Solving the noise in Germaniums

[pqt_bach]

Hello,

Is there any method to quiet down germanium transistors?

I heard that potting is one way, but I can't see how it changes anything. Are certain Germanium trasistors noisier than others?

thanks

P.Q.

[R.G.]

There is no way to "solve" the noise of germaniums or any other active device, other than using a different and less noisy device.

Individual germanium transistors vary widely in the amount of noise they produce.  Notice that I did not say "different type numbers" but "individual tranistors".

That's because the origins of noise in transistors was poorly understood when germaniums were in vogue, and semiconductor processing procedures that would have helped were also in their infancy, so germaniums varied - a lot - even within one type number. Because they were made later, the higher the type number the better the device, but that's an awfully crude yardstick.

In many cases, germanium hiss is a product of surface contamination on the transistor pellet itself, caused by humans not yet knowing to glass-passivate their semiconductors, and germanium oxide not being a good passivating glass.

It's easy to "hear" things on the internet. We're developing a new saying, I think. "Lies, damned lies, and then the internet".

[Sir H C]

Sometimes you can increase/decrease the bias current and find where the transistors might be less noisy, but often that is only when the current is 0.

[markm]

Sometimes you can increase/decrease the bias current and find where the transistors might be less noisy, but often that is only when the current is 0.

[Gilles C]

I am wondering if the military versions of the germanium transistors have better specs (less noise) or if the number is only different?

Gilles

[Phorhas]

I've tested some CV counter parts of quite a few OC transistors and result weren't conclusive. Although most CV (mil spec) had better leakage/gain characteristics, the amount of noise varied from quiet to noisy like h*ll even with the Mil spec parts.

D

[pqt_bach]

Yeah, I figured thjat the potting thing was web-myth

So... nothing to do but going through a large pack of trannies... no "after-market" solution, huh?

[R.G.]

Well, you see - taking a large batch and testing for which ones were good enough at several different levels was the industrial solution. That's what the makers of germanium transistors did. They could not hold tolerances tight enough, so they made a batch, sorted it into batches, and labeled each batch.

Once again, the maker guarantees ONLY what is on the datasheet. If you look at the datasheet for older devices, it probably would not surprise you to find no specification for noise, would it?

[DDD]

The gain to noise rate in Ge amp stage heavily depends on the operating (collector) current of the stage. The leakage current of the transistor makes the main tribute to the noise figure. Since the leakage current is proportional to the collector-to-base voltage and practically does not depend on the collector current, you need to reduce the collector-to-base voltage and make the collector current higher. So you'll make the gain higher with the less noise.
Actually the above procedure is quite complicated by itself and, moreover, it might drastically change the distortion characteristic of the stage.