Need help with TIA circuit for SiPM sensor

[Aluz]

Hello, I'm designing a transimpedance amplifier for a SiPM (Silicon Photo Multiplier Array Broadcom AFBR-S4N44C013) sensor. Basically I want the output to be 3.2 V max. so I can input it into an Arduino, with output voltages dropping as the brightness drops.

I'm using a 36 V bias voltage from a custom 3.2 V to 36 V booster.

However, the issue I'm having is that the output of the op-amp won't change; it's stuck at 3.2 V no matter the SiPM brightness. If I remove C4 and R4 the SiPM begins to overheat. Also, since the SiPM is operating in reverse bias I believe that the SiPM's C (cathode) should be connected to the 36+ input, so please disregard that error in the schematic. The op-amp is a TLV9101|DBVR.

I've attached the 2 TIA schematic I've tried (buffer part of circuit is currently not being used/connected) and the voltage booster schematic. SiPM data sheet: https://docs.broadcom.com/doc/AFBR-S4N44C013-DS

Any help would be much appreciated.

[paul@yahrprobert.com]

So my understanding of your SiPM is that at 36 V bias it has a lot of gain and will give you around 10 microamps per square mm of dark current. That dark current will charge up C4 until the bias on the SiPM is significantly reduced. So I think you'll want a resistor to ground on the Anode, much bigger than R4 but small enough so that the dark current doesn't raise much DC voltage there to compete with your bias.  So maybe 10k ohm.   Then, on your bias to the + pin of the op amp, the DC level of the output is going to go to the + input's level, so modify your divider to divide the 3.3V  bias by 2, so that you're using more of the opamp's swing (is your opamp capable of running with such a restricted supply range?).

Also, with a high gain sensor make sure when the bias is on that you have very little light on the sensor or you'll overheat it.

[jmelson]

First, C4 blocks the 36 V bias from reverse biasing the SiPM.  Second, U3 is powered from 3.3 v but sees the full 36 V.
Finally, I would increase the value of R3 to at least 10K Ohms to get better filtering.  I used two RC stages in series in one design.  The SiPMs generally leak nanoamps in their usual operating range.
Jon

[Aluz]

Thank you both for your help, much appreciated. I tried removing C4 and then increasing R4 to 10k and grounding it. I also tested a higher supply on the op amp to get more swing as suggested (4.5v), but also tested connecting the + input level to GND as well. But I realized for some reason my output is not changing even if the + input is not connected at all!

The output I'm getting right now is strange, but at least it's a changing signal. It seems like when the light is bright the detector voltage output actually drops, and if the light is too bright the current will shoot up super high and start overheating (which makes sense). So I tried putting it in a dark box and controlling dim light input, it seems like the voltage drops when light gets brighter (even though current goes up slightly) and goes up (but only to .3v) when darker, far from the 3.2v max.

Also jmelson, I'm not currently using the U3 buffer part. Just trying to get the detector part working first.

Wondering what the issue could be? I also increased R3 to 10k but had same behavior.

Thanks again.

[paul@yahrprobert.com]

Well, that's an inverting connection on the amplifier, so the output will go low as more light comes in.  Come to think of it, thats probably why you have a high bias on the + terminal, so that you have the most range available for a negative going  signal.

One more point of caution: any noise on your supply will be sent through the + terminal.  Often people will put a separately regulated supply there, and plenty of capacitance, to eliminate that.