All-Around "Best" Ge Xstr?

[Joe Kramer]

Hi Friends,

Just an informal opinion poll.   Yes, "best" is very subjective and conditional, but with all the experience around here, maybe some kind of consensus has emerged.  If you had to pick one "desert-island" Ge transistor, which would you say is the most generally all-purpose, more-or-less foolproof, decent-sounding, works-well-75-out-of-100-times type?  The 2N3905 or MPF102 of Ge transistors, if you will.  PNP or NPN, no matter.  
 
Regards,
Joe

[Ge_Whiz]

IMHO, NSB.

(No Such Beast).

[Stevo]

One is a general purpose silicon trans and one is a feild effects transistor MPF 102 is n channel fet and 2n3905 I think is npn silicon trans....If you want some good germanium trans go to AMZ-FX I like the ones he sells

[R.G.]

The only reliable, good-nine-times-out-of-ten germanium?

A preselected, pretested one.

All others are a one in 3 - or worse! - crap shoot. The type number holds only moderate indication.

[nelson]

what about military spec germs? werent they made to higher tolerances?

[R.G.]

what about military spec germs? werent they made to higher tolerances?

They were **tested** to whatever specs the military needed. If the military happened to order a batch that by chance were the specs you need for a fuzz face or whatever, and if by chance the transistors did not degrade in the intervening forty years (glass passivation was unknown at the time), then they may be fine.

You gotta understand - germaniums were HAND made. They took a melt pot of germanium and the operator zone refined the pot until it was pure enough, then the operator ran the crystal puller to pull out a boule of germanium. After that the operator sawed off slices of germanium, and when that was done, diced the slices up into dies with a diamond saw.

Then the operators stacked little sandwiches of indium/germanium/indium and put them on a cookie tray in the diffusion oven, twisted some gas gauges to the right (?) concentration of diffusion gasses, and cooked them for the right (?) time while watching the oven temperature carefully.

Once the batch came out, they tested the sandwiches. Some were high voltage, some low voltage, high gain; some batches were all high leakage if we didn't watch the oven atmosphere very well. The good ones were selected out and manually glued into metal cups and wires bonded to the die.

It was common for transistor makers to not be able to make enough transistors to a given military spec.

Semiconductor fabrication was unbelievably crude in the germanium days. The wonder is that they worked at all. It really was like baking cookies.

[nelson]

wow, that clears up alot.


One question R.G what is glass passivation?

[barret77]

hmmm... that's a great hook for some marketing hype "hand made custom-baked mojo germanium transistors"


by the way, thanks RG for adding some information to my brain.

[R.G.]

One question R.G what is glass passivation?

One of the things that Mother Nature taught us as a species about semiconductors is that pure crystals of anything exposed to highly reactive nitrogen-oxygen mixtures at temperatures over 270K ( that is, air at room temperature) tend to get schmutz on the outside.

In the case of germanium, the schmutz is (a) partly conductive and (b) noisy so it makes germanium parts work and sound bad. This is where the phrase "heremetically sealed" came in with semiconductors. The first way we thought of to deal with Mother's little problem is to fill the little cans with inert gasses and seal them up f-o-r-e-v-e-r. At least we tried. Sometimes it worked.

When planar processing (Yah! We'll diffuse a bunch of transistors onto one slice of semiconductigook before cutting it into individual dies!) came along, it also got people to thinking about cleaning up the dies after they were sliced. Near as I can tell, it was only with silicon where some bright guy said - "Hey, if we put this back in the oven with some really pure oxygen, wouldn't it grow a layer of glass all over it?" Germanium is chemically different from silicon; it may not make glass, or its oxide may not work that way.

But glass passivation is the process of making the surface of a semiconductor chip inert (that is, passivated) by growing a layer of silicon dioxide all over it. Seals out the contaminants better than we can at the package level, and coincidentally makes plastic packaged parts **possible**. They were'nt before that. No plastic packaged germanium.

by the way, thanks RG for adding some information to my brain.

My pleasure. The store is always open.  

[Joe Kramer]

The only reliable, good-nine-times-out-of-ten germanium?

A preselected, pretested one.

All others are a one in 3 - or worse! - crap shoot. The type number holds only moderate indication.

Given a choice between, say, an OC44, a 2SB54, and a 2N388, you would have no preference whatsoever?

[petemoore]

Given a choice between, say, an OC44, a 2SB54, and a 2N388, you would have no preference whatsoever?
 The part #'s arent really the important ones, "What' sounds not bad is low leakage.
 What sounds great is a matter of taste.
 What sounds great [and of course is most reliable] is also the most definitive test for an audio transistor application.
 Build a batch of circuit into which Ge transistors can be auditioned.
 EZ face is a hybrid Ge/Si, great sounding circuit and great for Ge auditions.

[squidsquad]

Thanks RG for some fascinating insight!

[R.G.]

Given a choice between, say, an OC44, a 2SB54, and a 2N388, you would have no preference whatsoever?

Honestly, no. Beauty and ugliness both lie hidden under the (supposedly) heremetic seals of those devices.

[aron]

archive

[MartyMart]

RG, thanks again for "more valuable" info
As far as "quick test and pop 'em in" my OC140's (NPN) are "almost"
foolproof.
I've found three "bad ones" out of around 50 ! the rest are perfect.
These are "military" spec, same as small bears.

Marty.

[petemoore]

Low hfe Q1? Bigger gain R ?
 revoice ?
 I built a FF board with trimmers/sockets all over it, same for Rangemaster...may help if you only have a few Q candidates for those positions. Only the 100k was fixed, and was made 150k, but has lead length above board to easily trim it with a fixed resistor or 500k pot.
 I think for your or my purposes, the discussion is limited to what is on hand, and of those Q's which sounds best after 'preparing circuit work areas' for each one, tweeking the circuit to 'fit' 'X' transistor[s.
 'These things work differnt"
 If you have enough transistors, you probably need all that ^ even less.
 I jsut did all that to help me learn what 'this' resistor does when it's bigger, and to be able to tweek in limited candidates of Ge's...
 circuit compensation [tweeks] for working with different sounding and gain Hfe Q's yields 'some' different tones, this and tweeking the gain of the transistors by swapping, piggybacking..etc. may help you find 'your' FF.  
 My FF testboard also has an adjustable YAFF mod...Q1E > 100ohm to ground.

[Joe Kramer]

As far as "quick test and pop 'em in" my OC140's (NPN) are "almost" foolproof.  I've found three "bad ones" out of around 50 ! the rest are perfect.  These are "military" spec, same as small bears.
Marty.

Hi!

Thanks for your reply Marty.  This was the kind of answer I was hoping for, with all due respect to RG and others.  I'm still a little surprised no one else has offered up testimonials about their most prefered Ge's.  I thought it was a pretty straightforward question that would have yeilded some interesting/informative results.  

BTW Marty, what make are the OC140's?  

Regards,
Joe
 
PS: Maybe I should have posted a starter preference: I happen to like the 2N13XX's a lot, though they're on the bright side.  They seem to behave almost as consistently as Si xstrs. . . .