Which transistors for Superfuzz clone - if it matters?

[Shea]

I'm thinking of making a pedal around the Superfuzz circuit. One website says that 2SC828(Q) transistors were used in one particular make of the Superfuzz. I Googled 2SC828(Q), and pulled up a thread on another forum where it was described as a "plain jellybean" transistor. Whoever posted that compared it to the NTE 199.

Another website says to use 2N2222A or 2N3391 transistors.

So, my question is: does it matter which type of transistor I use, or will any ol' high gain NPN transistor do? I don't know much about transistors or how to pick substitutes for them. But I'm getting the impression that the transistors in this circuit might not be as critical to the tone as the transistors in a Rangemaster or Fuzz Face clone.

Shea

[Enzo]

Transistors are remarkably non-critical in most applications, as long as basic needs are met for current and voltage.

Most modern small signal silicon transistors still come in the TO92 size, that is the little half-round plastic guy. Asian type numbers - 2SC828, 2SD945, 2SA1815, any 2Sxxxx - have their legs arranged E-C-B, emitter-collector-base. American numbers such as 2N5210, 2N3904, and like MPSA06, MPS8089 have the arrangement E-B-C, emitter-base-collector.

The two types are electrically interchangeable if the specs are the same. But with the legs in different order, the parts would have to be mounted facing a different way and the leads not bent the same way. Keep that in mind. I mention it because a 2222 might replace that 828, but the leads might be in different order.

[Shea]

Thanks!

[Mark Hammer]

I would agree with Enzo, but also disagree.

The problem is that transistors can vary considerably from batch to batch, even though they have the same number printed on the outside. So, while it can often be the case that a certain tone depends very much on the transistors used, you have very little assurance that using a transistor of a given part number WILL necessarily deliver that tone. The key may be some particular property of the transistor that occurs with some regularity in that transistor number, but also occurs often enough in others. So, maybe 60% of 2SC828 trannies will meet the needed spec, and 40% of 2SC945s will. The 945 can sub for the 828 if you measure and confirm the spec, and in some instances, using 828s simply won't deliver if they are fro the 40% that don't meet the spec.

That's one of the reasons that DIY folks often recommend using sockets for transistors. That not only permits you to flip the pins around with ease (as per Enzo's very wise comment about Asian-American differences in pinouts), but lets you "season to taste" with transistors, by trying out other numbers.

It is somewhat impractical to do these days, given the vast number of devices out there, but years back you could easily buy transistor substitution books that would list all the commercially available devices that could sub for a given tranny. I have a couple of these, and you'd be surprised how many transistors, even 35 years ago, were deemed to be functional "substitutes" for others. Of course, exactly what set of parameters the substitute was subbing for was never described in the listing. What you could basically derive was that the tranny wouldn't melt down if inserted into that same spot in the circuit. - more of a survival criteria rather than a tonal one.

Keep in mind that many "classic" Asian-made pedals, although produced in reasonably large numbers for a number of years, were not made in humungous volumes (hence a smaller lot of trannies may have been bought from a single supplier). They also could be quite variable from unit to unit (though this is more true of those using germanium devices than silicon). Finally, the service manuals and schematics never included any recommendations about transistor specs other than part number. Remember, this IS rock and roll, and there was never any perceived need to assure adherence to some consistent tonal quality. It was always "discover a nifty sound" -> "draw up the circuit" -> "stuff them boards and ship'em out".

[Shea]

Thanks, Mark. So, the lesson I take from your post is that if I try this, I should put transistor sockets on the board, get a whole bunch of transistors, and see which work best by trial and error.

I'd have to get a bunch of transistors anyway, so I could match a pair for the differential amp in the octave circuit.

[Mark Hammer]

Well, a little more refined than that. I would recommend at least having a DMM on hand that lets you measure transistor hfe so you have some reasonable preview of what to expect. Not that the transistor hfe is a guarantee of anything, but usually you can learn a fair amount by measuring your trannies' hfe and listening for what difference that makes. So, for instance you may find that, even when the circuit calls for the same transistor in each of two positions in the circuit, what you get when the higher hfe one is in the second position may be preferable to what you get with it in the 1st position. Same two transistors, but different specs in each place.

Even if you place an order with a distributor and order, say half a dozen transistors of the same part number, there may be variation between them, and it is helpful to know what that variation is. In a great many instances it won't be important, but in the odd instance it WILL be.

[dai h.]

FWIW, when I did some websearching on 2SC828s, I found they were a plastic package version of a metal can one which was equiv. to Euro BC107. And for (the metal can package) BC107/8/9 there are corresponding plastic package versions (i.e. BC147/8/9). (Assuming that they give the same results,) maybe it makes sense for instance that Fulltone 70s and their wahs use 828s(since original Fuzz Faces and wahs used BC107/8/9). If there isn't anything special, then that should mean there are many more (gen. purpose transistors) to choose from.

re: pin outs of japanese transistors: some are the same as BC107/8/9. I'm thinking these are probably older designs?

also, there was something I vaguely recall (which I may not be recollecting correctly) about designers back then using RF transistors for inputs for low noise(?--easier approach with transistors back then), therefore a possible explanation why you might sometimes see a bunch of regular ones then a different one for the input (Univibe). So if trying to reproduce the effect today, maybe just use a low noise non-RF (assuming low noise was the goal and not some special quality that came from using RF transistors) and get the same result (instead of killing yourself trying to find exact old stock transistors).

sorry that I can't tell you if specific ones matter in this case or not--I wish I knew (would be interesting to know).