phemt diode

May 11, 2021 Last reply: 5 years ago 10 Replies

The late lamanted ATF-50189 phemt made an amazing diode. 1 amp and about 3 pF.



I tried a MiniCircuits SAV-541 as a diode.



formatting link
Cd is only about half a pF and Vf about 0.2. What's weird is that it's very noisy in the negative-drain "forward" conduction direction. It's the same at 20 MHz scope bandwidth, so it's not some GHz oscillation, and it's a 1x probe anyhow.



Gotta look at this some more. I may have a use for a very fast, low capacitance clamp.



There's another connection, drain to gate, but capacitance would be higher. Cds and Cgs would be in parallel.



Phil's SAV model looks pretty good. It predicts the reverse drain conduction spot on.


Typically, compound semiconductor devices that show a real lot of 1/f noise are gradually disassembling themselves. LEDs and diode lasers do the same thing--you can use anomalous 1/f noise as an indication of early failure.

"Blind luck and bloody ignorance", as me old dad used to say.

I'll take it. ;)

Cheers

Phil Hobbs

It seems that there is oscillation beyond the range of my 500 MHz scope. At -0.2 volts on the gate and drain!

I added a resistor in the gate, to ground, to see if the diode effect involves gate current. It doesn't seem to, at least until I hit the 12 volt drain breakdown.

Beats working.

There was a circuit of a jfet-based oscillator dc/dc converter that ran down to 80 mV DC input, from a thermocouple maybe. I wonder if a phemt could do even better.

Not your upmarket 20 MHz LPF, then. ;)

Okay, cool.

That was the "Joule thief" thing, right?

Cheers

Phil Hobbs

That was Jan Pantelje lighting a LED from a candle flame!

piglet

His schematics, being unreadable, make good kindling.

I draw on d-size velum and never throw them away.

[...]

You can make a detector with a small loop of wire and a fast schottky or germanium diode. Just wave it near the circuit to spot any unusual radiation.

There are plenty of radiation detectors on Amazon for finding bugs. I have no idea if they are any good.

Linear Technology used to make logarithmic detectors with very wide bandwidth and good sensitivity.

A SDRPlay RSP1A covers 1 KHz to 2 Ghz with 14 bit resolution and sensitivity in microvolts. It will find the weakest of oscillations.

Your spectrum analyzer should be good for detecting oscillations.

You should be able to pick up radiation from your celltower. This may confuse the readings from a parasitic oscillation.

What happened to your SD-32?

Right, that should suggest what's going on.

Still around. I have the 3 GHz sampling probe too, SD14 I think, that's a great e-field probe. But the sampling scopes need an external trigger.

We have a 7 GHz LeCroy scope with fet probes too.

You can at least see hash if you use an arbitrary trigger frequency.

If you use a DC block, you can probably put those on your SA safely. I commonly use P6201 and P6249 FET probes on mine. (I have the matching TekProbe power supplies.)

It's super easy to blow up the first mixer if you aren't careful about the DC!

Cheers

Phil Hobbs

I think they're great! Worthy of inclusion with the other great Dutch Masters.

He sent me an original, complete with coffee and cigarette stains. One day I'll have to get around to framing it.

Let's start up a crowdfund to buy him some grid paper and roller-ball pens.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required