LED on Photodiode step response

On a sunny day (Wed, 13 Apr 2011 05:55:39 -0400) it happened JW wrote in :

Yes, works in NewsFleX (Linux) too.

The only thing worse than the Usenet pussy you responded to is your Usenet pussy response.

The 'cruft' that needs trimming is you pussified, brainless fucktards.

Stupid children like you three should not be allowed on computers.

There should be an aptitude requirement. None of you three would qualify. It doesn't get much more pathetic than this stupid shit.

On a sunny day (Wed, 13 Apr 2011 04:13:04 -0700) it happened Bart! wrote in :

Then why do you write it?

English not your first language, dumbfuck?

On a sunny day (Wed, 13 Apr 2011 04:51:23 -0700) it happened Bart! wrote in :

And now you write shit again, why bother? You know nothing better?

Shaddup already, Dimmy.

rward

t.

hms

wire

up

e of

ces

ook'

if I

uoted text -

Hmm, Well it would give me a nice light pulse. I could control the width of the pulse by the rotation speed and distance. Optical choppers, or shutters are going to be much slower. At least the ones I know of.

George H.

I had an order for some photodiodes held up for a week and a half because I made the mistake of dealing with the US office of a European company.

We do have the clearances and documentation necessary to deal with stuff that's actually a real threat (ie. on the munitions list), but it sucks to have to hand over proprietary commercial information to foreigners for no good reason.

If it blows up by itself it might actually be something worth tracking.

Sounds good to me.

ward

ms

ire

of

es

ok'

f I

I looked at the pulse response too. (I kinda had less faith in getting a sharp pulse from an LED than a sharp step.) It's hard to see any sign of the long tail in the pulse response... I'm not sure if that's becasue it's not there or is the pulse response less sensitive to seeing 'low frequency' artifacts for some math reason. (You have to take a derivative to go from the step to the pulse... so that sounds like it reduces the low frequency signal by more... Is that correct?)

George H.

lly

of

ckly

on

ut

Thanks John, that looks like a 'normal' laser diode to me. (I mean don't all laser gain medium have to live in a cavity?)

Yup lots of cool things, I wish I had more time to learn about and play with them.

George H.

Hi John, I'm not sure what equalize it out means. But if there is an issue with a 'slow' response from the PD then that throws a bit of a monkey wrench into the shot noise measurement. The 'slow' electrons are going to add to the average current, but they only contribute to the shot noise up to a frequency of 1/(slow time).

George H.

You'd get the step response by integrating the impulse response. And there are all sorts of lasers that will give you a clean ps or fs wide light impulse, so that becomes the surrogate for a harder to make reference-step-flat pulse.

Just apply money.

John

I meant that you can add an RC circuit somewhere in the amplifier chain that will flatten your droopy step response.

John

That's an interesting question. My first guess is that you'll get the shot noise of the instantaneous photocurrent at all frequencies, regardless of the mechanism by which it gets there. Diffusion currents have full shot noise, so the same rolloff that suppresses the one shot noise contribution will add back the same amount of noise due to diffusion. I wouldn't swear to that absolutely, but it seems pretty plausible to me.

Cheers

Phil Hobbs

We're just getting the 'mericans back for making us promise not to build a biological weapon every time we order a reel of resistors from Digikey.

if

t- Hide quoted text -

Hmm, you mean I'm wrong yet again? I'll admit I don't have a good picture of how an e-h pair inside a diode gets turned into a pulse in the external circuit. I imagine them created and then pulled apart by the electric field. (Ignoring the diffusion electrons for the moment.) Does the outside circuit "see" this process. (Image charges on the conducting end plates of the diode that get larger till the e or h annihilates it.) That kinda makes sense. So the pulse lasts as long as it takes the pair to drift across the depletion region. Now to the charges that have to diffuse first. There=92s not much happening as they diffuse through the neutral regions. There=92s not much charge separation then... and it=92s only when the e or h reaches the depletion region that there is charge separation and any current pulse.

Is that how you imagine it Phil? In which case I must agree. No reduction in the shot noise... That=92s darn interesting. I will say that the data I=92ve taken show no change (to 1% or less) of the full bandwidth shot noise, if I change from a white LED to an IR LED.

George H.

I've never thought about it before, and I don't know the answer for sure, which is why I think it's an interesting point.

One way of looking at it heuristically is that there's nothing you can do to a Poisson process to make its variance any worse (except apply gain, but that doesn't count). Thus a stochastic detection event followed by stochastic diffusion will result in no more than full shot noise.

The diffusion current has exactly full shot noise--in fact it leads to a fairly pretty derivation of Johnson noise for PN junctions at zero bias. The Johnson noise is numerically equal to the RMS sum of the shot noise of the forward and reverse diffusion currents. That's also why the differential resistance of a diode has an equivalent noise temperature of T_J/2--the reverse current goes to zilch, so you only see the forward current's contribution.

So given that you can't make it worse than full shot noise, and both mechanisms supply shot noise on their own, I would expect that the noise would be the shot noise of the instantaneous short-circuit current of the PD.

I can't see how it would get better in the middle someplace--with a photocurrent there's no negative feedback mechanism the way there is when you put an emitter resistor on a BJT.

So I'd lay probably 4 to 1 in favour of it being exactly full shot noise.

Cheers

Phil Hobbs

e if

ugh

an

he

nts

as

No

t- Hide quoted text -

OK I'm liking that. In the pn diode the diffusing charge carrier doesn't 'do' anything until it diffuses into the depletion region and then drifts to the other side... again there's no charge separation until the carrier makes it to the depletion region. It's just like the photodiode diffusion.

So here's a related question. Is there a device dependent cut-off freqeuncy to diode shot noise. That is, no frequencies higher than one over the transit time of the device. (My simple model would predict something like that.. but I could be totally wrong.)

George H.