So more spad stuff, I was trying to trigger the one shot right from the comparator,cless delay. But the comparator output only goes to ~1V of the rails, which works for one shot.. but if another comes right away, then there's issues. (oh this is with an AD8611 as comparator.)
formatting link
well it worked, maybe this post is just to share my rat nest.
formatting link
And here's the 'scope traces,
formatting link
upper brown (1) is the comparator output. Green (2) is the voltage into the quench cap. the ghosty saved trace is the spad node. Most of the delay is RC into that node, I think I can reduce R some.
piglet.. I haven't tried (what I'm calling) piglets ladder. At the moment the fet (2n7000) is fast enough. the delay is in the HC14, which I guess I need to replace. I was looking at JL's LVDT
formatting link
But only 3.6 V supply. 3.3 Vout. (is that enough to turn fet on fast?) There's some slower 5V ones,
Right complimentary outputs. I use one for counting/ looking and the other quenchs
Spad is cathode is at -120V. anode is near ground. Cap is connected to anode.. and reduces reverse bias when it fires. (ghosty ref trace on scope shot.)
Right, but I don't get enough of a voltage pulse from the AD8611.. so I wanted to boost it's voltage.. and then keep the same limits on the HC14.
Hmm, well all the 'hysteria' is to deal with pulses that come right after the quench/ reset. These happen so fast that the coupling cap into the one shot hasn't fully recovered, and then doesn't get low enough to trigger it again.
I don't know if these asap pulses are after-pulsing or twilight things... It doesn't really matter at some level. (I've done no measurements to try and figure out the number of after pulses.)
Suggest something simpler. I guess I could use a comparator to make the one-shot.. that would let me set the voltage limits.. wihtout having to work around the built in 'hc14 ones.
All you want is an approx 20ns quench pulse with 20ns delay after active edge of the SPAD trigger? What's the voltage range worst case max/min limits on the SPAD trigger?
The dashed box "mario magic" is the embellishment you can add later whereby comparator hystersis is temporally widened after pulsing to reduce sensitivity to twilighting.
I don't really understand the SPAD electrical behavior, but it seems to act line a zener diode that forgets to zener until a photon hits.
So,
formatting link
The current I sets/forces an average firing rate. You're going to have a background count rate anyhow, so you may as well control it.
It probably doesn't need R2.
If you are doing entangled photon correlation, use two of these. Compare the hit correlation rate with and without the entangled photons. Play with polarizers.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
The spad voltage is coupled via a shottky diode to the comparator input. Here is at about 1 V over voltage. Brown upper is the comparator input. and green is spad anode. both these are x10 probes coupled in with
2.2 pF caps.. so the voltage scale is about x10 that listed. So min is about 50 mV on comparator.
formatting link
And here is at ~15 V over voltage. (That's about the max I can get to, without more quench voltage.
Hmm, direct quenching. So the fet is normally on, and then I pulse it off to quench? OK. Is the FET capacitance a problem? I guess I don't care about that when the FET is on...
I'll try drawing that the other way.
Why are you worried about twilighting? (From my limited understanding twilighting is when a breakdown happens while the comparator input is disabled, but still active... comparator dead time.) I don't think I have any comparator dead time.
Yeah that's the first circuit I used. Works great. And it may be fast enough for the entangled photon stuff... which has to have a very thin magic x-tal.. .and this means few photons. ~1-10 kHz rates.
But then there are also correlated photon experiments... where I'm just getting two photons at the same time, but not spin entangled. and there I can have much higher count rates... ~100k to 1 M Hz.
Oh a current source feeding it? (I've got a voltage source and current limit.) You need to keep the over voltage on the spad constant. The over voltage determines the dark count and QE of the detector.
Well I think for what you drew R2 is what I'd call the Quench resistor. Without R2 there will be too much current flowing and the zener just stays in the conducting state... or stays there for a long time at least.
I had in mind that C1 would be big, so the servo action would be slow, microseconds to milliseconds, and the average current would be controlled to force the average fire rate to be essentially constant. You could even close a loop on that. So the useful information is in the correlations.
Small C1 is a different mode.
There's R3, too. R2 could be moved below the SPAD to make a voltage divider with R3. Without division the voltage spike could well be too much for a comparator, or for a logic gate!
I never could learn if a SPAD can latch on like an SCR if there's too much current available, or if it just acts like a zener.
Zeners can go over-voltage for a while and then remember to fire, making a noisy relaxation oscillator, which maybe the SPAD does. It's an optically triggered zener.
We're designing some laser drivers for LIDAR, so the SPAD thing is something we should know about too.
I guess that a LIDAR could be a statistical correlation system too.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
I've got 100 uA current limit and then 1000 pF of capacitance at the 'top' of the spad. More than 100uA CW and you fry it. I did some guesstimate for the 1000 pF.. how much energy if it all lets go at once. So far with all my mucking about I haven't blown it up...
There's a forward biased diode separating the spad from comparator. This gives protection and hides the comparator C from the spad... else the HV supply has to supply more charge with each discharge (C*delta V) and that would limit the maximum count rate... (given 100 uA current limit.)
I'm going to say more like a zener. With too much current it just stays on longer... sorta random square waves rather than spikes... It's very much like a zener above the breakdown voltage.
Sure... it triggers by itself too. I've got the overvoltage up to
20V and the dark count is 100k Hz! (well 3/4 dark count, if I turn off the room lights it drops to 80k.
formatting link
I kept hitting single shot till I got one with two pulses.
The top is the comparator input. I made a mistake on the scale... it's set to x1.. so I'm getting ~1V pulses into the comparator... (I'm confused, because I didn't think it could be that big.)
The guys selling me the spad said they are doing lots of Lidar buisness. (The company makes APD's too.)
I don't know lidar... you pulse a laser and time the return pulse? You need an APD 'cause there aren't many photons coming back?
FET capacitance could be problem depending on the part used, to charge the drain and SPAD sense node capacitance fast enough may require a low value pullup resistor - that is why I labelled it "hot"!
A spiffier high voltage source like the capacitive booster you so flatteringly named "piglet's ladder" would generate less heat.
I worried that with AC coupling the quench pulse the problem can be that at the end of the pulse the coupling capacitor recharge actually increases the reverse bias on the SPAD and so proactively encourages after-pulsing events?
Everyone should have a circuit named after them. :^)
Ahh, yeah that doesn't seem to be a problem. If you look here,
formatting link
Green trace is the spad node (anode) and it's firing 1/2 way up to recovering. (75% of those are dark counts.)
There's a 'back dip' on the comparator input, upper trace. I'm not sure where it comes from... and it will give a double trigger at the 'wrong' reference voltage.
Mostly I could use faster. Right, ordering JL's LVDS bit tomorrow.
I see a dip at 10ns and another at 75ns, which one is the bother? Isn't the 10ns dip caused by the quench pulse? The 75ns dip followed by a very slow ramp is weird.
Small diode forward recovery time ought to be very fast, much under 1ns and well into "Larkin-land". Larkin land is alien terriority to me, I still think a micro-second is adventurous!
Is that a small signal schottky? You could try substituting a gold doped PN switch diode like 1N914 or 4148 or even a collector-base junction, I'd expect 'em all to behave differently?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.