Single photon avalanche detector (SPAD) difficulties with photometer

[Benschu]

TL;DR Summary

The SPAD and circuitry have been developed over a substantial time. However, the output drops to zero whenever a non-pulsed optical signal is applied.

I have been developing a SPAD based photometer having a 10 ns deadtime, near-zero dark counts, near-zero afterpulsing, and that is temperature independent (within a reasonable range). It works well with photon level pulsed optical signals, (10ns to several microsecond pulse widths at rates up to about 50 kHz), but at higher rates or whenever the illumination is continuous (at any level), the SPAD goes completely blind. Early on I've experienced the same issue with ordinary APDs. The ECL based detector circuitry is happy up to 30 MHz, so not believed to be a factor here.

There is, of course, difficulty observing the SPAD-detector junction because an extra picofarad or two throws measurements completely off.
I am concerned there may be something strange going on there, but at a loss as to how it might be checked. The bias resistors at the SPAD junction are carbon film and I plan to replace them with metal film resistors in the next day or so. Any ideas out there about possible causes for this limitation will be appreciated.
Ben

 

[mfb]

whenever the illumination is continuous (at any level), the SPAD goes completely blind

What does that mean? A very dim continuous illumination should look like individual photons arriving at random times.

Did you test the single photon deadtime experimentally with two pulses separated by x ns?

 

[Benschu]

What does that mean? A very dim continuous illumination should look like individual photons arriving at random times.

Did you test the single photon deadtime experimentally with two pulses separated by x ns?

Hi mfb,
Yes, using short driving pulses (<15ns), an NIR LED doesn't begin emitting until nearly 15 ns. Also, it is not difficult to approximate the number of photons expected from an LED given the power, distance, beam divergence, efficiency, etc.
I have been using photomultipliers for years with routine single-photon detection sensitivity. SPADs and APDs a bit new to me. The problem with SPAD blindness at high count rates has been found. I have been trying to operate too near the SPAD breakdown voltage. So, that problem is mostly solved. Many compromises are required.

The use for these detectors is for optical SETI searching for pulsed laser signals. If one does all the calculations, it is likely that, for my telescope, a received signal with be less than 100 photons and maybe a lot less. So, the work continues.

Ben.