PNP vs. NPN ---?

[Jneely88]

In general. Is it true a PNP fuzz will have lower noise than a NPN.

I understand it is based on other factors. So let's just say your average fuzz face?

[R.G.]

It's not possible to say. There will probably be many answers on both yes and no sides, but the "other factors" have so much more weight than the NPN/PNP-ness that it's the wrong question to ask to get any useful information.

[Jneely88]

Hmmmm what is the right question?

[antonis]

Maybe something like: "Lower noise comparing to what..??" or "Is lower noise willing in a Fuzz..??" or "Lower audible or not noise..??"

[Transmogrifox]

Most of the noise I get from a fuzz usually turns out to be noise external to the fuzz from some other pedal or stuff coupled into the cable or power supply.  The noise from the fuzz itself generally is not too bad unless it's feeding into some other high gain monster (but then how much more gain do you really need?).

My take on RG's comment is that if you are trying to decide whether to build an NPN or a PNP fuzz, probably the noise characteristic difference between the NPN and PNP devices is not a factor that will make an audible difference in the noise performance.

If you're trying to make a low noise fuzz, then the right questions would be more general about what factors cause noise in a fuzz.

If you're trying to decide whether to make an NPN or PNP fuzz, then the right questions (again) would be more general about the pros and cons of NPN and PNP fuzzes.

As far as my ears tell me there isn't a difference when the NPN and PNP transistors are reasonably matched complements.  When they aren't reasonably matched complements, then the comparison isn't a valid test of PNP vs NPN.  It's just a difference between transistors and you can get the same variation between different PNP devices.

I would vote NPN as the low noise choice on a negative ground system.  I haven't experienced it myself, but there are a lot of posts/threads about oscillations and other power supply noise issues if converting a PNP fuzz to negative ground.  If you leave it positive ground, then you have to either keep it on its own isolated power supply/battery, or use  a charge pump inverter to get +/-9V.  All of these add complications, so I vote NPN.

[R.G.]

The right question is "how do I get a low noise fuzz-face style circuit?"

The answers are
1. Use a low noise guitar. No, that's not facetious. Guitars vary in how much noise they produce and send down the line.
2. Use metal film transistors of as low a resistance as you can. Actually, using wirewound is the lowest noise, but it's not all that practical. Carbon film is a bit worse, and carbon comp is the worst as it has excess noise that the others don't.
3. Use modern transistors; old, old PNP germaniums were made at a time when we were lucky to get transistors to transist at all. Those first-generation Ge devices were diffused and packaged with what amounts to stone-axe techniques, and many were not surface passivated so leakage currents ran over the surfaces of the pellets in the cans. The older a transistor is, the more likely it is to be noisy. PNP versus NPN germanium is only modestly an issue, in that decent NPN germaniums were much later than PNP and often better packaged.
4. Use silicon. Germanium leaks about 1000 times as much as silicon, and leakages always introduce some noise. Note that this is directly at odds with the prevailing wisdom that ffs should be made from germanium.
5. Make a quiet power supply with appropriate decoupling and wire routing.
6. Do not use reverse-polarity wiring to run PNP fuzz faces from a negative ground supply.
7. Minimize RF pickup.

For severe cases, you can graft on a noise gate.

[PRR]

> In general. Is it true a PNP fuzz will have lower noise than a NPN.

No.

Have you been reading books?

This assertion (PNP less hiss than NPN) appears in several semiconductor writings. Usually without derivation.

However if you skim the plans for many-many low-low-hiss amplifiers, designed by guys who actually built and precision tested, there seems to be no preference. Or the preference changes with impedance level and particularly with decade. There have been slightly-better devices over the years, and these do not always stay in production, so the "best" design depends on what low-hiss (for the impedance) device is available for production right now.

So what is that bumfodder in the books? It seems to associate hiss with intrinsic conductivity. Which is correct IF our device size is very small, and we can't readily (or economically) make it bigger.

But what matters is not intrinsic conductivity, but total conductivity. Intrinsic conductivity times area.

And since 1953, we can readily buy a 50mA device to do a 0.5mA job. Or a 500mA device to do a 2mA job. (And we can buy that excess size without the surface leakage which R.G. mentions.)

It's like: which conducts better, Copper or Aluminum? Well, in the same size, copper is better. Around 2X as good as Al. But copper is expensive. About 10X the price/pound of Aluminum; and Al is fatter per pound. If I re-ran my power line in new cable, for the same losses, Al is less than 1/5th the price of copper. (Same argument even stronger if I considered a Silver wire. Silver is better per diameter, but 100X the price.)

Getting back on the Fuzz track.... all of what R.G. sed. With modern clean transistors, in audio, the transistor contribution to hiss should be almost unmeasurable. It's mostly about the other parts. Both the simple circuit, and the too-too-real practical factors he enumerated.

I do think that with selected 1959 stock, or some of the late-production Russian Germaniums, reasonably low hiss may be possible. Not for-sure, or easy. But even with the bad old stuff (after we got past 40dB NF parts) preamps of 2dB NF were built routinely.

[Jneely88]

Ok . Yeah I had no real reason. Just been reading a lot. Wondered if it was true and worth a try.

[Renegadrian]

The main concern on the sili fuzzes is they use high hfe transistors, I never understood the bc108s in that, never liked them in the FF...you have to get some lower hfe trannies (I love pn2369, cheap and good) and put some smoothing 100p cap between C and B. A good addition on the FF is a 10k trimmer in place of the 8.2k and a 1k in place of the 330 feeding the trimmer. that will make a nice sounding fuzz.

As for the PNP trannies, I just finished my first rangemaster, the OC44 is good, but I got a very close result with a Si BC327...it costs less and it's more stable and predictable.