Electrical Engineering

i am liking the cost and behaviour of my current processing chain so i'm going to do the 2nd order channel-select filter and an envelope detector and then finally the comparator

The circuit is gone now, but I had ~3uA from the photodiode, producing 15mA through the NJFET and PNP pair

but all the current was going down into the NJFET and my PNP load resistor never got any significant current, so it sat at 3.3V or whatever

Asmyldof

@KyranF That is expected behaviour, depending on the size of that resistor

pull out of the base of a PNP and it will conduct, if you made that resistor large, it will hardly create any Collector current

KyranF

yeah

Asmyldof

And thus go to high

I would have expected Rphoto to be 100's of k's at most in that one and the collector resistor in the 100 ~ 1000 ohm range depending on the PNP

Anyway

KyranF

but how low can one go? I tried 1000 ohms on the collector, plus a few others

Asmyldof

If you want cuter bias stability the upcoming picture may inspire

Might have to get not a 2N3906

models at gains of 300~700

KyranF

10:31 PM

i'm liking the 2n3906

way better than the 2n4401/3 models i have in my office for breadboarding

Asmyldof

better is a matter of what you need

There's a reason all those others are still made

KyranF

well yeah in this case, i need the 3906 or similar

of course, i'd not use the 3906 for high power or "Switch" style things but for low current small signals it's great.

Asmyldof

@KyranF unless it isn't

In the case of option 1, the high gain of the 3906 is exactly what kills power efficiency

I also did the following:.......

BTW, these designs in no way really need a JFET, but I found it fun to keep it in

With the low photo currents the mirror should have a collector voltage no less than 1V under 5V and only mildly variable, which gives you higher bias stability

I really don't think the R and C extra are a good idea, but it's getting late again

So I thought I'd just upload it and let someone else choose

If with some coupling setup there you can reject static currents, you only amplify and transimpede dynamic current signals

The Photon

Why is it that whenever I throw some old junk away I end up needing it 3 months later?

Asmyldof

but, again, I'm not very happy with it as drawn

Law of conservation of lack of rack space

It's a self-oscillating law

One of the few

The Photon

10:50 PM

Product A:

KyranF

@Asmyldof what's the zener looking thing near the photodiode, connected to the NPN base?

Asmyldof

@KyranF a zener

The Photon

Product B:

Suspiciously similar?

KyranF

@Asmyldof 6.8V Zener?

Asmyldof

@ThePhoton artistic coincidence

@KyranF 's what it says

KyranF

10:51 PM

just checkin!

Asmyldof

Maybe...

This is what I mean:

Zener is just there to make sure the current mirror stays somewhere near its mirror domain

Direct base connection would pull the right collector so low it would wack the whole mirror out of balance

In the "update" it's there to limit the DC peak voltage on the left of the capacitor

Actually, for a .... sort of right setpoint, you'd want:

Though a diode would do

most likely

maybe

Brain stopped

KyranF

11:10 PM

haha

so a current mirror was one way i thought to adventure upon for this challenge

it seems like a good way to avoid loading down the photodiode

Asmyldof

And had you also thought of rejection of DC current before initial TIA action, I would have slept a lot more this week

:-P

KyranF

11:30 PM

DC current can be cancelled by using an op-amp and trimmer pot going into the non-inverting input of the TIA amp and you can just servo out the DC error, but thats really only for the dark current

by AC coupling the transistor gain stages it would indeed cut out the DC current in general

Asmyldof

@KyranF Good luck when you work in semi-shaded sunny rooms

I made a guesstimation

For what might be okay

KyranF

yeah.. i mean it could saturate my photodiode but honestly, it's hard enough to get signal out of this thing let alone the sun adding to it

the sun would just increase the DC noise/background level, but all my signal chains are AC coupled to hopefully avoid that

Asmyldof

Hmz

The sooner you get the current out, before TIA-ing, the less chance you force/shift operation points out of a nice domain

KyranF