Seeed Looks For Help Developing An Open Source Radiation Detector

Seeed Studios, makers of the Seeeduino and fabricators of small-run PCB orders have put out a call to help develop an open source radiation detector. Will it be of any help to people in the area of Japan that is at risk? We really can’t say. But if you can lend some expertise with this, it can’t hurt. We’ve already seen a simple dosimeter project but this one sounds like it’s more on the level of a DIY Geiger counter. We know it’s possible, but the hacked together unit we saw back in 2007 had very little documentation and used parts that may be hard to come by.

The specific information needed is what type of sensor to use, what supporting circuits should be included, and what method is best to calibrate each unit. There’s a discussion going in the comment thread of that post which should be interesting to read even if you think you don’t have anything to add.

[Thanks Michael]

68 thoughts on “Seeed Looks For Help Developing An Open Source Radiation Detector

  1. It is possible to do this with semiconducting materials, search for >>scintillation counter<< on the web, Wiki does also help
    In german you can buy the book "Experimente mit selbstgebauten Geigerzählern"
    i don´t know if this is sold in other languages.

  2. An open hardware sensor board including a Geiger tube to detect alpha, beta and gamma radiation is currently being developed by Libelium. Once the first prototype is finished it will be sent and tested in the Hackerspace at Tokyo. This new sensor board will be compatible with both Waspmote and Arduino
    platforms. The idea is double, on the one hand, with the Arduino platform people will be be able to have easily running their own detector at home; on the other hand, with the Waspmote platform authorities and will be able to deploy autonomous sensor networks to send the radiation levels from dangerous areas using ZigBee and GPRS

    If you are interested and want to collaborate with this project please post in the forum thread:

  3. @Squant funny I was looking at that maxim example myself. Nice use of amplifiers and a noise reference.

    I’m curious how sensitive a PIN diode is to non light based stimulation.

  4. “- To compare a photo multiplier tube to a photodiode and declare the diode as the more sensitive or better suited for this particular application is silly at best. You either have no idea as to to how either of them function or you’ve experimented with both and failed to consider “basic rules of thumb”, for one or the other. The PMT will always be able to see light that you cannot. Scintillation/PMT detection circuitry is, as far as I know, the best (most efficient and expensive)way to detect ionizing-radiation.”

    Please re-read my comment. I declared photomultipliers as the more sensitive of the two. The most expensive method for detecting it are semiconductor detectors such as HPGe.

  5. Scintillation and semiconductors are two separate ways of detecting radiation. Scintillation is indirect because it first requires the radiation interact with a crystal (which is the detector) that then emits light proportional to the energy of the incident radiation. The photomultiplier or silicon photodiode, avalanche photodiode, etc, then absorbs this light and creates a charge that is then created into a pulse using a charge sensitive preamplifier. A semiconductor detector like a germanium or CdZnTe, directly absorbs the gamma radiation and produces a charge proportional to the incident radiation. The charge is then read by a charge sensitive preamp and passed through a signal processing chain. Comparing scintillators and semiconductors as the best way is kind of a fruitless endeavor since both have obvious advantages over the other, but really depend on the application.

  6. tl;dr but I am in Tokyo at the moment and have not died of radiation poisoning yet. The western media overhyped the situation a lot. There is no danger here, the levels as measured by experts with highly accurate Geiger counters are very low.

    Japan is one of the, if the /the/ most modern country in the world. The Japanese also love data and statistics. Every hour on the TV news they show the exact readings (to five decimal places on occasion if the amount is low enough) for each area.

  7. We could actually use a good wikibook on measuring radioactivity.
    And with everything in it you can find.
    Gamma and neutron counters and dosimeters
    All the possible ways to do this.
    There are quit some things possible with PIN photometers, GaAs diodes, ion chambers made from cans with air in and many other things. This is so valuable, I think we really should make a wikibook from stuff like that.

  8. *Radiation Sensor Board for Arduino RELEASED*

    Quick Overview:
    Detect Alpha, Beta and Gamma radiation integrating any Geiger Tube which works in the range 400V – 1000V and read this levels using Arduino. As well as from the terminal, the radiation levels can be shown using different actuators:

    – Piezo: it allows us to hear the typical “chirp” common in the radioactivity counters
    – Leds: 3 green and 2 red let easily to show low, medium and high levels
    – LCD: it displays the counts per minute (cpm) and the equivalent absorbed energy levels in Servants (µSV/h).

    Read the Tutorial:

  9. I’ve developed a good alpha and beta detector that gives audible clicks using a PIN diode (non-IR filter case clear plastic) feeding a JFET OP amp with as large a feedback resistance as you can find (~ 30.Mohm in my case) followed by another OP amp gain stage (X 100) which has a low impedance emitter follower transistor that drives a ferrite core step-up transformer.

    then a sync trigger which pulse stretches to a driver transistor with a LED and small speaker

    the critical part is that the PIN diode is covered with a phosphor coating consisting of camera viewfinder phosphor (best) or B/W CRT phosphor (almost as good) or color CRT phosphor (fair – only the green dots seem to work well) stuck to clear tape *with the phosphor facing outward* toward the radiation source (the tape blocks alpha particles) this phosphor glows slightly, and with a strong magnifying lens focused on it in the dark you can just make out sparkles.

    the PIN diode has no bias voltage, is AC coupled (solar cell mode)

    this setup produces clean clicks when a radioactive source is near it. the element from a smoke detector (.9 microcurie AM 241) produces a strong clicking sound. this must be used in complete darkness UNLESS you cover the PIN diode with extremely thin foil that blocks all light but allows the alpha particles through.

    ordinary aluminum foil is much too thick, however candy bar wrapper paper backed foil *if you carefully rub off the paper* with water (which loosens the paper enough to remove it, using a flat clean surface like a piece of glass) leaves behind foil thin enough to allow most of the alpha particles through. This foil is very fragile like gold leaf and check for pin holes by holding it up to the light and looking through it.

    If you use this foil, this detector will operate in daytime uncovered. (before this I would place a tin can over it to get a reading)

    the PIN diode barely detects radiation directly but you can’t really separate the pulses out from the noise. the phosphor coating is the key part that makes it work well.

    The phosphor that works best is:

    ZnS:Ag (P11), blue (455 nm), 80 µs decay, low afterglow, for alpha particles and electrons.

    Zinc Sulfide, Silver doped..

    ZnS:Ag+(Zn,Cd)S:Ag (P4) (Zinc Sulfide, Silver doped, Cadmium doped (WARNING TOXIC) is next best, this gives a yellowish glow under UV light (these are all fluorescent BTW)

    P22G ZnS:Cu,Al Green
    P22B ZnS:Ag+Co-on-Al2O3 Blue
    (red doesn’t seem to respond but it’s hard to tell since the dots are so small)

    If you break open a color CRT be very careful and snap the vacuum seal first, otherwise it may implode in your face.

    all these phosphors have an aluminum coating over them, you need to remove this (I used sticky tape gently barely sticking it to the coating then lifting it off exposing the phosphor powder underneath. I then firmly stuck tape to this powder to get phosphor coated tape. don’t press too hard or rub otherwise the phosphor turns black for some reason (maybe graphite under it or maybe the phosphor is damaged by pressure – not sure why) Use an UV light (UV LEDS work fine) to “see” what you’re doing, you want a clean semi-transparent coating, not too thick.

    if you don’t want to build the PIN diode circuit, just use a lens to see the sparkles, people have been doing this at parties since 1903 (Crookes spinthariscope)

    “You can see the flash from a single atom decaying is nothing short of amazing. Your eyes must be totally dark-adapted, which means sitting in a pitch-black room for at least 10 minutes. If you were a dashing young man of science in 1903 and you brought one of these scopes to a dinner party, you would have been rude not to invite the ladies to sit next to you in a dark room. This might account for some of the early popularity of the device.”

    Before I built the PIN diode circuit mine worked well enough to see a glow in the daytime, and immediately see the sparkles at night, but I got tired of having to look at it and wanted to hear clicks. I tried different phosphors like fluorescent tubes, VFD’s etc. No other phosphors or glow-in-the-dark stuff worked at all. The viewfinder phosphor works best I found..

  10. I’ve improved the design so phosphor is no longer needed, using the same basic circuit but with a ‘decapped’ TO-5 style transistor connecting the collector to ground and the base to the preamp input. This gives clean clicks with much lower noise so no sync separator stage is needed. I tried different die sizes (medium power to small signal), larger die have higher noise but pick up more particles, while smaller die have lower noise, and pick up less particles. The same transconductance stage followed by the same X 100 gain stage, but now feeding a small audio amp (LM386) directly.

    Soldered some fine copper mesh over the open end of the TO-5 case since this circuit is very sensitive to EMI.

    With a small speaker and 9 volt battery it’s now portable..

    Holding an Americium 241 element up to it gives a robust clicking under normal indoor light levels, while sunlight saturates it, I noticed when I took it outside, so it’s somewhat sensitive to light, but the prior phosphor design requires absolute darkness to function at all.

    So this is better, cheaper and simpler, requiring no phosphor or light shielding, yet yields a much better output with much lower noise.

  11. I’ve made further improvements to the radiation detector. It’s now micro power CMOS running off 3 volts and only draws 1/2 milliamp total. Using a hex inverter with the first as a linear amp using a 22 M ohm feedback resistor, with the adjacent inverter pin tied to rail to prevent noise pickup or feedback, followed by 2 inverters in series then the last two’s inputs tied together, with one driving a micro speaker using a one shot bipolar transistor and the other the same, driving a small white LED.

    No loading of the chip is permissible as this causes instability and feedback, so both outputs use isolation resistors to the high gain bipolar drivers of 5.6 k ohm minimum, any more loading than this courts instability.

    To maintain low power consumption under all power supply conditions, AC coupled one shots are used to prevent excessive power draw in case the chip’s outputs ever float above ground (which can happen for a few seconds if the supply voltage is falling as the first linear stage tries to balance itself)

    I tested this with a .22 Farad memory cap which will run it for 10-15 minutes but which has a rather high internal resistance allowing the voltage to fall somewhat when radiation is detected, from the extra current draw caused by the light and speaker – without the AC coulpled one shots, it would go into a high current draw mode pulling the voltage down below 3 volts which prevents detection, then the voltage would rise again in a slow cycling loop.

    My intention is solar powered, a small hand crank stepper motor as a generator charging a memory cap or a small ni-cad pack, tiny button cells, a 3 volt lithium coin cell, etc.

    Micro power opens options, avoiding dependency on primary cells which will eventually become unavailable I figure..

    The OP amp and audio amp design above while relatively low power, draws far too much power and too high voltage for solar or other alternative supplies, and while CMOS OP amps are available they also require too high voltage.

    Using CMOS digital the performance actually improves from 5 to 3 volts (one of my specialties in design is micro power so this was the obvious next step)

    These chips are also very common and cheap (4000 series CMOS) while micro power OP amps are a special order part – one of my design goals is always to use common readily available parts.

    Knowing CMOS digital chips have the same required > 100 M ohm input impedance it was only natural to give it a try. Using only a single linear stage allows the use of the more common buffered output chips (although unbuffered draws less power and is optimal – my first attempt using a HCU type hex inverter worked but drew too much power defeating the purpose, the chip I’m using now for 1/2 ma draw is a MC14049UBCP)

  12. I decided to try just a couple of bipolar small signal low noise audio transistors in darlington configuration, stabilized bias by filtered feedback from collector. This works as well and is cheaper than a JFET OP AMP

    Also I found not all JFET OP AMPS work well even though they have the same part # – some are too noisy and you can’t hear the clicks over the noise. (texas instruments I’m pointing at YOU)

    I’ve done further die testing and found some die geometries work much better than others.

    Tested TO-72, TO-39, TO-5 and TO-3’s (dies are sized to the case style)

    TO-72, very sensitive but small detection area.
    TO-39, good sensitivity, good detection area.
    TO-5, fair sensitivity, very good detection area.
    TO-3, poor sensitivity, extremely good detection area.

    TO-39 best overall, Germaniums don’t work at all, even with dies not flooded with silicone grease (almost all are).

    The best silicon dies are high voltage non-glass passivated, (very clean clicks well above background noise) the worst are multi-emitter RF (very poor to not detected at all)

    Some silicon die contact geometries work poorly (it appears the contact metal layer is too thick, blocking alpha particles, and/or covers too much of the silicon) Also checked whether base or emitter connection works best, both tied together gives best signal output.

    Also tried small signal FET, MOSFET (don’t work at all) and SCS (silicon controlled switch – basicly two complementary transistors wired as self-latching, these are rare and hard to obtain, haven’t been manufactered in decades, but thought the die being somewhat darlington-like might offer improved detection, but it doesn’t work any better than the poorest of normal silicons)

    The best overall if you want to buy as new, by part # is 2N3440, high voltage medium power silicon epitaxial planar NPN transistor in jedec TO-39 metal case.

  13. @ st

    no, I don’t.

    I’ve posted above all the info I have, except I’ve tested adding a bias voltage to the emitter, while using the base as the sensing element.

    at a certain voltage range (-4 to -6 volts) this improves output, but above that it adds noise (thermal-shot) which swamps the signal

    I’m still evaluating beta particle detection, as I have no good beta source other than K-40 which is very weak. I am able to detect particles from the sky however, which are around 80 % muons (similar to beta in detection), and about 10 % alpha

    beta is 5000 times weaker than alpha so the clicks are buried in noise, however using computer processing (a de-click plugin, running in monitor mode, running through a VST plugin, using WINAMP’s line-in plugin) allows me to hear them – barely

  14. Recently another inventor figured out how to detect gamma rays with the basic circuit I described above.

    He posted a somewhat sloppy schematic, and a not well optimized selection of parts, but that crude attempt provided me the info I needed to build a good gamma detector. Gamma is almost all the radiation you will see from nuclear contamination – you’re very unlikely to ever see any alpha emitters (which I haven’t while checking rain samples this past year)

    So this gamma detection represents a huge step up in ability to accurately track radioactive contamination.

    He used a discrete low noise audio FET feeding a low quality dual OP amp (LM358), with the second one set as a comparator – all DC coupled.

    So to improve this, I used a dual Bi-JFET OP amp (LF412), and AC coupling.

    This results in greatly improved operation and stability. The gamma rays appear as extremely fast pulses, and in fact it’s only the parasitic capacitance of the PIN diode that stretches these pulses enough to be detectable with audio frequency parts. The pulses are well into the hundreds of megahertz range…

    So using a faster OP amp makes it more sensitive (specifically the OP amp’s slew rate), and using a FET input OP amp eliminates the need for the discrete FET stage. I tried different load resistors from 1/2 meg to 40.meg, and found ~ 10.meg to be the best balance between noise floor and signal output. This is important because the pulses are partially buried in noise, and a very sharp threshold is needed to separate them.
    With DC coupling the whole thing is very sensitive to temperature variations. AC coupling makes it much less so, however with an ordinary comparator stage, tightly regulated voltage must be supplied, at least to the reference and bias inputs of the pre-AMP and comparator.

    So my improvement is to use AC threshold insted which only requires the gain of the pre-AMP to be stable. This means only the feedback resistor needs to be metal 1% with balanced source and load resistors but the source AC bypassed (then the two non-precision resistors cancel each other out for DC drift)

    So that completes the improvements yielding far improved performance, I’ve found.

    PS – I confirmed it’s definitely detecting gamma rays, by blocking the alpha emissions with copper foil, and I’m getting some detection of background radiation, from objects like rocks, plaster, brick, etc.

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.