First Test of SiPm Detector

[sgt_bear]

Hello Folks,

I've built myself a small SiPm Detector with a crystal CsI(Ti).

SiliconePhotomultipiers basically work the same as PMT Tubes, but are alot smaller (the size of a photo diode) and do not require high voltage. They run between 20-30V DC which can be generated from low voltage (like usb or Battery) and the electronics do take a small space. SiPms come in different sizes, usually 1x1, 3x3 and 6x6 mm, also arrays are available which are combinations of single sipm's on a pcb board. The only disadantage is their size limitation, to operate on big crystals usually arrays of multiple sipms are required, and these are currently very pricey (~400$ for a 2x2 6mm2 Array).

For my experiment I've chosen the MICROFC-60035 from SensL/Onsemi which already comes assembled on a pcb board https://uk.farnell.com/on-semiconductor ... dp/2949118
Sipms are very difficult to solder at home since they have only leads on the underside (BGA Device) and are sensitive to temperature and chemicals, so i decided to go with the eval board that already has the Sipm assembled.

The used crystal is a 8.5x8.5x30mm CsI(Ti). To make the detector robust i printed a little enclosure to put in the PCB and crystal. The crystal I've got from luuk, and he also installed the crystal inside the holder.

For the Readout i made a small pcb with the following components:
- Power Supply (based on LT3482, Analog Devices, APD Power Supply IC)
- Micro controller (STM32G431, ST Microelectronics, USB Connection and DAC for controlling Bias Voltage)
- Signal Processing (copied from physicsopenlab.org)
- Auxiliary Parts

Sipm's do not like hard power on, so i use the digital-to-analog converter of the micro controller to ramp-up the bias voltage, and also control the voltage. The voltage can be controlled from 24 to 32V and will ramp up from zero to the set voltage in ~10 Seconds.

Images of the Detector (it was wrapped in black tape afterwards) and Control PCB:

detector.png

Control_Pcb.png

the detector was tested by luuk with a resolution of 6.4% for Cs-137. My tests with Theremino resulted in 6.1%, but since i've have no knowledge how to set the parameters of the SiPm and theremino usually "lies" a bit about resolution, I'm sure this is also 6.4% or close to it. For a quick view, i also run pra, but i did not do any calibration or settings for the moment. Here are the results:

SiPm_Theremino.png

SiPm_PRA.png

The next step will be to try building a different readout, which consist of a peak-and-hold detector, so the MCU can directly sample the data. Also i will experiment on a "playback" feature which will use the dac to emulate recorded pulses to PRA and Theremino. Also i will do some effort to make this system "portable" and might even update to a 2x2 Array of the same Sipm to use with a bigger crysta

[GigaBecquerel]

Hi Jonathan,

I really like your detector!
The result is very good, can you guesstimate a detection efficiency, what was your Cs activity?

[Sesselmann]

Jonathan,

Very nice project, thanks for sharing, also on my list of things to do.

You have three cables coming from the evaluation board, does this mean that there is a coupling on the board?

Steven

[Conor Whyte]

Wow, that is super cool! Very good efficiency for that tiny a crystal. I wonder what would happen you coupled the output to a RedPitaya?

[luuk]

Hi Jonathan,
It turned out really nice and it looks like all is working great, I like the design of the pcb.
Resolution is maybe a little to good but so you should check linearity that could give problems.
Luuk

[sgt_bear]

The result is very good, can you guesstimate a detection efficiency, what was your Cs activity?

The aciticity is 0.25uCi +-20%. The Runtime was 2642second, with 350670 counts and 132.8cps. Since the backround is around 10cps, i would guess the true counts of the Cs-137 source is 325k Counts

You have three cables coming from the evaluation board, does this mean that there is a coupling on the board?

No coupling. The SiPm evaluation pcb contains a "fast out" which is directly connected inside the SIPM and gives incredibly fast pulses, much faster than the energy pule (which is clamped between the restistor on +Vbias and the sipm itself, just like the PMT connections are.

The brown wire is the fast-out. It does not lead anywhere, i just soldered it to the pins so it does not fly around

[Polon]

Jonathan,

does the output voltage changes from the power supply accurately correct the SiPM temperature coefficient?

[sgt_bear]

Currently it does not, but since this is my personal "eval board" i will experiment with it and then create a final design ;)

[Polon]

I think that a good solution would be to place the detector in a thermally insulated casing and stabilize its operating temperature with a Peltier cell.