Conundrum: MPPC pixels vs pulse heights

[tim.hbn]

Hi Everyone

I have a small gamma spectrometer which uses a MicroFC-60035 MPPC which has 18980 pixels. The following are two sceenshots from a spectrum generated with it:

Screenshot 1.jpg (23.01 KiB) Viewed 4189 times

Screenshot 2.jpg (78.8 KiB) Viewed 4189 times

The numbers along the bottom of each screenshot are channel numbers.

As you can see, it seems that the MPPC is able to generate as many as 62479 or more different pulse heights even though it only has 18980 pixels. Does anyone here know how this can be the case?

Thank you very much.

Tim

EDIT: In case anyone tries to repeat what I have done, they may not get the same results without the following Becquerel Monitor settings:

Config.jpg (9.51 KiB) Viewed 4188 times

[NuclearPhoenix]

Tim,

that's certainly something software or noise related, I'd guess. What ADC are you using to measure the SiPM pulses?

For example, if you're using a 16-bit ADC with a reference voltage very close to the maximum SiPM output voltage, it would generate results like this just because of noise.

Another thing to double check, does it truly use the full output range of your ADC or MCA? I.e. are there pulses at the very lowest bins and at the highest?

Also possible is that your integration time or general response time is really slow and signals might overlap giving you wrong results. SiPMs are very fast devices after all.

[tim.hbn]

Hi NuclearPhoenix

Thank you very much indeed for your very informative reply.

What ADC are you using to measure the SiPM pulses?

I actually purchased the small gamma spectrometer online so I don't know what ADC it uses.

I think that I can now guess what is going on. As you will have seen from the "EDIT" section at the bottom of my original post in this thread, I have set the "Channel Pitch" in Becquerel Monitor to 0.0016. This means that the spectrum is stretched across many more channels than it would otherwise have been stretched across. This means that theoretically (not taking noise into account), there will be gaps in between the channel bins which are being filed with counts where there are a few empty channel bins which are not being filed with counts. However, if there is some noise in the analogue part of the circuitry then this might add random amounts of height to the pulse heights thus causing the otherwise empty channel bins to be filled.

Does anyone on this forum have any ideas about whether or not the above explanation is correct?

Thank you very much.

[NuclearPhoenix]

Yes, that sounds reasonable. I think you would see the same effect if you tried to use a 24-bit ADC on a 10-bit signal with the same voltage ranges. At least to some extent.

But after all, I can only guess without knowing any detail on the hardware.

[tim.hbn]

Hi NuclearPhoenix

Thank you very much indeed for your reply.

But after all, I can only guess without knowing any detail on the hardware.

The gamma spectrometer is an AtomNano 3. It was made by a Russian eBayer. I purchased it from him in July of 2021 which was a few months before Putin invaded Ukraine. (I think I should point out here that I don't think private individuals, such as the Russian eBayer, should be punished for the acts of their governments.)

You can see details of it here:

[broken link removed - Steven]
"CsI(Tl) 10*10*30 mm USB Gamma spectrometer for PC or Tablet AtomNano 3 Res. 6.5%"

You may need to scroll down past the "Similar sponsored items".

I have opened it up and have taken the following photo:

Atom Nano 3 open.jpg (25.84 KiB) Viewed 4144 times

For scale, the white box that contains everything is 6cm long.

I have seen that the ADC is a Texas Instruments PCM2900C. You can see details of it here:

https://www.ti.com/product/PCM2900C

It is 16-bit.

Thank you very much.