why can't we make our own transistors?

[pinkjimiphoton]

couldn't you just connect to diodes together and use that as a transistor?

if it would work, i wonder what it would sound like?

[merlinb]

couldn't you just connect to diodes together and use that as a transistor?

No, because the two diodes would not be sharing a single lump of silicon (the base region).

But it is possible:
http://www.youtube.com/watch?v=vmotkjMSKnI

[rousejeremy]

ultra mojo. these make NKT275 look like a 2N3904

[fuzzo]

actually you can make your own semi-conductors at home   

http://www.youtube.com/watch?v=PdcKwOo7dmM

[R.G.]

Yep, it's the injection of charge into the shared base region that "poisons" the ability of the reverse-biased collector-base to hold off current flow that makes transistor action. Gotta be a shared base.

But yes, one can make transistors. FETs are actually easier than bipolars.

But the original bipolars were made with a chunk of purified germanium crystal which was carefully im-purified with N-type  impurities. Then, a chunk of metallic indium was put on either side of the base  ... yep, that's where the term came from... slab of germanium and baked. The heat caused the indium to diffuse into the germanium base, causing one region that ... emitted ... charge carriers and one that ... collected ... them.

Germanium was mankind's "learner kit" for making semiconductors because the conditions for diffusing it into semiconducting junctions is less demanding that for it's low-leakage cousing, silicon. Germanium slabs were sandwiched with indium and baked on trays in atmosphere-controlled ovens. The resulting transistors had lead wires soldered to them and were tested and sorted into grades.

It's more demanding than many garage-shop operations, but not terribly. Of course, the early transistors were not terribly good, either.

[armdnrdy]

There! You now have it Jimi....now get to work and make us all a batch of J201s.

After you "master" that....I'll take a few hundred PHOTON1024's and JP3005's   

[Jdansti]

You need some dope, dude...  

I can get you a hit of high grade gallium...   

[pinkjimiphoton]

*I* need DOPE??!?!?!?!?!?!?!?!?!?


lololololololol.....

[armdnrdy]

"If you don't finish yer stompbox, you can't have any dope!.
How can you have any dope if you don't finish your stompbox?"

[pinkjimiphoton]

YOU!! YES YOU!! WITH THE SOLDERING IRON !!! STAND STILL, LADDIE!!!!





all in all you were all just parts in my board???   

[petey twofinger]

roger , caroline is on the phone , congratulations , you have just discovered the secret vero , please send your germanium to old pink care of the funny farm at .....

[Thecomedian]

Yep, it's the injection of charge into the shared base region that "poisons" the ability of the reverse-biased collector-base to hold off current flow that makes transistor action. Gotta be a shared base.

But yes, one can make transistors. FETs are actually easier than bipolars.

But the original bipolars were made with a chunk of purified germanium crystal which was carefully im-purified with N-type  impurities. Then, a chunk of metallic indium was put on either side of the base  ... yep, that's where the term came from... slab of germanium and baked. The heat caused the indium to diffuse into the germanium base, causing one region that ... emitted ... charge carriers and one that ... collected ... them.

Germanium was mankind's "learner kit" for making semiconductors because the conditions for diffusing it into semiconducting junctions is less demanding that for it's low-leakage cousing, silicon. Germanium slabs were sandwiched with indium and baked on trays in atmosphere-controlled ovens. The resulting transistors had lead wires soldered to them and were tested and sorted into grades.

It's more demanding than many garage-shop operations, but not terribly. Of course, the early transistors were not terribly good, either.

different types of doping agents for the regions also affected it's behavior to some degree, which is why I think there may be subtle difference in "sound" from early germ transistors to ones made either today or in the 80-90's.

Germanium switches slower, silicon is faster. lately they're finding more use for GeSi composite ICs to take advantage of the properties of both materials and do stuff neither could do alone. Kind of like how Tubes are having a resurgence in use in specialized functions, as very very small objects.

They still use tubes at least somewhere and wire wrap in space planes afaik.


As far as chaining diodes into a transistor, I tried that in LTspice and it wouldnt do anything. It has to be physically connected sandwich layers and not by wire.

[wavley]

Yep, it's the injection of charge into the shared base region that "poisons" the ability of the reverse-biased collector-base to hold off current flow that makes transistor action. Gotta be a shared base.

But yes, one can make transistors. FETs are actually easier than bipolars.

But the original bipolars were made with a chunk of purified germanium crystal which was carefully im-purified with N-type  impurities. Then, a chunk of metallic indium was put on either side of the base  ... yep, that's where the term came from... slab of germanium and baked. The heat caused the indium to diffuse into the germanium base, causing one region that ... emitted ... charge carriers and one that ... collected ... them.

Germanium was mankind's "learner kit" for making semiconductors because the conditions for diffusing it into semiconducting junctions is less demanding that for it's low-leakage cousing, silicon. Germanium slabs were sandwiched with indium and baked on trays in atmosphere-controlled ovens. The resulting transistors had lead wires soldered to them and were tested and sorted into grades.

It's more demanding than many garage-shop operations, but not terribly. Of course, the early transistors were not terribly good, either.

Of course indium is REALLY expensive these days, we use it at my work in cryogenic dewars for it's excellent thermal transfer capabilities, basically as a really expensive gasket that we have to scrape off carefully and reuse.

Germanium switches slower, silicon is faster. lately they're finding more use for GeSi composite ICs to take advantage of the properties of both materials and do stuff neither could do alone. Kind of like how Tubes are having a resurgence in use in specialized functions, as very very small objects.

We have been researching the use of SiGe devices for use in the low GHz with mixed, but promising, results.  Not as nice as discrete InP HEMTs, but better than MMIC at these frequencies.

[J0K3RX]

Really? 

[tca]

You could try and build a zinc negative resistor: http://www.sparkbangbuzz.com/els/zincosc-el.htm And make an amplifier out of that!

Cheers.

[brett]

Hi
just thinking that the dimensions of the required layers are about the thickness of paint. Put some micro-fine Ge in an aerosol can with a volatile carrier like acetone or ethanol and ....spay a JFET?
Aluminium is a bit cheaper and easier than Indium as a doping agent. Alfoil is very thin (100 atoms!) and small areas (1mm^2) contain very few atoms. Perfect for doping a gram or two of Ge.
Of course, the Ge will be also be + or - doped by common contamination. With luck, the doping would vary spatially, producing a variety of devices. And the transistors might not necessarily require multiple layers. They might exist 'across' a sheet of sprayed Ge.
just 2c worth...
cheers