fast(ish) one-shot

That lower circuit is a fast one-shot, about 12 ns width. U3 is actually a 2G07. Someone created 2G06 in our library without the bubble, so I used it.

Even the Tiny versions of xx123 are pig slow.

Another trick is to use a flipflop that sets on a rising edge and clears itself through a short delay. An R-L-C makes a pretty good delay for that. Or even a PCB trace. Or another gate or two. There are some 1 ns Tiny flops.

Been there, done that.

Then Motorola introduced the

which could get down to 10nsec, but sadly they made it obsolete a few years ago.

You can make your own monostable with 5 GHz wideband transistors.

Ghiggino, K.P., Phillips, D., and Sloman, A.W. "Nanosecond pulse stretcher",Journal of Physics E: Scientific Instruments, 12, 686-687 (1979).

It wasn't very good, but worked fine where we used it.

You can get packaged delay lines - either active or passive.

The Farnell/Element 14/Newark search function found 52 active parts, mostly from Maxim, but there were a few ON-semiconductor parts too.

I've also used bits of coax cable, but twisted pair would have worked just as well.

An emitter-coupled monostable would have been better (as is mentioned in the note).

Thanks! I've got a faster comparator to try (AD8611) and that may give me a fast enough edge to drive a little cap.

George H.

AD8611 is a single 5 ns comparator that costs $5 in singles. An LVDS receiver, like DS90LV012 maybe, is better and costs under $1. It's not as analog precise as an official comparator, but should be fine for the SPAD detector or a one-shot.

OK, (money doesn't matter) The AD8611 as the first comparator.. Even though I've got monster sized pulses, there's only ~1-200 mV at the comparator input 'cause of the forward biased diode. So I'd like a nice comparator there. But I'll try some of the above line drivers as one shots.

The other idea floating around is to drive the fet-> quench cap directly from the comparator and then use a one shot--- or some other delay to reset the comparator. Mario does something like that here...

George h.

If you do not care about input values/amplitudes, then one might try a one-transistor one-shot.

Two transistor one-shots are a lot more designable.

I've ordered some of the LVDS parts ala JL... they look faster than my spendy AD8611.. and way better than hc14's!

George H.

LVDS is essentially ECL hung between +5V and 0V, rather than 0V and -3.3V. Like ECL it can drive transmission lines at their characteristic impedance.

You don't need a one-shot for the SPAD active quench.

Right, I could directly quench and then have a delayed reset. But I worry about latching...

George H.

I agree, I think. Can't be sure without knowing more about the SPADs, which are poorly characterized. An AC-coupled quench has a potential latchup mode which can be avoided by limiting peak current, which slows things down.

I think I posted a DC-coupled SPAD quench somewhere around here.

George, can you re-post the link to the SPAD data sheet?

'scope shots are here,

I'd have to explain them somewhat. But even with a low quench R (1k ohm) and relatively high over voltage, the spad fires, then conducts, but slowly the current decays until it gets close to zero.. at which point it shuts of randomly. There's other stuff I don't understand... in the last pic (S13.bmp) there is some longer recharge time I don't get.

George H.

We have dealt with Laser Components before. They don't seem to know much about the stuff that they sell. I asked them what the capacitance of a laser diode was, and they didn't know.

arp

de to positive rail. I think R_min is 500 ohm. Well, that's more delay th an

You're going to have two types of feedback, positive (PFB) and negative (NF B), no matter what, and this feedback is derived from either the SPAD curre nt or voltage. Since the voltage is so much larger and a more certain magni tude than current, it is advisable to use voltage. Your comparator should t rip on SPAD voltage dipping below a fixed threshold, and this trip should c ause an immediate PFB to be applied to the comparator which 1) endures for the entire quench process and 2) establishes a new upper threshold for the SPAD voltage at which the comparator turns off. Since this takes time, it a dvisable to introduce a gate delay between application of the PFB and the N FB. The comparator also sets in motion the active quench, which by definiti on induces a NFB on the comparator. Since the PFB has set a new threshold, the NFB due to the quench in process will not affect the comparator output until the SPAD is actually quenched. The comparator can now flip and the SP AD stays quenched.

It varies ridiculously widely depending on bias. A medium sized laser diode can look like 1 ohm in parallel with a few nanofarads.

Cheers

Phil Hobbs

I think the technical people are over in Deutschland, you need to ask auf deutsch. :^) The US is mostly sales people.

George H.

te:

sharp

iode to positive rail. I think R_min is 500 ohm. Well, that's more delay than

NFB), no matter what, and this feedback is derived from either the SPAD cur rent or voltage. Since the voltage is so much larger and a more certain mag nitude than current, it is advisable to use voltage. Your comparator should trip on SPAD voltage dipping below a fixed threshold, and this trip should cause an immediate PFB to be applied to the comparator which 1) endures fo r the entire quench process and 2) establishes a new upper threshold for th e SPAD voltage at which the comparator turns off. Since this takes time, it advisable to introduce a gate delay between application of the PFB and the NFB. The comparator also sets in motion the active quench, which by defini tion induces a NFB on the comparator. Since the PFB has set a new threshold , the NFB due to the quench in process will not affect the comparator outpu t until the SPAD is actually quenched. The comparator can now flip and the SPAD stays quenched.

Right, I think I get all that. With the one-shot I just quench the spad.. with then a passive (RC) reset t o the set bias voltage. I don't have to disable the comparator. But I do th ink about other ways... A fast quench would stop the current flow sooner.. (For high count rates I can only afford so many electrons per event.)

George H.