Gamma Spectrometry Forum

Hi All,

The subject of Silicone photomultipliers has come up several times recently and I wasn't sure how to deal with it, so I invested in a tiny CapeSym CsI detector just to have a play with.

The model number is CSI-14X25C-SIPM-T

https://capescint.com/product/macropixe ... 5c-sipm-t/
(No association with the seller.)

Unlike regular PMT's the SiPM only needs around 30 Volts bias, so connecting it directly to the GS would kill it for sure, but I came up with a simple hack 😉

Due to the 1M internal load resistor in the GS you can bridge the output with a 100K resistor, this results in a voltage drop of around 90% so 300V becomes ~30V.

Since the SiPM also draws a bit of current I had to turn up the voltage on the GS to just over 400V but that worked almost perfectly. I say "almost" because the small 100K pull down resistor I used slowly warms up, causing upwards voltage drift. Thinking of ways to prevent or compensate for this, maybe a 30V zener diode in parallel with the resistor could work.

Only three of the pins on the CapeSym are used, the other 4 are exclusively for the connection of a digital temperature probe. I suspect the actual detector will also suffer some drift as the temperature changes.

Other than that, the only change I made was to replace the recommended 200 ohm load resistor from anode to ground with 2K ohm, this has the effect of making the pulse longer.

As you can see, the signal connects to the BNC input just like any 2 wire detector.

Important to note, unlike normal PMT's the pulse from the SiPM is positive and the GS inverts the pulse, so in PRA you need to set the Threshold to a negative LLD, (suggest -0.5)
The resolution of this little detector is great, were it not for the drift it would be in the range 5.3% @ 662, but as it drifted a bit over a 10 minute run it slowly increased to 6.3%.
The next stage would be to make up a little probe handle containing these parts with a socket for the CapeSym. probably with an SHV and a BNC so connection to the GS becomes the same as any other probe.

Ultimately I don't see too much of a barrier to making my own SiPM detectors, other than the raw cost of the actual SiPM chips.

Steven

Here is a follow up to my last post...

After testing the SiPM detector on a breadboard, I built the circuit onto a small circuit board. this improved noise and made it easier to work with.
I solved the voltage drift problem by adding a 30V zener diode in parallel to the 100K resistor, so when the resistor heats up the zener bleeds away the extra current and keeps the voltage stable at 30V.

The end result was a very satisfactory stable spectrum with 5.3% resolution.
Here is another spectrum of Th_232, not bad considering how small the crystal is.
Next I intend to 3D print an enclosure to hold everything together.

Steven

Funny, I was just looking at these, so I'm glad you can share your experiences. Some questions that come to mind if you find any real-world answers:

- I'm wondering what advantage a SiPM has over a PMT in a hobby lab application. I suppose the small size would allow for a smaller shield.

- I'd be interested to see what your shielded background looks like compared to a PMT based detector. I'm curious if the SiPM adds significant noise and what that noise looks like on the spectrum.

- I'm assuming the crystal is the limiting factor for resolution and that a SiPM doesn't make resolution significantly better or worse than it would be with a PMT.

- I'll be interested to hear how much impact temperature has on the SiPM and if there's an easy way to compensate for temperature induced drift. Since it has a temperature sensor built in and temperature compensation is part of their integrated probes, I have to assume it is a meaningful factor.

Only tangentially related, but while "window shopping" I see that, while pricy, the SrI2(Eu) version would seem to be a sweet setup. It appears to have better resolution than LaBr3 at lower energies and doesn't have the internal activity. Though a SrI2(Eu) crystal on a PMT would probably provide the same benefits.

Mike S

I just use three 9-volt batteries in series to power the SiPM with the CapeSym detectors. This is discussed with photos in another thread on this board. See link below -

viewtopic.php?f=14&t=885

Mike,

Mike S wrote: ↑

23 Jul 2022, 19:55

Funny, I was just looking at these, so I'm glad you can share your experiences. Some questions that come to mind if you find any real-world answers:

I am no expert on SiPM's at lest not yet, my initial findings are they are still expensive compared to PMT's but I think it is likely that the prices will soon drop below PMT's, can't see why not, because it should be much cheaper to make than a PMT.

- I'm wondering what advantage a SiPM has over a PMT in a hobby lab application. I suppose the small size would allow for a smaller shield.

The only way to find out is to get the price down so hobbyists can start buying them, once they are out there in garages all over the world, who knows what inventive hobbyists will come up with ? Cant wait to see !

- I'd be interested to see what your shielded background looks like compared to a PMT based detector. I'm curious if the SiPM adds significant noise and what that noise looks like on the spectrum.

My next step is to package it up in a more convenient housing, probably 3D print some kind of enclosure. In regards to the noise, most of it disappeared once I shortened the cables, so I don't think the signal to noise is a big problem. At 30 volts with a 2K load, the pulse is as good as a PMT.

- I'm assuming the crystal is the limiting factor for resolution and that a SiPM doesn't make resolution significantly better or worse than it would be with a PMT.

The crystal is always the limiting factor, but since the SiPM's are still very small one can afford to use a higher quality crystal.

- I'll be interested to hear how much impact temperature has on the SiPM and if there's an easy way to compensate for temperature induced drift. Since it has a temperature sensor built in and temperature compensation is part of their integrated probes, I have to assume it is a meaningful factor.

Don't know, but since they have gone to all the trouble of including a temperature sensor in the Capesym, I guess it might be a problem in some applications. The digital circuit with 3.3V power seems a bit overkill, maybe a simple thermistor or thermocouple could also do the job.

Only tangentially related, but while "window shopping" I see that, while pricy, the SrI2(Eu) version would seem to be a sweet setup. It appears to have better resolution than LaBr3 at lower energies and doesn't have the internal activity. Though a SrI2(Eu) crystal on a PMT would probably provide the same benefits.

Mike S

I agree, but definitely a site for deep pockets.

If I can find a source to buy the raw SiPM's I might try and design some affordable SiPM detectors. Since they only need 30V bias, the entire package including the USB codec can be very compact. Would be quite a fun project.

Steven

Thanks for the information. It will be interesting to see your future results.

Mike S.

Just a brief update on this project.

The tiny Capesym detector is somewhat awkward, so I have been thinking about what to do with it, so today I made up a 3D printed holder for it. I made it like a square cup with double walls. My intention is to fill some lead shot into the 6 mm wall cavity and close it off with a lid. It's almost like a little Marinelli beaker with a 40x40 mm chamber.

Unfortunately I don't have any lead shot here, but I have ordered small bag.

I also attach my circuit diagram, important to note that the 1M Load resistor is already inside the GS-USB-PRO, I am just showing it for clarity.

Steven

SIPM and Scintillation crystals are temperature sensitive, depending on used crystal type. This might be an issue for a portable identifier device or similar application because that instrument should be usable in various temperatures, for lab application it might be not a huge problem.

I've found this file online about it: http://www.ritec.lv/files/publications/RitecRAD2018.pdf

Steven,

I have one of these on order now (in SrI flavor), so I'm doubly interested in your progress. I'm going to try it first with the GS-USB-Pro as the A-D input device but use a separate power supply (or 9V battery stack).

Mike S.