Negative resistance with BJTs

dakupoto wrote

Negative resistance can be done with only one diode, it is called a tunnel-diode or Esaki_diode:

Just bias the diode at the right current where I versus V is negative

You can do it with one or more tunnel diodes, but not with ordinary ones because they have no gain.

You can also do it with a single transistor at AC.

Cheers

Phil Hobbs

Maybe if you sawed them open at exactly the right point & fused the silicon dies together... otherwise they're just 2 diodes, with the middle p&n layers not in contact with each other.

NT

Some diodes have a negative resistance region, but it's obscure. You can tease a zener diode to oscillate badly. RF diodes, gunn and such, oscillate. I think an avalanche photodiode will oscillate in the right circuit.

An LED and a photodiode in series? In other words, an optocoupler with CTR > 1.

As a kid, I tried using two diodes as a transistor. Apparently lots of people did that.

I think I've still got a commercial one somewhere, left over from my research in the 1960s. With the right bias they would oscillate at very high frequencies (for those days!)...

Mike.

Gunn is not even a diode, or -- if you prefer -- a junctionless one. :-] Amazing construction, BTW.

When you grow up, you start using transistors as diodes...

Best regards, Piotr

Out of curiosity: do what?

Best regards, Piotr

Make a negative resistor. Start with an emitter follower and put a small capacitance on the emitter. The input resistance goes negative. That's why you have to use beads, anti-snivet resistors, and so on.

Cheers

Phil Hobbs

You used to be able to buy four-layer diodes (also sometimes called Schockley diodes - not to be confused with Schottky!) in a range of voltages. Became obsolete mid 1970s. I had to service gear that used them.

If you accept fudging the rectifier part of the name of an SCR into what you'd call a diode then that is a four-layer diode. Leave the gate terminal open and at a high enough voltage it will exhibit negative resistance !!

piglet

You mean the DIAC? They are not obsolete, just niche. Still used in TRIAC dimmers and self-oscillating half-bridge converters for halogen lamps, etc. The immortal DB3, for instance.

Best regards, Piotr

Two discretes, no; there have to be carriers passing through one layer into the next without recombining, and a metal wire always recombines 'em. The two-transistor connection is just an SCR, and the negative resistance comes from its breakdown voltage. The commercial diode device that contains that circuit, a diac, is commonly used for SCR triggers, though I recall it also being used for fox& hound type wire tracing.

Yeah, you get a PUT or SCR depending on which node you nominate as gate.

sure, but you'd need to decapsulate them grind the dies and somehow fuse the silicon together just right...

The first transistors were made by adding a cat's whisker base connection to a diode.

Cheers

Phil Hobbs

Snivet?

Funny.

Some NPN transistors can be used as "reference zeners" to make 6.2 volts at a very low TC. That's the zenering b-e junction in series with the forward-conducting b-c.

Why? An electron enters the wire and another comes out the other end.

The Fox and hound I looked into just had the antenna going into an LM386.

The "base" got its name because it was the semiconductor slab. The emitter and collector were the point contacts.

When I was a kid, I got a tour of Bell Labs and had lunch with Walter Brattain. Nice old guy.

Hah, Phil's showing his age\\\experience? :-)

"Snivets" was an oscillation that sometimes affected vacuum tube television sets, where the horizontal sweep driver had UHF oscillations for part of its operating cycle, apparently due to the cavity formed between screen and plate, and the exact electron beam density and electrode voltages applied.

(I know this from books, certainly not from seeing the phenomenon in person. I don't know that it was something that happened randomly in any set; I suspect it was more of an early problem, and/or a Muntzing problem. Back in the era when, y'know, Muntz was a real brand and everything. :-) )

Tetrodes are well known for having a negative resistance region on the plate curve (due to secondary emission diverting plate current to the screen), and beam tetrodes less so but still to some extent in some parts of their operating range. But I don't know that it's a fast enough phenomenon to cause UHF oscillation. The usual description implies it's a problem with the internal design of the tube itself, which doesn't suggest it would be a universal problem, only an early one.

Tim

One of the uses for the term was the microwave emissions emitted by tetrode s, pentodes and beam power tubes when used as switchers, such as a horizont al output (or LOPT) in TV sets. It was pretty much eliminated by putting a positive voltage on the suppressor grid, which of course would not work wit h tetrodes.

It only became a problem with the introduction of the UHF TV band. The sniv ets were not detected by common test equipment but interfered with the UHF band.

They used to cause mostly vertically oriented black oval "holes" in the pic ture, almost always to the right. They were modulated by line frequency whi ch indicates that the filament voltage was involved because everything else was filtered.

As far as I know, this phenomenon did not occur with semiconductor devices. But then...we have this.