NPN in unusual quadrant

Nothing until you exceed the emitter base reverse breakdown (or collector base breakdown) voltage,

Bitter-base is usually about 5V

Behaves kind of like a zener diode and damages the transistor.

The collector becomes a better OFF switch; the collector breakdown Vceo applies when the base is open, but higher Vcbo applies when base is held negative.

But what happens?

? ? ?

RL

If the collector is already positive and pullup resistor is not extremely high then I think not much detectable will happen even once the Veb junction breaks down. But if the collector is open circuit or very lightly loaded then the collector should show that photo-electric effect "the pease conundrum" we discussed back in 2015. Whereby the E-B avalanching junction emits light that the C-B junction does a photo-diode act on?

piglet

That was a stated condition

"Detectable" can include amps or picoamps of currents.

Curious that nobody seems to address this case.

How could the information be used? It seems to have not been needed in the past, yes?

I'd expect that understanding transistors is a generally good thing to do. Looking at schematics on the web, it seems like few people do.

If it behaves as I suspect it may, I might have a use for it.

Back in the 60's/70's there were some germanium transistors marketed for avalanche pulse generator work. May have been Mullard / Philips, but don't remember any circuit details.

Years ago, found myself fixing more than a few high power audio amps and built a simple breakdown tester with a variac, stepdown transformer, limiting resistor and full wave rectifer. Got some interesting effects at breakdown, very high frequency oscillation on the scope. 2n3055 class power devices, normally quite low frequency use. Interesting effect, perhaps similar to other weird stuff of the time, like impatt diodes...

Chris

Early Tek and HP sampling scopes used an avalanche transistor to drive the sampler. I think maybe Lumatron did that first.

Transistors do all sorts of fun stuff that's not on the label.

Zetex makes SOT23 silicon transistors that are designed to avalanche. As in <1 ns, 300 volts, 60 amps.

Some other tricks:

unclamped inductor flyback avalanche

zero volt saturation

step-recovery effects

stored charge effects

inverse beta

temp-compensated b-e zenering

They can be used as current-controlled resistors in reverse-saturation mode, where the collector-emitter resistance becomes generally proportional to the base current.

You missed noise generation from eb breakdown. John

Ohmic region? Cute.

Right. I was thinking that my original quadrant comment could suggest a mode where the b-e zener makes noise and the c-b junction amplifies it. That depends on where the zener-inspired carriers go.

Somebody should try that. I can't do that this week or so.

Oh, that's obvious; breakdown of B-E is mainly dominated by carriers from the emitter, because there are so many of them (base is lightly doped). Those carriers are repelled by the collector (+ bias on collector, and emitter of NPN is shooting holes).

Main effect I'd expect is aging of the surface, leading to base leakage if you ever revert the bias to normal transresistance operation.

but would the E-B reverse breakdown produce photons?

It always seems to. The packaging determines whether you can see them. There was a thread where the issue came up about a decade ago. mainly about whether zener diodes could oscillate - theye don't seem to - but avalanche diodes - anything that breaks down at much over 5V - can look as if that's what's going on.

Yes. Cut the top off a TO-18 or TO-5 metal can transistor and see for yourself!

Back in fall 2015 we discussed this and I posted some photos. I cannot recall the current but I think was in the region of 10-20mA the light was clearly visible with ceiling lights off. To the eye it looked silvery-gray but the phone camera showed orange, maybe because there was a IR component to it?

Robert Pease described the effect after seeing little speckles of light as op-amps were offset trimmed by the zener-zap technique (AFAIK an array of sacrificial junctions selectively zapped, a kind of anti-fuse, the zapped structure becoming a short).

piglet

So, to expect this in future, should we assume the B-E breakdown creates light, and the collector current is a photodiode detecting that light? That does explain the effect, and predicts emitter current proportional to collector current.