Avalanche transistors

Good to know. The old-style double-diffused 2N2369s in the metal cans have all gone away, AFAICT.

Cheers

Phil Hobbs

The Zetex things avalanche beautifully. Their breakdown voltages are very predictable (up to 300 volts), they trigger nicely, and when they avalanche they short out like an SCR, pretty much down to zero volts. A SOT23 can output 5 kilowatt pulses. But they're slow... I recall getting turnon edges above a nanosecond.

We get fast, tens-of-volts/100 ps edges from step-recovery diodes.

The fastest all-electrical edges come from shock lines, nonlinear transmission lines. PSPL sells the only available parts, as far as I can determine.

John

Sure, the Zetex cannot rival that. But: They have datasheet values for lifetime. Do you know how long an avalanching BFG541 will last until it goes phut?

Probably depends on the area under the curve to a degree. Do some accelerated life tests ?.

In the old days, they made ge transistors just for that job and tunnel diodes were quite interesting as well, though don't know if anyone makes them anymore...

Regards,

Chris

How well did the old germaniums do at that? Ive got plenty of PNPs, but they're so leaky, how would you even tell if they've avalanched? ;-) You'd have to switch the capacitor into it, there's too much leakage for RC charging like you'd use a 2N2369.

(Fact: the old TO-5's were so leaky and so bad thermally that a transistor, base open, in series with 1kohm and supplied by 15V, will show hysteretic behavior like an SCR. As it heats up, it holds itself right at the maximum power point (half +V) from leakage. Disconnect it for a second and it cools off; forward bias it slightly and it turns on.)

Tim

That's the best biasing scheme ever: no resistors at all.

Sure Joerg's drooling at that :-)

In some of the old germanium transistor circuits, the base bias resistor network *reduced* the collector current.

I once, when I was a kid, mounted a black Ge transistor in a flashlight reflector and unintentionally biased it just at the edge of themal runaway. It would detect my hand, from thermal IR, from a few feet away.

John

On a sunny day (Sat, 24 Oct 2009 01:21:53 -0700) it happened John Larkin wrote in :

Or, maybe it was an old _glass_ black painted transistor, like OC71 / OC13 etc.. and the paint was damaged, so it would detect *light* variations caused by your hand.

I used the OC71 in an audio preamp, until one day I noticed an increasing and decreasing noise, very irregular, found out it changed every time a cloud came in front of the sun.

Just a little damage of that black paint.

Silicon transistors do that too. BV_CEO (base open) is about where the base current changes sign, and is always lower than BV_CES (BE shorted). Nowadays high BV_CEO values are available inexpensively, so there's usually no need to work in that range unless you're building avalanche circuits. (Coincidentally on-topic.) ;)

Cheers

Phil Hobbs

etc..

your hand.

decreasing noise,

sun.

At one point, lots of phototransistors were just normal BJTs mounted in transparent packages. I think they're worse now.

Cheers

Phil Hobbs

hand.

No, it was a metal can, 2N107 I think. It wasn't photosensitive.

The early GE silicon transistors were packaged in a brownish epoxy and were photosensitive.

John

etc..

your hand.

Yup, sort of concave on the top where they poured the tan epoxy into a mold.

They were quite good transistors. I made some opamps, with a pair of them shrink-tubed together and got below 1 uV per degree C with a bit of current trimming.

GE made cool parts, like their symmetric zener diodes, in that package.

John

etc..

your hand.

Gawd, I can barely remember those- brown 'D' on the top with a black disk at the base, wasn't it?

Best regards, Spehro Pefhany

Some Si transistors had weird effects as well. Wasn't there some HP gear where they had screwed up a reverse Vbe situation and the common "fix" was to extinguish a Marlboro butt on top of the transistor once in a while?

Wonder what they do now that smoking is banned from all labs.

Initially (in the '70's or so) they were ordinary transistors like you mentioned, mounted in the standard small metal can, but with the top cut off (actually formed that way - not really cut). Then in the '80s or so, Fairchild designed a transistor optimized as a photo detector. The base took almost all of the die area, and the emitter was a dinky square area off in one corner (just large enough for bonding). They were sold in epoxy packages that had an integral clear top; one version was flat-top and the other was a lens: FPT100. About TWO orders of magnitude more sensitive. The market must not have been all that great. I think i have a few left.

I never used them, but remember seeing them listed in Mullard or Ferranti short form catalogs in the early 70's. Can't remember circuits, but would guess the base was grounded or even reverse biased.

We used to select si power transistors for breakdown for audio work by running a half wave rectified voltage through a limiting resistor across collector emitter, with base shorted to emitter. It's quite a crude test, but curve tracers were expensive at the time. Same thing works for si diodes. You can see loads of rf as the device breaks down.

The ge transistors were not that bad, especially types like the 2n305 series, with leakage currents in the low tens of microamps, if that. You need to short base to emitter to measure leakage current accurately though, due to the leakage between collector / base biasing the device on otherwise. Once you get a biasing network round the device, the leakage is swamped for all practical purposes...

Regards,

Chris

No, I don't. It's not specified for avalanche operation. It's just a curiosity that may be useful.

Anyway, I also own a PSPL 4015. :-)

Jeroen Belleman

etc..

your hand.

decreasing noise,

sun.

Philips cottoned onto this 'problem' with the OC71 and fixed it by flooding the OC71 (and other transistors using the same encapsulation) with opaque silicon grease and repackaging the old OC71, without the black paint, as the OCP71 photo transistor at a premium price!

On a sunny day (Mon, 26 Oct 2009 14:54:59 -0000) it happened Terry Casey wrote in :

etc..

your hand.

decreasing noise,

sun.

Jup, I scrapped paint of the OC13 so I could use it as photo transistor :-)