When did they stop making Ge Transistors?

[scott1568]

Just curious

[Gus]

I believe you can still have them made.  IIRC Someone posted the company name here or at ampage in the past.

[R.G.]

On a technical basis, their performance is poor.

Germanium leaks approximately 1000 times as much as silicon if all else is equal. Leakage is a Bad Thing if what you're trying to do is control current.

Silicon can tolerate much higher temperatures before dying than germanium.

Those two, all by themselves are enough.

You can still get germanium devices made. Some places specialize in it. But it's not cheap, because the volume is low, and all the technical sophistication that's been lavished on silicon for the last 50 years was not spent on making germanium better.

So - leaks worse, dies easier, and costs more. What would YOU do? 

[Philippe]

Germanium transistors are often overated due to their antiquity, availabilty (in certain cases) & past usage in certain fxs of the past. Proper selection of Hfe is more critical.

Case in point...when a virtuoso like EJ (who apparently has the audible hearing range of a dog) elects to run with silicon transistors in his FF, the hyperbole over germanium becomes more apparent.

EJ or no EJ, let your own ears be the judge.

[zombiwoof]

I think the answer to your question is that they are still being made, they just aren't in wide use in electronics these days except for by pedal freaks and such.  The technology is obsolete, but then so is tube technology, but don't tell that to guitar players and hifi nuts!

Al

[PRR]

> When did they stop making Ge Transistors?

Soon after better-performing Silicon parts got to tolerable price.

You can make "a transistor" with commercial Germanium and one re-refining. Silicon requires MUCH better refining to make anything useful.

A low-purity Ge transistor is pretty bad. However for battery uses it is SO much better than a tube that bad Ge parts were widely used.

The telephone company had millions of mechanical relays, needed more, but with less wear and maintenance. Computers used LOTS of tubes. Big heat, and 2,000 tubes at average 10,000-hour life means daily crashes. The advantages of a lower-power and potentially longer-life transistor were huge. But low-purity Ge parts weren't a ton better. Also there were instrumentation uses (notably oil-wells) where heat made leaky Ge useless.

So Bell Labs, Fairchild, IBM, Texas Instruments, and a lot of others worked on the purification problem. Better Ge devices were very good at room temperature. Along this path they also learned to purify Silicon to a useful degree, and make high-temperature transistors. These were produced in HUGE numbers for high-buck computer and instrumentation uses. The fallout from tight-spec production came on the market for less demanding applications. Once everybody was making Si parts, competition drove prices down.

The period 1962 to 1969 saw Ge go away and Si dominate. The last small Ge were cheap transistor radios and such. By 1972, IIRC, even these holes were mostly filled with Si parts. So obviously most small Ge production stopped in the 1960s.

There was a loooong tail of specialized Ge parts. Ge can be made for VERY high current, because it costs less and has lower voltage. Within the last decade huge Ge power transistors were still in production in Florida. Probably making spares for B-52 bomb-door motors and other legacy hardware. That company seems to have moved totally into opto-electronics and other more sophisticated uses of Ge compounds. I don't doubt that Ge foundries exist here and there around the world, not busy but not yet cut up for scrap. There seem to be ample supplies of "new" Ge parts, though whether freshly-made or found in warehouses is hard to know.

[brett]

Hi
RG is totally correct -

Silicon can tolerate much higher temperatures before dying than germanium.

However, many of our small signal circuits don't generate much heat, so the main issue isn't dying, but shifts in bias and increasing leakage with temperature.

The period 1962 to 1969 saw Ge go away and Si dominate. The last small Ge were cheap transistor radios and such. By 1972, IIRC, even these holes were mostly filled with Si parts. So obviously most small Ge production stopped in the 1960s.

Literally hundreds of thousands of Ge devices were made in the 1980s.  While the US and Europe had swung to Si in the 1970s, the eastern bloc and China stayed with Ge production (I presume because the factories could be low-tech).

Germanium leaks approximately 1000 times as much as silicon if all else is equal.

Again, RG is right on the money.  However, this information might make people think that Si intrinsically leaks 0.1 uA and Ge leaks 100 uA.  Not so.  Almost all of the leakage of older Ge devices (AC, NKT, etc) is associated with design and manufacture.  Modern Si devices often leak in the pico amps range (or maybe a nanoamp).  Well-designed and manufactured Ge devices leak a microamp or two. But were these devices ever made, and where can you get them?  They were made in the eastern bloc and China in the 1980s and  Japan in the 1970s. 

I have a couple of thousand soviet Ge transistors that typically leak a couple of uA in RG's tester.  I'd say that 99% leak less than 25 uA, which is only 1/10th of what many people consider the leakage limit (250 uA). These are GT308, GT313 and similar devices were mostly made from 1982 to 1985 in Russia.  The small number of Japanese devices that I have (50 or so) are almost as good.  It seems reasonable to assume that no matter where they were made, devices from the 1950s are extremely leaky, from the 1960s are very leaky, in the 1970s are somewhat leaky and in the 1980s are not leaky.  The DOA rate also falls dramatically.

Lastly, none of this is important if Ge devices don't offer any advantage in terms of tone.  IMO the 0.3V base-emitter voltage is important in the tone of circuits such as the fuzz-face.  An Si device in position Q1 of a fuzz-face can't and won't create the fuzz that a Ge device will (same hFE).  For that reason alone, there will always be a great love of Ge devices.

just my 2c...

[analogmike]

When I started in the 90s, and until recently, you could get germanium transistors from Germanium Power Devices in Mass. Founded in 1973 as a company manufacturing germanium transistors, in the last 10 years it has specialized in the production of detectors for infrared. Those are the "NKT275" you see in recent Dunlop fuzzface pedals with the G brand on them.



I don't find them useful for effects pedals.

DSI in Europe seems to still be selling transistors but not sure if they are still manufacturing them. They are not useful for fuzz faces but can be used for more extreme sounding pedals.



have fun!

[Eb7+9]

I don't find them useful for effects pedals.
They are not useful for fuzz faces but can be used for more extreme sounding pedals.

they're in the power-device biz, so maybe it's tough for their transistor process to keep a consistently low Is

what were they like in RM type circuits ??

[R.G.]

Mike's note brings up an interesting point. It's almost impossible to separate the poor processing available when "vintage" germaniums were made from what germanium itself does.

Semiconductor manufacturing today is massively improved over semiconductor manufacturing in the 1960s. For silicon alone, transistors today are so much better than they were fifty years ago that it's almost impossible to get devices as bad as they used to be. When you take modern standards for processing, even "ancient" processing done on used, left over and otherwise surplus processing machinery, it's so vastly better than germanium processing was half a century ago that you can't really compare a modern germanium to a "glory days" germanium. Modern ones are "too good" in the technical, industrial-use sense.

Making vintage style germaniums is conceptually easy - just don't do a very good job of refining germanium, diffusing the impurities, passivating the surface, or packaging it up. Getting the perfect mix of doing just the right amount of "not very good" may be challenging, though. We don't know what value of "not very good" is right for good sounding ones. And doing the technical work to find that out so you could make new just-ugly-enough germanium transistors would be expensive.

It's far cheaper to mine the obsolete parts bins of the world, and that is what the zillions of newly-minted effects experts have been doing for over a decade.

[frank_p]

For the curious.  (not BJT)
http://www.sciencedaily.com/releases/2010/02/100217093936.htm

[petemoore]

 Still have the pocket radio that still works, 7 or 8 Ge's in there, sounds good !
Somehow I managed to poke around inside an old radio with 3legged silver cylinders in it, enough to attach an input for guitar, that and a DIST+ made the cool project amp !