Why predominantly PNP Germanium transistors?

[Radio Wrangler]

Ah, easy!

Cast your mind back to those little plastic hand-held games with an array of sliding tiles with either letters on them or bits of a picture. One tile space was empty. A sort of 2-dimensional predecessor to the Rubik's cube.

Your mission was to slide tiles around to put them in the right places to spell a word, or get the picture right.

You can't move the space, because the space doesn't exist as a tile, but you can move the tiles (which do exist) around it to open up the gap somewhere else. Several tile movements are needed to shift the space to an adjacent position.

Therefore the mobility of the space (hole) can only be a fraction of the mobility of the tiles (electrons).

David

[John10b]

Ta David, I am really trying hard to understand it, give me time for it to sink in
John

[ajgriff]

To my mind is it's just as easy to move a space as it is to move a tile. It is after all a one to one relationship. Having said that I've never found the whole 'hole' analogy very satisfactory. I think, probably wrongly, that it could be argued that there are equal numbers of electrons and holes to begin with anyway. I'll join John and scratch my head.

Alan

[TonyDuell]

I once tied a physics teacher up in knots with the following question :

Consider a slab of a conductive material, with a current passing through it (say in the 'Y' direction). Now apply a magnetic field at right angles (say in the 'Z') direction. It will deflect the charge carriers in the material, so the in the 'X' direction you should get a voltage set up between the 2 faces of the material.

You do. It's called the Hall Effect, and it's how those magnetic field sensors work.

Now if you think some more it turns out the the direction the charge carriers are deflected does not depend on their sign. So the polarity of the Hall voltage depends on the sign of the charge carriers, n-type and p-type semiconductors should give opposite polarity hall voltages.

Again they do. This is an observed effect, you can distinguish n-type from p-type materials this way.

But if the 'hole current' is really an electron current flowing the other way, why does a p-type semiconductor appear to have positive charge carriers. Why do 'holes' (a lack of an electron) appear to have positive mass?

As I said, I got a teacher very confused with that. FWIW I believe there is no explanation of it in classical physics.

[ajgriff]

Originally Posted by TonyDuell

... I believe there is no explanation for it in classical physics.

Agreed, it requires a quantum leap and I'm certainly not going there.

Alan

[Radio Wrangler]

An awful lot of telly CRTs were made which used the fact that ion beams were deflected by a magnet in the opposite sense to what the electron beam was, in order to prevent ions burning the screen....

I've never had to get to the bottom of the hall effect. The bit that intrigues me is that hall sensors seem to be made of semiconductor material. A simple electron current deflection model would seem to have to work just as well in a sheet of resisrive material... something interesting is going on.

David

[AC/HL]

Given that PNP was easier to fabricate in the Germanium days, and the early applications were fairly basic and tended to be analogous to the existing valve designs, there was little incentive to push it. The earliest NPNs I can recall were switching transistors, OC139 etc, in early logic circuitry. Complementary circuits changed all that, or was it the other way round? Changing to Silicon was the subsequent advance.

[Radio Wrangler]

The predominance of PNP in the early days of transistors threw a bit of a googly at people used to valves, not only did they get hit with big reductions in voltage, they got lower impedance circuits and on top of it all, everything flipped to positive earth.

I can remember as a kid looking at my jars full of laboriously reclaimed and tested resistors, only to think that they all were either too low or too high in value for transistor circuits.

David

[kalee20]

Originally Posted by Radio Wrangler

I've never had to get to the bottom of the hall effect. The bit that intrigues me is that hall sensors seem to be made of semiconductor material. A simple electron current deflection model would seem to have to work just as well in a sheet of resisrive material...

Ignore the resistiveness, that's just a red herring.

When you pass current through a slab of metal, electrons move through it (slowly).

Moving electrons in a magnetic field get deflected sideways, that's how magnetic deflection in a CRT works.

So the electrons get more abundant on one side of the slab of metal, and you get a small pd across the slab. (Note: we haven't needed to use any resistance effect).

Now swap the metal for an identically-dimensioned slab of semiconductor. Now we have all electrons locked in place, except for the very few contributed by the dopant (perhaps one atom of dopant per 10,000 atoms of material).

With 10,000 times fewer electrons available to pass our current than in a metal, they have to move 10,000 times faster.

Electrons moving 10,000 times faster experience 10,000 times more sideways deflection. Hence, 10,000 times more pd is produced.

So the output is much more useable.

But... I'm with everyone who says that P-type materials producing opposite-polarity Hall voltage, is just mind-boggling! I also got pulled up short by the concept in A level physics. It doesn't make sense. One day, perhaps, I'll understand it...

[G6Tanuki]

I've always pondered this, particularly in the context of low/medium-power Germanium output-stages.

AC127/AC128, AC187/188 as used in portable radios, and the later AD161/162 which were ubiquitous in 1960s/1970s domestic 'hifi' and car radio/cassette-players.

The precise physics of Gernanium vs Silicon escapes me, but at a practical level the intrinsic band-gap voltage of Germanium at around 0.1v no doubt had some serious advantages against Silicon's 0.6v in low-power-supply-voltage battery-power applications.

The ability of a quad of Germanium NKT404 P-N-P power-transistors to turn hard-on and present minimal forward voltage-drop in 1960s power-inverters was really important in the days of Pye Cambridge/Vanguard mobile radios.

[Nuvistor]

A very comprehensive book from RCA, warning nearly 700 pages with lots of in depth info, too much for me but others may enjoy it.
https://www.americanradiohistory.com...1-RCA-1956.pdf