QM and diodes (2023 Update)

Semiconductor detectors were discovered at almost exactly the same time Planck formulated the first quantum hypothesis. Bose patented the galena/cat whisker detector in 1901, and Planck published his seminal paper in 1900.

Cheers

Phil Hobbs

I think building a practicable PN-type diode without knowledge of stuff like direct band gap and indirect bandgap semiconductors (or what a "band gap" is in the first place...) would've been difficult, how would you know which semiconductor and dopant materials and how much of the latter to use in the first place?

There are a lot of semiconductors and dopants to pick from, sounds like you could fly blind around there a long time without a map.

... Would engineers not

Discovery confers knowledge. Science means knowledge and understanding (and that understanding is how we can make predictions). So, three scientists that discovered the transistor, are represented in our knowledge mainly by the one that wrote the book on it... with all that theory. Understanding is always theory, and as such can reach farther than our past observations. Theory gives us a wider scope than tinkering ever can.

I had lunch with Walter Brattain! He and his pals were trying to make a jfet and discovered the bipolar transistor mostly by accident, by tinkering.

We just had a whiteboard session on PIN diodes, specifically how hard I have to bias them on and off to steer some fast pulses. The conclusion is that we don't understand them very well so should overkill.

My board only has -5 volts available and I'd feel a lot better back-biasing them more. So C will add a wire in our board-board ribbon cable to get me some of her -15.

If it's just a short pulse you need, can you do the capacitor-discharge-ignition trick with a +15 supply? Sometimes the add-a-wire solution is opening up a ratsnest, and 'her -15' seems to indicate some responsibility encroachment. Like, is RF exposure going to visit someone else's power?

I could add a charge pump or some sort of inverter to my little board, but that might be a jitter hazard. C already has some -15 on her board, and I only need a couple mA, and neither board layout is final, so I'll steal a bit of hers and filter it hard just off the connector.

One issue is how long the PIN takes to forward conduct, which I don't know. So a lot of reverse bias is prudent. That reduces the off capacitance a bit more too.

If you can deplete the full thickness of the die, and switch the bias nice and fast, that ought to help a lot.

Cheers

Phil Hobbs

The pin diode bias will be switched infrequently.

Here's the idea: we have a fast amplifier and we want a very clean gain path. But occasionally we want to inject a test pulse, and sometimes may want to kill some gain.

The beauty of the pin diodes is that they basically disappear from the signal path in normal operation. As in 0.14 pF.

Yeah, they're super useful. As a former RF guy, I knew about that nice behaviour, though I've never actually used one in a design AFAICR.

Cheers

Phil Hobbs