Help with the design of a CSA

[Cai Arcos]

Hello:

I'm trying to use a photodiode for performing low energy gamma spectroscopy (similar to this: https://physicsopenlab.org/2017/06/22/x ... hotodiode/), and for that I'm trying to design and simulate a CSA around a low noise op-amp. I attach the schematic as an image.

It consists of an integrating stage and then a single order low/high pass with the same time constant of 1 us (according to what I've read, this minimizes noise for this set-up) with an amplification of 10.The noise itself (which I also attach) is inside what I expected.

However, what surprised me is that the amplification factor (mV/keV) does not seem constant, and decreases the higher the charge is. I'm injecting pulses with a duration of 1us and amplitude of 1nA (6,25keV), 5nA (31,25keV), 10nA (62,5keV) and 15nA (93,75keV), and for each of those I get sensitivities of 0,6455mV/keV, 0,441mV/keV, 0,397mV/keV and 0,408mV/keV. I'm at a loss because the time constant of the integrating stage is much higher than the duration of the pulse, and thus no ballistic loss should occur.

Any help is much appreciated:
Cai

[Joe Gayo]

One issue:
You are not injecting charge correctly in the simulator. You need an AC coupling capacitor. Though I would do it differently all together so that it can be replicated easily on the bench.

[Cai Arcos]

Joe:

I'm using the set up shown below. In the simulation I used the equivalent circuit. It comes from here: https://www-physics.lbl.gov/~spieler/He ... index.html.
The entire page of Spieler is an amazing resource

[Joe Gayo]

Ok, then you should know everything you need(except for an understanding of of how ltspice works or the macro models). I regret trying to help. Good luck.

[Cai Arcos]

It was not my intention to be rude. English is not my first language and often write in short sentences or similar because it is easier. I only tried to explain why I had made the circuit the way it was. Because I have very little experience I tried to copy what seemed a reputable source. I very obviously need help in understanding why this doesn't work the way it is supposes.

[Richard Hull]

I guess I can see the photo diode (PIN) as a simple detector, but not as a truly viable or reliable energy indicator for gammas at room temperatures. Nice effort though.

Richard Hull

[Cai Arcos]

Hi:

After some trial and error, I determined that the problem was caused by R3 (the pole-zero cancellation resistor). When I removed it the undershoot became apparent (to combat this, I decreased sensitivity and increased the integrator time constant by using a 2pF capacitor), but the linearity was as expected. I'm quite puzzled at this, and I would be grateful if someone could explain to me why this happened. In none of the nuclear electronics textbooks I've read is it mentioned that pole zero cancellation cause a degradation in linearity or anything similar. Is this some problem with the simulator itself? Am I missing something obvious or the motive is more obscure?

Richard:

While my little project is little more than a toy to help me get experience, there have been serious efforts to repurpose PIN photodiodes as low energy (<80/100keV) gamma, beta or alpha/proton detectors.
See for instance "The perfomance of low-cost commercial photodiodes for charged particle and XRay spectrometry", "Response of PIN Diodes as room temperature photon detectors" (by the same team), "Nuclear Spectroscopy with Si PIN diode detectors at room temperature" and many more!

[Richard Hull]

I have known of this PIN diode feature for some years and monkeyed with the effort myself back in the 80's. It came to little or nothing compared to normal accepted nuclear detection devices that I had on hand. In the end, it was nothing worth writing home to mom about. A momentary amusement, followed by blunted enthusiasm, cast on the trash heap of "been there done that".... Didn't like the experience or the result. Admittedly I was grossly under-impressed by my shot at it. Perhaps I just got bored by early underwhelming results and just didn't push the effort, feeling my time might be better spent chasing other rainbows. All the best with your effort with PIN diode detection, of course.

Richard Hull

[John Futter]

I have done this at work using Hamamatsu pin diodes to5 cases that I turned off the lids.
The bugger is the hama diodes are light sensitive the ortec ssds put the light sensitive layer at the back metalisation
yes the hama diodes did work but no where as well as the ortec ssd's this with a one on one setup in one of our detection chambers.

to stop a couple of photons getting in to the detectors you have to go to a lot of trouble, light bounces around several 90 degree bends. Also the depletion layer is thinner so the detection layer is thinner.

My gut feeling is even the ortec ssd's at US$700 are not good enough for fusor use let alone a good PIN photodiode that is at least an order of magnitude behind the Ortec ssd