Hysterical* power supply.

So the wiggle (at 75 ns) is some ~Q=2 step response in my one-shot->fet. It seems mostly to be the late, turning off edge. With a longer pulse width I can see the whole ugly thing. With shorter pulses it's buried in the R_qench * C_spad-node response. And I see the tail peeking through.

(my layout is rather sprawling. Several inches of lead wire inductance.)

George H.

Whose SPAD are you using?

From what I gather from the vague crappy descriptions from the manufacturer s, the SPAD diode has a breakdown voltage VBR, the SPAD is biased at someth ing "much" greater than VBR to achieve a certain single photon avalanche se nsitivity. When the SPAD avalanches, the terminal voltage then drops from " much" greater than VBR to VBR plus a "small" differential required to susta in the magnitude of the avalanche current, said current varying "greatly." The diode is then "quenched" (turned off) by either reducing the terminal v oltage or stealing its avalanche current, take your pick, typically this ta kes something on the order of 10ns. The biasing and detector threshold are usually adjustable to achieve a compromise between missed photon count and false avalanche count caused by noise triggering. It looks like 1 in 10,000 is commonly achieved.

SAP 500 Laser components

George H.

ay,

ers, the SPAD diode has a breakdown voltage VBR, the SPAD is biased at some thing "much" greater than VBR to achieve a certain single photon avalanche sensitivity. When the SPAD avalanches, the terminal voltage then drops from "much" greater than VBR to VBR plus a "small" differential required to sus tain the magnitude of the avalanche current, said current varying "greatly. " The diode is then "quenched" (turned off) by either reducing the terminal voltage or stealing its avalanche current, take your pick, typically this takes something on the order of 10ns. The biasing and detector threshold ar e usually adjustable to achieve a compromise between missed photon count an d false avalanche count caused by noise triggering. It looks like 1 in 10,0

The big advantage (over a pmt) is decent QE at 800 nm (down converted from ~400 laser diode.)

George H.

Looks like their self-quenching fig. 7a is messed up. The capacitor should attach between COM node and the junction of RS and APD cathode.

No that circuit is fine. You get a little stray C on the Rs- spad node and that gives a bit more signal, but adding C to that node just lengthens the reset time.

George H.

It's actually an avalanche diode that gets taken out of it's avalanche regime whenever it happens to avalanche (and the avalanche can be initiated by a photon).

A SPAD is just an avalanche photodiode installed in a circuit that kills the avalanche whenever it happens.

Most "Zener" didoes are avalanche diodes - the Zener mechanisn only dominates for knee voltages lower than about 5V.