I'm building a Tone Bender Mk2 type 3 transistor fuzz pedal with old germanium transistors. Currently I'm swapping the Q1 transistor between various transistors I have. Some produce white noise like a gale down a chimney, others mostly quiet. These are for transistors with similar gain and leakage ( around x25 , 0.02mA ).
I can appreciate as you add more gain, the noise goes up, but why the huge difference for similar gain? Is it to do with their freq response, so the gain my tester measures is not at the freq of the noise?
I can appreciate as you add more gain, the noise goes up, but why the huge difference for similar gain? Is it to do with their freq response, so the gain my tester measures is not at the freq of the noise?
Ge. transistors are very old and used tin in their production.
Over time, tin 'grows' tiny microscopic whiskers and in the end will form a short circuit; as we find with circa 2010 BGA lead free soldering.
Whilst the 'whiskers' grow, they form leaks in the flow of random electrons when powered up. The Hfe cannot be accurately measure with a transistor tester.
This is why Ge. components are obsolete.
A silicon BC159 will work just as well.
Over time, tin 'grows' tiny microscopic whiskers and in the end will form a short circuit; as we find with circa 2010 BGA lead free soldering.
Whilst the 'whiskers' grow, they form leaks in the flow of random electrons when powered up. The Hfe cannot be accurately measure with a transistor tester.
This is why Ge. components are obsolete.
A silicon BC159 will work just as well.
Thanks for the explanation. I thought the whiskers would show up in the leakage measurement, but I can see they might also be resulting in white noise. I have a 1960s Siemens Ediswan transistor tester that you can set the load current on. It is pretty good at getting Hfe.
Perfectly clean semiconductor has a characteristic noise level. This can be computed from physical constants.
Dirty (impure) semiconductor has a higher noise level.
Early Germanium's noise could 40dB (100:1) higher than theoretical.
Sorting was costly and would yield a few "better" parts which were snapped-up by telephone, hearing-aid, radio, and audio factories.
It was 20 years and another technology before we could routinely expect noise within a few dB of ideal.
Dirty (impure) semiconductor has a higher noise level.
Early Germanium's noise could 40dB (100:1) higher than theoretical.
Sorting was costly and would yield a few "better" parts which were snapped-up by telephone, hearing-aid, radio, and audio factories.
It was 20 years and another technology before we could routinely expect noise within a few dB of ideal.
That's what you do.
This is likely mainly due to (surface?) defect density. Old manufacturing lines were nothing like as clean as modern ones and I think noise mechanisms were less well understood too. Basic Ge purity could be a contributing factor too - semiconductors really want to be ultra-pure (think 99.9999999% or better) for the best electronic properties.