Transistor selection based on hfe

[Yazoo]

I knew germanium transistors had widely varying gains (hfes) within a specific type but I've always wrongly assumed that silicon transistors were fairly standard. Now I've actually tested some for a Tycobrahe Octavia, I've seen how widely transistors of the same type can vary. I've built several pedals and I think I've just been lucky that they all worked.

So the next question is for a given stompbox, how can you know which transistor to use, e.g. how would you pick a 2n5088 transistor with the right gain for say a Big Muff. A list of recommended gains for specific effects would be great if such a thing exists.

[LucifersTrip]

I knew germanium transistors had widely varying gains (hfes) within a specific type but I've always wrongly assumed that silicon transistors were fairly standard. Now I've actually tested some for a Tycobrahe Octavia, I've seen how widely transistors of the same type can vary.

the biggest differences will usually be between vintage & newer versions of the same transistor. for instance, the new 2n3906's have hfe's ~ 200-300 and the vintage ones I have are in the 100's. It is more common that the newer ones have higher gains

I've built several pedals and I think I've just been lucky that they all worked.

So the next question is for a given stompbox, how can you know which transistor to use, e.g. how would you pick a 2n5088 transistor with the right gain for say a Big Muff. A list of recommended gains for specific effects would be great if such a thing exists.

If you use non-defective transistors that are called for in the schematic, it should work...but the different ranges in gain will of course give you different tones (sometimes)...and change the voltages (see below).

In the end, it is only your ear that counts. Experimenting with the different ranges with each transistor is part of the DIY fun. If you don't enjoy that part, do searches here, and you will porbably find recommendations.

Also, note that the voltage readings of the transistors (ebc) are many times better indicators of overall sound.  Ie, try to get the recommended voltage readings first, then worry about tweaking with different gains.   You will most likely have to alter surrounding resistors when using different gain transistors to get the desired voltage.

[R.G.]

I knew germanium transistors had widely varying gains (hfes) within a specific type but I've always wrongly assumed that silicon transistors were fairly standard. Now I've actually tested some for a Tycobrahe Octavia, I've seen how widely transistors of the same type can vary. I've built several pedals and I think I've just been lucky that they all worked.

Maybe not so lucky. Germanium transistors in general and the fuzz face in particular are sensitive to hfe and leakage. Silicon sidesteps leakage entirely since as a material it is about 1000 times less leaky. And about the time silicon came in, EEs were heavily pushed to figure out circuits that made gain variations not matter much if at all, so long as it was enough.

It was a different KIND of circuit, as well as different transistors.

So the next question is for a given stompbox, how can you know which transistor to use, e.g. how would you pick a 2n5088 transistor with the right gain for say a Big Muff. A list of recommended gains for specific effects would be great if such a thing exists.

Now that you're all super-sensitive to the issue, it may not matter. It is (probably) true that low gain silicons, from 40 to 100, do *exist*, and in fact were much more common before silicon diffusion expertise became a neverending gold mine. But modern devices are uniformly higher gain, sometimes greatly so.

And there is that circuits thing again. For a surprisingly large number of circuits, the actual gain of the transistors used makes a very small, and sometimes vanishingly small difference. I'm guessing that one of the next few posts will say "yes, but I used hfe = 80 silicon devices in my [insert pedal name here] and it made a big difference in sound."

Maybe so. It's truer for gains below 100 than above. And over 200, the actual current gain makes a rapidly diminishing difference.

And for the 2N5088 especially, it makes little difference, to the point that I don't know (a) how to tell you to pick one gain over another in most circuits and (b) how you would reliably *measure* the gains in a way that means anything.

You'll undoubtedly get a lot of recommendations for best gains for X circuit. Unless you're more interested in honing and polishing pedals than playing, don't get too wrapped up in "perfect" gains for silicon.

[lascaux]

In terms of a list, another variable I've read about – if someone would like to debunk this I would welcome it – is that germanium was temperature-sensitive.
In other words, if you were looking for a particular sound, e.g., from Hendrix's 'Red House' in San Diego, CA, you might have to look in a 1969 Farmer's Almanac to see what the weather was like.
[This doesn't seem to be too well documented, let alone a popular idea; and I would have to hunt for that source to cite it.]
Either way, it does seem like the "science of fuzz" has a lot of fuzziness to it.

[R.G.]

Less than you'd imagine. The temperature sensitivity of germaniums is well documented, and was mentioned in the first Guitar Effects FAQ at geofex, back in the 90s. It was so well know that it was being forgotten as ancient history before that in the EE community.

[tubelectron]

It's truer for gains below 100 than above. And over 200, the actual current gain makes a rapidly diminishing difference.

I agree : I noticed that experimentally while testing different transistors with different Hfe on a basic Si Fuzz design for evaluation.

I usually noticed also that "the more the gain the more the fizz" but it may vary along with the transistor model, i.e. in my tests, a 2N5088 hfe 400 gave less fizz than a MPSA18 hfe 900, but gave more fizz than a BC109C at hfe 390, while in term of saturation amount, the difference was little. The PSU voltage and passive components remained the same for all tests.

I tested 2N1613 hfe 40 with no fizz and much more smooth tone, but a bit less saturation and sustain. The frontier seemed to be hfe 100-130 (i.e. 2N3904).

A+!