Bennett says:

        Hmmm… good question. Using an FTDI FT232 must increase the product cost since the FT232 is an extra IC (and consequent board space used); a USB STM32 would do the same job without any external ICs.

        • Colin D Bennett says:

          After further analysis, I see that Bunnie has done something more clever than just dropping in the FT232 for plain vanilla UART-USB interface.

          First, he’s using the STM32F100 Value Line series, so it will cost an extra $1.00 or so to upgrade to the ‘F102 or ‘F103 line to get USB. It would still be most cost-efficient, however, to use built-in STM32 USB since the FT232RL IC is going to cost more than the incremental cost to upgrade to a ‘F103 Performance Line MCU. Also, the attendant board space would be saved.

          Second, the FT232 has some of its GPIO control lines connected to the STM32 BOOT0 and NRST pins. This makes it super easy to take advantage of the STM32’s built-in ROM UART boot loader for firmware upload. The PC can set BOOT0 high and pulse the NRST pin to invoke the boot loader, then use the USB-serial converter to upload firmware to the MCU.

          So without the FT232 IC, you’d need to use a USB bootloader on the STM32F103 in order to implement USB firmware upload. But that’s already been done, e.g., by the LeafLabs Maple ( — MIT licensed), and others. So I’d still be interested to hear Bunnie’s reasoning on why he chose to use the ol’ FT232.

          • bunnie says:

            The reason is that the FT232 drivers are built into a large number of systems already, whereas the other drivers are not. So it’s a customer support issue. If a driver installation problem generates a customer support query, that can cost a couple dollars of effective man power to remedy, and is a huge hassle especially for a volunteer-run organization.

            So the cost of drivers & support is essentially bundled into the cost of the FT232 IC, and passed on to the customers within the product, as opposed to potentially burdening the organization down the road.

    9. Curtis says:

      Did you include a facility for geotagging the logged data? If you are concerned about the credibility of the measurements, including a location stamp with the signed log data would seem to be pretty useful. Adding a software interface for an external USB GPS dongle seems like it would be pretty straightforward, but that would probably not help the battery life.

    10. Tony says:

      If put into production, about how much would it cost to buy the device?

    11. Justin says:

      Nice work! Also really interesting to go through the design process step by step, explaining why decisions were made, why they changed, etc.

      I’ve designed some (hobby) geiger counters myself. I know feature creep is both undesirable and too late in this instance, but I had a suggestion and thought I’d throw it out there for possible future devices – a geiger tube uses so little power at background levels, that you can run it off a very small and cheap solar panel – smaller than this size of the device. That powered-forever feature could make a device timeless. The panel can obviously also keep the battery topped up.

    12. […] counter to help citizens in Japan detect radiation in the wake of the nuclear disaster in Japan, Huang writes on his blog. After several design iterations, Huang writes that he created a Geiger counter design that he […]

    13. […] Geiger counter to help citizens in Japan detect radiation in the wake of the nuclear disaster, Huang writes on his blog. After several design iterations, Huang writes that he created a Geiger counter design that he […]

    14. […] Bunnie Huang, a member of the MAKE Tech Advisory Board, tasked himself with the goal of designing a civilian-friendly Geiger counter. He was inspired by the efforts of Safecast, an organization whose goal is to build an open sensor […]

    15. […]  |  Andrew “Bunnie” Huang  | Email this | Comments __spr_config = { pid: '4ebf8e94396cef5c9c00016b', […]

    16. […]  |  Andrew “Bunnie” Huang  | Email this | Comments Share this:DiggFacebookStumbleUponEmailPrint […]

    17. […]  |  Andrew “Bunnie” Huang  | Email this | Comments Posted in Uncategorized | Leave a […]

    18. […]  |  Andrew “Bunnie” Huang  | Email this | Comments Engadget Tags: Builds, Bunnie, civilians, counter, […]

    19. jj says:

      Nice but i think at this point a discrete device doesn’t have much of a future and chances are phones could gain the capability soon enough (DoCoMo has a prototype phone case for example).

    20. Ryan says:


    21. Toby says:

      This project is doing some intresting work along the same lines as your project.

    22. psuedonymous says:

      How is calibration handled? The super-cheap-ebay-dosimiter-of-questionable-quality that have become popular can produce wildly inaccurate readings, leading either to unnecessary panic (and poisoned/unreliable community survey maps) over a false-high reading, or (unlikely for the general population who are distant to Fukushima itself) potentially failing to indicate a high level. You touched on the readings being traceable to a counter to allow poorly calibrated devices to be excluded from datasets, but for a DIY device there must still be some initial calibration involved.

    23. AlexT says:

      Very nice effort.

      What’s next ? Is there anyone interested in mass producing this ?

    24. ank says:

      Good work and great idea. Though didn’t I read somewhere that there is a cheaper alternative namely a closed box with a German cat in it.

    25. Chris says:

      Hi Bunnie,

      This looks like a really great device and I would like to check it out by building a few. I work for SE International, Inc. and have been involved with the development and marketing of radiation detectors for a bit. You might have seen our Inspector Survey Meter around Japan over the last year. If everything seems to test out okay, what would we have to do in terms of producing this instrument on a commercial scale? I read through the creative commons license and from what I can glean, it appears that it’s an open source unit that anyone can make, but if any changes or improvements are made to the design, we share those developments in turn. That, and we give credit to your design. Would a mention of this on the label be sufficient? If there’s more to it, please let me know. I think this unit has some potential and would like to look at the possibilities of mass producing it. We already have many of the components in house, so I think we could get something going relatively quick. Thanks for your contribution to this important endeavor and I look forward to talking with you.

      • bunnie says:

        You basically have it right. That’s how open source works — you can use it, build it, but if you improve it you must contribute back to the community. Also, if you patent a derivative work you must also cross-license that back to the community.

        For the CC licensing, you should note next to the copyright notice, wherever it is, with the text “Licensed under a CC3.0-BY-SA + XL1.0”.

        For the acknowledgement aspect required by the CC license, I think the most tasteful way to handle it would be to put my name alongside the mandatory F/OSS license notices, and note that yours is a derivative work of mine (by using my work, you also require attribution for the embodiment of the effort). I wouldn’t want my name brandished on the outside of the products, that’s a bit ostentatious. F/OSS-compliance notices are typically included in the user manual and in the “about” section in the software menu tree; if you use Firefox you can see an example in Help->About Firefox->Licensing Information. Also, you need to retain the OSHW logo somewhere visible on the PCB: .

        I think that’s about it for the compliance bits.

        • Chris says:

          Sweet! Thanks for the comprehensive reply. Sounds like a great formula for innovation and development. We’ll check it out and if we run into any ideas or solutions along the way, we’ll be sure to contribute back to the effort. Would accessories for the unit be covered under the license as well? We tend to get requests from customers that need a specific accessory for an application.

    26. Barbara says:

      I’d like to know if this device will allow the user to test food to check the level of radiation. Most of the other devices available cannot do this.

    27. Sandra jacobs says:

      AS a complete novice I’ve been researching this, and it’s plain that geiger muller counters, i.e. pancake ones, are not sensitive enough to pick food radiaiton anywhere near the standard for food. Scitillation counters are better – but they have to have a big enough crystal, and you need a lead covered space to separate from background radiation, and you need to be able to test it over 24 hrs +.

      I saw a YouTube demo of a scintillation counter detecting radiation in a rock near atomic test site area, but the user pointed out that it was 1000s times more radioactive than the food standard allows.

      Basically, if food is at risk, you need to eat imported food and take several nutrients proven to either help eradicate nuclides from the body or protect from the radiation coming from the nuclides into the body cells. There are some that do this: spirulina and apple pectin remove nuclides (minerals/metals) and alpha-lipioc acid and vitamin C either does this or protects, and gamma tocotrienol (a form of vitamin E) will protect. These are the easiest supplements to procure and take.

      You need to avoid green vegs and dairy from suspected areas, as these contain strontium which replaces calcium in the bones, and also I think caesium does too. Cesium also replaces potassium which tends to get eliminated from the body – but obviously if you are ingesting this all the time it will do damage.

      Most nuclides are gamma emitters – but the whole pint is that even a scintillation counter (which detects gamma) made for ordinary people doesn’t detect food levels.

      Bee research make scintilliation counters for under $2000 that they say are will detect small amounts of C134 and C137, but I think not down to food levels. But this would be the best t get if you want a sensitive counter.

      My understanding is that the normal geiger muller (pancake) counter will only detect fairly gross levels – high spots etc. But if it’s in the air, then you probably want to move out before this. Or, as I said, take the supplements.

      Someone with a scintillation counter was sent dirt from a Tokyo gutter by a friend in japan the other week, and it tested radioactive. They pointed out that this was much much higher than what you would want to eat.

      But, probably a geiger counter would hardly register the radiation.

      I presume the radiation was brought down by the rain. You don’t need to play in the gutter, but plants are growing in the soil and so on.

    28. Cody says:

      Hi Bunnie, I’m working on a senior project that may incorporate a Geiger counter and have been looking at your excellent work as reference! I saw earlier you pointed somebody asking about the HV board to International Medcom. Before I go bugging them, can you tell me if that part of the design is open-source and whether it’s worth asking for? I have to ask because everything in my project must be published under some kind of open-source license and easily reproducible by others.

      My application doesn’t have to be as small or portable as yours so I was considering either using a built HV supply, like Emco makes, or “rolling my own” (or letting one of my more EE-inclined teammates do it). It’s important that my measurements are scientifically reliable, this is intended for science after all. What are your thoughts?

    29. Geiger Dude says:

      I have always understood that OLEDs are poor in sunlight. What has been your experience?

    30. Visitor says:

      It seems that Medcom never got around to producing your design. Do you have any information about their reasons for not producing it?

      • bunnie says:

        They are actually well into the process of producing it, it’s just taking them a while to get it through all the final certifications, calibrations, etc. etc.

        Should be up for sale soon.

    31. Very hot device, and very good the idea to use a pancake tube that makes the instrument very sensitive to low levels of radiation, but I think that this choice certainly will very increase the final price. Whereas it should be an open source project and therefore should cost little, some components could be used more economic, such as the Geiger tube could also be used a cheap and good tube made ​​in China whose price does not exceed 40 USD. Also the choice of expensive OLED display that sure does a great scenic effect could be changed to equally functional and more economic alphanumeric lcd display.

    32. Nicholas Tenhue says:

      Fantastic read. I’m doing research on responses to the nuclear disaster in Japan. I’d really love to get a time-frame on each of the iterations so that I can get a better idea of when this all happened in the evolution of RDTN & Safecast.