[Swlug] DIY ionising radiation detector -- getting that loop area right down & shielding against EMI

[Rhys Sage]

I was going to work on the Geiger counter yesterday but 'er indoors needed to be taken out in her pickup truck (it has a wheelchair lift). That's always an adventure as it has no functioning shock absorbers, a bad wheel bearing and an electrical fault that seems to have progressed to flattening a battery in just 2 hours. We had to get a jump start! That pretty much burned up my free Monday.

Anyway- I found somebody in California selling 10 x 10m resistors (1/4W) and they claim to have posted them to me with an anticipated delivery date of Thursday this week - Tuesday next week. I was about to order a boost converter from eBay that'd arrive faster and what came up in the picture looked awfully like the board that I assumed was just a Buck converter and had used as a Buck converter. I'll have to play with that (maybe later today) to see if it is a Buck-Boost converter.

I've got some trackboard so I'll try to cram in the components on the trackboard as per the circuit diagram though the sensor assembly and boost converter will be off the board. If nothing else, I can wrap the boards in a plastic bag and then in aluminium foil with just the power wires, sensor wires and data wire extruding. It'll be interesting to test with and without the circuit's tinfoil hat to see which is best.

I don't recall from the PDF which sensor the guy used. The symbol looked a bit like an ionisation tube but 27v is way too low for that. Most Geiger-Muller tubes run at about 300v - pretty much the same as a Xenon tube. Given that the differentiation between dark and presumably single photons from the CsI:Tl crystal will be somewhat challenging, the photo diode probably needs the full 27v. I assume the 10K resistor on the diode is there to prevent the Darlington from over-saturating.

I know what you mean about it being hard to keep up enthusiasm for a project when it takes a long time. This is why I have so many electronic bits and so few completed projects. 2-weeks between idea and the bits arriving is an eternity.

Tom Whitehouse in Swansea market closed then reopened under a new name- same staff, same stall, same stock. I think (reading company balance sheets) that it was a way out of paying taxes. There's now a Makerspace opposite one of the sex shops in the old Knitters and Sewer's World location. I used to buy a lot of components in Tom Whitehouse and before that, Tandy.

Companies seem to go under in different countries at different times. Radio Shack is still around but all the Radio Shack shops closed in the USA, 5 years ago. Tandy (part of Radio Shack) had a shop in Swansea but that closed in probably 1997. C&A is still going just not in Britain. Woolworth died out in the USA in the 70s but it stayed on til 2009 in the UK though it's still going in Australia.

The resistor that blew up on my voltage reducer was a 1/4W and the circuit needed a higher wattage resistor. The ATTiny85 uses 2ma at 5v but the relay used too much current. I've forgotten how much but that was the culprit. As I didn't have any high power resistors I put one of the cheap Buck converters in, used a TIP120 to switch the relay and drove the entire thing off an ATTiny13. That has been running now for an hour a day for the last year.

The clicks on the speaker seem pretty regular which makes me think it's charge-discharge of the capacitor. It's too slow to be anything else. I could put a 2n2222 on the output and run that through my scope to get a pulse frequency.

If I grounded the bus (easy enough), perhaps it would work better as a shield. It is quite difficult to pick up wifi and cellphone signals inside. The further back, where the windows are sheeted over with 14ga steel sheet is particularly difficult. All I'd have to do with be to drive in a ground pin connected to a crocodile clip connected to the bus.

Yes, I'm working in the USA. I own a couple of acres that I'm living on, here. I have some family properties in the UK too which occasionally I need to sort out. Neither country is "best" - they have there own advantage/disadvantages. No 25mph speed limits here except for housing estates and I get to drive my SUV with its 4L v6 engine at 18mpg without being bankrupted by the taxman. On the other hand, healthcare here is terrible but better in the UK.

As usual - not much sleep last night and I worked all morning. If I get time after work and after my midday nap, I'll look at the buck-boost converter.

Rhys Sage







On Monday, 17 March 2025 at 18:12:38 GMT-4, James R. Haigh (+ML.LUG subaddress) <jrhaigh+ml.lug at runbox.com> wrote: 





At Z+0000=2025-03-17Mon12:12:05, Rhys Sage sent:
> This is the latest edition - it's as close as I can get for the moment to the PDF. I've ordered some 2M, 5M and 10M resistors. I don't want to use all my 1M resistors on one circuit. I've got the Buck converter turned off at the moment so it's getting about 9v to the circuit. I ordered a boost converter with the resistors since the circuit on the PDF needs a 9v to 27v boost converter. I checked and the BPX61 should be OK up to 32v. Being the cheaparse that I am, I ordered everything from AliExpress so it'll be a week or more before it gets here.

    Seems like the requirement of voltages as high as 27v might be unavoidable due to the physical phenomena involved, but I am still not sure about that -- one of the articles that Alan found seems to imply that 9v is sufficient.  It might be something to try while you are waiting for the boost convertor.

    When the boost convertor does arrive and you set it to 27v, you'll find that the magic genie that you seem to have become well acquainted with becomes a lot less forgiving, even developing a bit of a temper! :-P

> Here's a link to a picture of the circuit:
> https://yt3.ggpht.com/2siN7_k3wjj-HK4c5i7p705Ojs-Je9_DBwp1Ys70hI-VJ1sOlXn_6fKlls5uk5tUSidIzttkuzsJOg=s4096-rw-nd-v1
> 
> I dislike he webp stuff but scaling works well. The actual image is 8mp so i think 4096 is about right. This is the same image that's on my latest YouTube post.

    Got it. :-)  Thank you.

    So first thing I notice is that you have switched to using adjacent power rails, which is good -- both rails are together, they are the near-side rails of the breadboard.  After that, things get messy again with lots of loop area.  Expected for a breadboard, but the question I am wondering all along is even whether this circuit is "breadboardable".  Some circuits are not breadboardable, and a sensitive circuit like this that needs low noise and high signal bandwidth at the 1st stage is almost definitely a circuit that will misbehave on a breadboard.

    That said, there are still a lot of small changes that would give the breadboard a fighting chance.

    You could put a low-ESR electrolytic capacitor across the near-side breadboard rails.  This will truncate the current-change loops of all loops that use power, making the loop area of the supply mostly irrelevant.

    You could twist and tape together the brown wire to the pale-pink wire.  Then again, this on the output side and doesn't matter so much for now.

    You could remove the floppy black jumper and move the transistor into the middle, connecting its collector to breadboard node 16.

    The sensitive input signal coming from the photodiode on the yellow/green pair, would be better connecting directly to an adjacent pair of short nodes, not long rails, and avoiding the floppy jumpers between the rails and the nodes.  If they connect much close to where used, you will eliminate several square centimetres of loop area, which will reduce noise pickup and inductance.

    Replace all floppy jumpers with the flat sort of jumpers.  Bunch stuff up as much as possible.  You could easily use only half of that breadboard, or even a quarter or less.  When you bunch stuff up, you minimise loop area for free.  Some people find it less convenient, but that convenience comes back to bite them when they attempt any RF stuff like this.

> Somewhere on my YouTube short videos I tested a circuit in action that had a voltage reducer that I'd designed. It worked really well in my microcontroller lab. It gave me a safe 5v for my ATTiny85 and 12v for my big relay.

    Nice. :-)

> When I put it in place and switched it on, it was all perfectly fine then I saw a trickle of smoke coming from the resistor that I'd paired with a zener.

    What was the diagnosis of this fault?

> Who knew resistors were also perfect fuses?

    They're not.  Fuses will reliably break the circuit, whereas many other components will actually fuse into a conductor, allowing more current through than when they were just a resistor or a semiconductor.  Diodes are notorious for failing to a short circuit, so watch-out for that.

> I put a small speaker on the output line and it gave a faint and consistent clicking noise. I have tried simply unplugging the photodiode and nothing changes with the clicking so I am assuming that the 100nf capacitor is charging then discharging.

    It has to be -- because its voltage is complementary with the buzzer (both add to the supply voltage) -- but why?  My guess is that the buzzer is emitting as clicks the zero-crossings of the amplification of the input-side noise.

> I have put the transistor on my scope and used the test function. It seemed to test OK. I'm encouraged by the clicking because it's clearly doing something even if it's not the right thing and because nothing yet had caught fire or gone boom. Fortunately things catching fire or going boom are very rare.

    At least this is indicating some kind of sensitivity, which does sound good.  But I'm wondering...  Are the clicks in a regular rhythm of equal intervals, or are they sporadic?  How frequent are they?

> My thought is that not having a 10M resistor might be the reason I'm hearing the clicking noise. I'm going to get the boost converter and 10M resistor in place before anything else to try to get as close as possible to the circuit diagram.

    Then again, I realise that given that you are running it at a lower voltage, you won't need as much resistance there -- until you get 27v to it, then it being 10MΩ is more important again to slow the decay rate enough.

    I'm not familiar with buzzers.  I'm wondering, if you give the buzzer some low-enough voltage, does it click slowly anyway?  Can you emulate its observed behaviour just by giving it a constant voltage?  If so, then you are using it more like an audible voltmeter, perhaps.  Why not just use a voltmeter?

> My oscilloscope is quite limited in what it does. It's a good beginner's scope and I'm happy with what it can do at the moment. I didn't want to buy one that was too complex or too expensive. I like that this one charges off a USB port. My microcontroller lab is limited in power - I have a 12v power supply and that's it. The lighting is 12v too. The actual level is something like 14.2 - 13.7v. Nearer to the charge controller the voltage is fractionally higher. The wiring system is multiple lengths of 12 - 14AWG connected with crimp spade connectors.

    I'm planing to do a similar thing for off-grid usage.  I already put together a 12v system for my 12ft bike trailer as well, a few years ago, which worked well, but now the trailer prototype itself is in disrepair after a couple of hundred miles of use on rural back lanes.

> Anyway, the scope is a DSO-TC3 from AliExpress.
> 
> I do notice the circuit is supposed to be shielded. That will come in later. I can always get an aluminium box to put it all in, once the circuit is working.

    Do you think it will work without the shielding?  It would be a nice surprise, but you are trying to amplify a signal that is smaller than the noise that a breadboard circuit will typically receive to much greater proportion, so you're going to make things harder by ignoring the elephant in the room.  Noise getting into your circuit is a really big problem in this project.  You're trying to detect alpha particles and high-energy electrons.  These are very subtle signals.  Inductive pickups of radio waves consisting of much larger energies is very likely to get in the way here a lot.

> It is kinda shielded at the moment as my microcontroller lab is not connected to mains electricity

    That does help a lot, actually.  My friend Simon's lab is an EMI nightmare.  Really cool, though.  But absence of mains is probably not enough given that even your local radio transmissions might have more impact on your breadboard loop antennae than beta particles would have on your photodiode.

> and is inside a steel body converted bus.

    That's not shielding.  I was surprised to discover a few years ago that my Nokia 3310 gets decent signal inside a solid steel shipping container with the door closed.  Why?  Isn't it meant to be a Faraday Cage?  Since then, I watched several YouTube videos on the matter, and it turns-out that Faraday Cages and electrical shielding are a lot more tricky than at first meet the eye (based upon a high-school level of understanding of the topic).

    From what I can gather, the biggest flaw in a shipping container's shielding is that its doors are not electrically connected all the way around.  Aligned polarisations of electromagnetism can pass through the slits for a start, but also, the poor electrical connection at the hinges means that the container doors are free to be at a different electrical potential to the body of the container.  Radio waves on one side can induce a changing potential difference between the doors and the body, then that changing potential difference across the container drives a changing current and magnetic field which basically retransmits the radio wave on the other side (outside/inside), allowing radio waves outside to get inside, and vice versa.

    Essentially the container itself becomes an antenna, not a shield.  And this is true of any half-arse attempt at a Faraday Cage -- if it is compromised, it becomes an antenna, which is the exact opposite of what you actually want.  So people who used to wear foil hats to supposedly protect their brains from EMI, apparently may have been increasing the gain of that reception by placing not a Faraday Cage on their heads but an antenna.

    Your converted bus is more likely to be a bit of an antenna than a shield.  Do you get mobile phone coverage outside your converted bus but not inside it?  More likely, it will be about the same.  Can you communicate via a walkie-talkie or DECT phone between inside and outside?  Can you get Wi-Fi signal from the house?  Any of these things reveal that it is not behaving as a Faraday Cage.

    Compare this with placing a mobile phone inside an unplugged microwave, which is actually designed to be a very good Faraday Cage.  Even my Nokia loses signal in a microwave.

    That said, Faraday Cages assume perfect insulators.  In practice, most Faraday Cages can only attenuate the EMI, albeit significantly.  They can attenuate by many orders of magnitude, but some EMI still gets through.  A Faraday Cage made out of a poor conductor would not be such a good shield.  Depending on what material, and whether it is actually an otherwise complete Faraday Cage avoiding the big flaws mentioned above, a Faraday Cage made of an inferior conductor may still be good enough for most reasons.  I guess steel would be sufficient, so I don't think it was the material that compromised the shielding of a steel container, I think it was the gaps around the doors lacking much electrical contact and forming an antenna with the rest of the container.

> I can't pick up much wifi nor phone signal inside.

    Hmm.  What about right outside it?

> It's not a Faraday cage but it's close to it.

    Does it not have any windows?

> I suppose I could always put the circuit in a plastic box and wrap the box in aluminium foil.

    One of the videos I watched on this tried the wrapping in foil technique and it acted more like an antenna, because of poor electrical connection between the piece of foil.  It literally needs to be a complete cage of electrical connection for it to work properly as a shield and not an antenna.  Holes smaller than radio wavelengths are fine, but slits leak polarised radio waves, and completely electrically-disconnected pieces are more likely to be antennae rather than shields.

> The tritium won't be awfully radioactive. It's enough though to be detectable over background. My understanding is it will give maybe 6 detectable events a minute. I remember I almost bought a tritium torch some 30 years ago. I had no particular use for an expensive torch at the time and didn't but by now the tritium would have decayed enough to be useless. It has something like a 25 year half life.

    I'm not sure what a tritium torch is, but surely being just under half strength would not make it completely useless.

> I don't know whether the power lines that I run my microcontroller lab from carry any electrical noise. If they do, it's coming from my Renogy Rover 30 MPPT unit. The only other thing connected aside from my LED light strip is my door controller (which I do aim to replace with something home-brew).

    There will be some noise from that, but like I say, supply capacitors will really help with this.  You can add as many as you want in parallel, here.  I sometimes see a small, high-speed capacitor in parallel with a large electrolytic on a power supply, if it really matters to keep it as clean as possible.

> As a thought, my microconroller lab is parked underneath a power line.

    Ahah, that explains why you got a hum when you are actually in a lab absent of mains.  Unfortunately being this close to an overhead line means that your lab is not free of the associated EMI that causes the 50Hz hum.

> That should be a 60hz signal.

    Only in the USA and some other countries.  Over hear it should be 50Hz.  It showed up on oscilloscopes sometimes when something comes lose or goes open-circuit and start being an antenna and receiving that hum.  Issues with common grounds can also introduce the hum into your circuit, but that won't be the case here.

> I'm not sure of the voltage. It could be 120v but I know I'm about 20 feet from a 16KV transformer supplying 2-phase power.

    The loop antennae that your jumpers make on your breadboard are enough to receive hum from this at a stronger level than the signal that you are trying to hone-in on.  It'll definitely help to get your most sensitive loop areas down to just a few square centimetres.

> My AliExpress order was made last night and has not yet shipped. It can take between a week and two weeks to get here from China. I could have ordered from DigiKey or somebody like that and had it in a week at ten times the cost. Sadly, Radio Shack went bust here, 5 years ago.

    5 years ago?  In the UK?  I did not realise they were still going here.  I wish I knew this like 10 years ago.  I'd have moved close to one and happily kept them in business! :-D

    I was sad enough when my local Maplin closed, and it has never been the same since.  Maplin was no Radio Shack or Tandy, but it served a purpose for me.

> Tandy in the UK went bust 30 years ago.

    That's more like what I thought Radio Shack did.  I had only ever heard of Tandy and Radio Shack from an era before I was born.  I didn't realise that in some places, Radio Shack lasted another quarter of a century, and I could have actually visited one if I had known about it.  Wow!  Er...  Crap!  That's sad.

    Wait, I thought Radio Shack was in the USA, anyway.  Over here they were Tandy, which means that they never were here as Radio Shack in the first place.  Which means that you must have moved to another country, one that had Radio Shack in branded as itself.

> I used to go to Tom Whitehouse in Swansea market for electronic bits

    Still around?

> but there's nothing that I've found here in Lexington. I don't think there's anything in Columbia either. There might be in Charlotte but that's 120 miles away. I don't mind driving but as I drive every day for work, I don't drive if I don't absolutely have to.

    *penny drops*  You're in the USA, aren't you!  X-D  Okay, that explains a lot of references which seemed odd.  Radio Shack...  60Hz...  American buildings built on sand...  Okay.  I get it now.  You live across the pond.

    So anyway, 60Hz doesn't change much in terms of EMI.  You still need to get that loop area right down, and your shielding needs to be electrically-connected in order to avoid it actually becoming an antenna.

    Also, given that you are in the same country as W2AEW, why not meet-up sometime?  He goes to amateur radio events, iirc..  W2AEW is his callsign and YouTube channel handle.  Alan Wolke is his name.

    Hmm, I was also thinking of sending you an assortment of components when you mentioned LM393 -- I have loads of LM392's.  Probably not the right choice in this project, but generally very useful.  I have plenty of several other components that you seem to be missing as well.  I was going to put an assortment together when I get over my support transition and support time being saturated on a backlog of problems, but by the time I have organised the faff of a postal thing, you'll already have everything you need from AliExpress, so nevermind.  Sending international is an extra faff and expense anyway.  There's lower-hanging fruit ways to help-out that doesn't run into the things that I am terrible at.

    I hope you manage to get your loop area down.  I guess we have to wait another week now to see what happens next, damn it.  I'm finding this a bit slow now to sustain my interest in writing answers, especially seeing as it feels like I have repeated myself a lot, especially about the loop area and EMI, and yet it doesn't seem to be the answers you are looking for.  Are they rhetorical questions that you keep asking about why it is misbehaving?  Anyway, it is still a really cool project and I hope it goes well.  Good luck! :-)


Kind regards,
James.
-- 
Wealth doesn't bring happiness, but poverty brings sadness.
Sent from Debian with Claws Mail, using email subaddressing as an alternative to error-prone heuristical spam filtering.