Transistor backwards... But it works ?_?

It is if they're inserted backwards.

That's the way you can usually tell C from E with a handheld multimeter (on the diode range). If it has an HFE function, that's a confirmation (but not so convenient to use on SMT transistors).

I knew there was a good reason not to try it.

If it has a diode range.

I was spoiled, I had access to curve tracers before meters with diode ranges became common.

About 50 years ago, I made a telephone by using a scraped OC71 as photodetector, and an incadescent bulb with DC feed and modulation added in series to the feed. As antennas served two automobile headlight mirrors.

I did about the same for a middle-school science project. The transmitter was a radio (Hallicrafters S38) speaker output in series with a flashlight battery, driving a flashlight bulb. The receiver was a selenium solar cell and a transistor amplifier. It really didn't sound bad, which is impressive for thermally modulating a filament.

An over-driven emitter follower, namely driving the base above the collector voltage, can be made to saturate to exactly zero volts Vec. I built a mess of 16-bit DACs from discrete transistors and a wirewound resistor ladder network once, using that trick.

Zenering the b-e junction gradually destroys beta. I don't think reverse operation is a problem otherwise.

I kind of thought that destroyed the beta in the device when using it in forward mode afterwards? This was something that took place over time of use in the reverse mode.

I guess it wouldn't matter if you never intend to extract the unit for another job.

I remember a circuit years ago that had a R on the emitter side and if the R got opened and the circuit had been operating for while in reverse state, the transistor had to be replaced because the beta was no longer where it should of been.

Jamie...

"Exactly" ?:-)

Yes.

Really small feature-size PECL carefully limits the reverse Vbe. ...Jim Thompson

Thermal forcing gives you a one-pole rolloff, so it isn't too hard to equalize it over a moderate frequency band.

At high enough frequency, thermal diffusion starts to dominate, and then things get harder exponentially.

Cheers

Phil Hobbs

I made a 4 GHz signal generator out of a 'RL Drake' voltage tuned 70 MHz downconverter, and added a video input to modulate it. It was handy to test & repair old C band equipment. :)

That circuit is working with an IC/IB ratio of ~ 0.5, about a factor of 20x less than usual, it only needs the transistor for inversion apparently.

That explains how this is somehow a low voltage reference:

Please view in a fixed-width font such as Courier.

. . . . . Isrc . | . | . | . .------+ . | | . | --- . | \ / . | --- . | | . | +----> Vref . | | . | |/ . '----| . |>

. | . | . ---- . . . . . .

inexplicable.

Shit... I guess somebody's already done it then. Possibly twice.

And the _value_ is ?:-) ...Jim Thompson

If you use a BJT upside down, it turns into a mangier version with rather low gain but an exceptionally low saturation voltage.

We used this stunt long ago in aerospace to minimize switching saturation offsets.

Cost effective MOSFETS eventually eliminated the need.

Except at low currents, where you die from capacitance with MOSFETs. RF transistors run upside down generally have lower output capacitance than when run right side up, and f_T is often a bit higher too.

Cheers

Phil Hobbs

So you're saying the heavier doped (emitter) will result in a lower saturation state when used as the high side element in a common low type circuit?

I'll have to put something on a test board and generate a bode chart on this. Something tells me the beta is lower because the emitter forward bias (lower forward). This means the base bias would actually be partially flowing through the emitter when saturation starts to drop below the diode drop? This would actually cause an opposite effect in Hfe, like seen in most cases where it goes up with higher currents.

That will give me something to do later unless we are talking about something else totally different here? I've never intentionally used a transistor that way, but have replaced a few that ended up working that wall in a circuit in a failed state.

Jamie

A Bode plot is gain versus frequency, usually on logarithmic axes. I think you meant a current transfer ratio plot.