Re: About TIAs

I just drive it into the emitter of an NPN. That's my Tough TIA circuit.

Also a strong choice if you’ve got at least a few microamps of photocurrent. Biasing the CB stage with a quiet current source is a win for slightly faster things.

Then if your photons are valuable, you can get into closed-loop bootstrapped bootstraps and fancy stuff like that, and wind up with something like our QL01 (10M ohms, 7 sq mm, 1 MHz, shot noise limited above 25 nA in the full BW) or its just-introduced little brother the QL03 (1M ohms, 700 kHz, 150 sq mm, shot noise limited above 60 nA).

Cheers

Phil Hobbs

No photons here: the Tough TIA is for a capacitive tank level measurement system. The signal is a KHz-range sine wave, and we use a synchronous detector, and tank levels don't change very fast, so there's no big s/n problem. But the signal might be coming in over a couple hundred feet of cable, and might accidentally get connected to

John, ISTR we made a deal about not feeding the trolls?

Troll? I think he sparked an interesting discussion. The TIA amplifiers you typically see in application notes are very poor. It's good to show people here that there are ways to do much better.

Jeroen Belleman

His post was reasonable and on-topic, and he didn't insult anyone, and TIAs are interesting.

Transistor betas are all over the place, but alphas are very predictable.

An oldie but goodie from your humble servant: “photodiode front ends: the real story”

Phil Hobbs

Yes, that one made me take notice.

Jeroen Belleman

For my third edition, I tried reproducing it with modern op amps, but they all had way too much input capacitance. I had to use a different circuit.

Cheers

Phil Hobbs Who wishes he could still get LF357s

More to the point, Cursitor Doom *is* an anonymous troll, and any "agreement" with him exists only in his fertile imagination.

MMICs are fabulous, fast, low noise, stable, and cheap. But be aware that some, the newer self-biasing kind, get very weird at low frequencies, and the data sheets hide it.

Yeah, you wind up fighting the bias FB loop, which doesn’t end well.

Speaking of “fabulous, fast, low noise, stable, and cheap,” Simon and I built a < 60-ps TDR, honest 10%-90%, with nice clean edges and no artifacts to speak of.

Including a dual 16-bit DAC to set thresholds for the Rx andTx pulsers, a quad 8-bit to set the beginning and ending voltages of the ramp, an LPC845 MCU, board, SMA connectors, the fully stuffed boards are $16 in quantity

2. The same 24-inch RG188 patch cord also measures about 60 ps with an SD-24, so the TDR may be better than that.

Alpha waves emanating. ;)

Cheers

Phil Hobbs

On 12/27/2024 3:14 PM, Phil Hobbs wrote: <snip>

Ed