Input voltage clamp

Tsk, tsk. Snoopy flies a Sopwith 'Camel'.

;)

Cheers

Phil Hobbs

-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510

That's fine, low capacitance and picosecond speed, as long as the average clamp current doesn't pull up the divider voltage.

--

John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

I dunno, walking on hands is more difficult than on your feet. But .. BFY90 (1 GHz NPN) or similar, collector to comparator input, emitter to ground, base connected to that same 5-10V step via resistor, prevents comparator input ging positive during reset.

5 minutes. But I have never used a Pdiode like that, so maybe I overlook something...

Right my mistake, I googled snoopy and spad and found a pic. Eddie Rickenbacker flew a spad... Maybe that's what snoopy *should* have been flying? My reset circuit seemed to be almost working... I then I f'ed something up, not sure what... now it's an oscillator.

Grumble.

George H.

Oh clamp the non-inverting input to ground? That might work.

George H.

The minimum pulse width - about 30nsec - is way too long.

Haven't we all.

Single chip bad solution. There used to be an ECL monostable - the MC10198 - which would go down to 10nsec, but it is long since obsolete.

--
Bill Sloman, Sydney

The Am685-style comparators have a fast-acting input latch that can be used to prevent the input from firing again.

The AD8465 has a latch input that might be used in the same way, and there's a also a variable hysterisis input ...

--
Bill Sloman, Sydney

George Herold

One design 'rule', well it is not a rule but something I always go for if possible, is that the input device must be ground referenced.

This is not always possible, but in this case why not?

• variable | [ ] R1 | C1 |-------------------||-------------- comparator | | |

--- c [ ] optional pulse shaper (diff) / \ NPN b-- | R2

--- e | /// | | [ ] R3 /// /// | SPAD | ------< reset

The SPAD thing works on sudden breakdown? From C1 R2 you get a negative pulse folowed by a positive one if the SPAD 'fires' into the comparator, use the positive one. | |\ ------ |/ |

Reset by discharging the SPAD (no idea if a SPAD reset works like that), This should work with a 74HC123 ?

Having a ground on one side of the SPAD makes circuits easier.

Really gotta get a SPAD to play with, what type number do you use?

On Friday, August 31, 2018 at 4:29:11 PM UTC+10, snipped-for-privacy@nospam.org wro te:

possible,

'fires' into the comparator,

A single photon avalanche diode is an avalanche photodiode that is normally biased a volt or so above the avalanche breakdown voltage.

An incident photon (or thermally excited charge carrier) that injects a cha rge carrier into the junction generates more charge carriers as it crosses the junction.

When the avalanche has built up to a detectable current you have to drop th e voltage across the junction below the level sufficient to support avalanc he multiplication, and keep it low until all the charge carriers have been swept out of the junction - and bit longer to make sure that any short-live d excited states in the junction have decayed back to a less excited state.

Like I said, the minimum pulse width is a bit too long for the application. The internal propagation delay (from trigger edge the start of the output pulse) is also a bit too long.

Tricky. I actually bought a couple some years ago and could probably find t hem if I looked hard enough, but the web is remarkably unwilling to produce a part number for an actual part, as opposed to expensive modules that inc orporates one along with a heap of electronics and a thermo-electric cooler .

--
Bill Sloman, Sydney

PS

may work 4 you. Pulse will have to be high enough to open the transistor. Can be made very fast.

Hmm OK, the LT1016 has a latch too... Maybe I'll try that... I'd like it to latch in the low state and not high... it gets to be a bit of a timing dance.

George H.

Hi Jan, there are some other threads about this. "It's the capacitance, stupid", for one.

and a nice article

the spad is from laser components SAP500

It cost me $250 for one. Talking with sales rep's at the company they are doing big business with self-driving car people and others.. so maybe some day in the future the technology will be cheaper.

You can make a spad from an AND113R and AND114R LED reversed biased to ~25 V. but the efficiency is the pits. (Like there is just one atom doing all the work... the numbers for that crazy idea sorta work... well you have to put in a number for how 'big' the atom is.

George H.

rote:

if possible,

AD 'fires' into the comparator,

,

ly biased a volt or so above the avalanche breakdown voltage. More like 10-20 V these days... efficiency goes up with over voltage till ~30V or so.. see final graph here,

George H.

harge carrier into the junction generates more charge carriers as it crosse s the junction.

the voltage across the junction below the level sufficient to support avala nche multiplication, and keep it low until all the charge carriers have bee n swept out of the junction - and bit longer to make sure that any short-li ved excited states in the junction have decayed back to a less excited stat e.

n. The internal propagation delay (from trigger edge the start of the outpu t pulse) is also a bit too long.

them if I looked hard enough, but the web is remarkably unwilling to produ ce a part number for an actual part, as opposed to expensive modules that i ncorporates one along with a heap of electronics and a thermo-electric cool er.

Thanks...

I don't have enough time today to get distracted, but the other day I got out my Data pulse 100A, and the pulse width wasn't working.

Turned out to be a bad switch.. but it got me looking at the four transistor multi vibrator (fig 3-5) ... there's a more complete schematic at the end of the document. Anyway, someday I want to puzzle out how it works.

George H.

at

oing

?

used to prevent the input from firing again.

685_96687.pdf

ere's a also a variable hysterisis input ...

65.pdf

I've not used the latch input on the LT1016. It seems to act pretty quickly - 2nsec - and might be good enough.

The LT1016 s a quite a bit slower than the 9685 and similar ECL compatible parts, and active quench on a SPAD seems to work better if the avalanche is quenched fast - the longer the junction is supporting an avalanche, the mo re excited states end up there and the worse the after-pulsing (or at leas t that's impression I got from what I read years ago).

Jan Pantelje is enthusiastic about the 74121 monostable, but - as I've obse rved - it isn't really fast enough. I ended up with terminated transmission lines to do fine timing in that sort of gear - mostly lumped element delay lines, but there were one or two lengths of minature coaxial cable. No RG-

405, though I was tempted - the connectors are expensive.

--
Bill Sloman, Sydney

George Herold wrote

Yes I skipped that after a quick look:-)

OK got it, well same idea I had in fig 8 to use a transistor to quench the SPAD. They add some LC and R..

The output voltage of that circuit you are using is very low due to the resistor ratios, you could bias the first transistor in that circuit I showed a bit, but then temperature effects become important, there are simple ways around that too tough, see below.

Yes, and I also searched for that SAP500 on ebay, not there yet :-) no pricing on that site.

Just got handed a nice 10 GHz LNB with PLL locked LO from the postman from ebay,,, Now to design a reference oscillator that will be locked to my Rubidium 10 MHz frequency standard.... PLL chips have not arrived yet...

Datasheet of AND113R says I is 10 uA at Vr 5V, maybe I will try to avalanche some LEDs I have :-) Plenty of those to test..

In the 2 transistor circuit I posted to temperature compensate,, add one transistor Q1

• | [ ] R | k spad a | | c |-------------- b | e R | | |---- R ---- + c ---| .7V | .71 V b--| [ ] R e Q1 | | /// ///

Something like that, but that fig 8 in 1706.06628.pdf why not try it?

Thank you for the links!

I bought some other (cheaper) APD's (avalanche photodiodes) But when biased above the breakdown voltage they didn't show any single photon type behavior.

I tested a bunch of different LED's the above is the only one that works well. I have a bucket load of AND113's I'd be happy to put some in the mail for you. (Or you can buy them from newark... Get the AND114's, less red in the plastic lens.)

Crap Crap crap! (that's triple crap) I see what's wrong with my scheme. If a photon hits my spad while I'm reseting (shorting) the top quench resistor then I'm getting a nice short from my HV to my comparator input... Bang! It's lucky I didn't blow up my spad. (I only put 1nF of C on the HV node.) OK I think my whole reset circuit is headed for the trash can. I'm going to do something else today and ponder it over the weekend. It would be useful if I could think of these things before I found they didn't work on the lab bench.

piglet posted some other low side quench/ reset circuit ideas.. (in one of the other threads) I might try one of those too.

George H.

Yeah, the key is whether the latch is in the input or output section. If the latch is in the input structure, AC-coupling to it works much better than AC hysteresis for pulse stretching--minimum pulse width doesn't depend on what the signal does, and no junk gets kicked out of the inputs. IME output latches always oscillate when you try that.

The LT1016 latch seems to be about halfway, so it might be worth a try.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
https://hobbs-eo.com