Using CMF55 Resistors to reduce potential noise... good idea, or pointless?

[Snufkinoob]

I've done a little searching, but haven't really found an answer to this.

Would there really be any point in using CMF55 resistors over 1% Metal Film resistors when building a typically noisy/hissy fuzz like a Super-Fuzz for example? Does their reputation for noise reduction extend to silencing noisy 9v fuzz circuits with clear results... or would it be a total waste of money?

[R.G.]

CMF55 are a variant of metal film, I believe. As such, the noise characteristics should be similar.

In general, getting away from carbon composition is the big step in combating resistors noise. Carbon film is quieter than CC, which is good, because CF is now cheaper! Once you go to metal film, there are minor differences in metal film resistor variants, I'm sure, but the differences are trivial compared to the step away from CC.

Very often, fuzzes with a rep for being noisy and hissy are that way because of the design of the circuit more than using/not using quiet resistors.

[Snufkinoob]

Yeah, they are essentially like "boutique" metal film resistors.

So they're probably useless with fuzz that uses a lot of components, and/or is noisy by nature, but what about in smaller circuits that would only use 5-10 resistors, and maybe not a fuzz? Does anyone swear by CMF55's for simpler stomp-box applications to reduce/eliminate even slight background noise over standard Metal Film?

I'm thinking they'd be useful in a simple booster, but perhaps not much else?

[Tony Forestiere]

Buy a couple few in standard values (1,10,100,1000, R,K,M...etc.) and test against similar values of Carbon film, Metal film, and your "Boutique Test Subjects" on an unshielded breadboard. I doubt you will hear much of a difference.

Stick some Carbon composition resistors in the same circuit, and you will hear a difference in hiss and thermal interference.

[R.G.]

Here's the problem: noise comes from many places, and it all adds up. When you reduce the noise from one source, it only helps until the lower noise from that source gets less than some other noise source, and then using zero-noise parts would not improve it more.

In my opinion, there is very little reason to go hunting for fancier metal film resistors once you've gone to metal film, no matter what the circuit. And since the input stage noise is amplified by everything after it, the noise in the input stage is critical; the rest of the circuit to a much lesser degree.

And that does mean that the higher the gain from input to output, the more the input noise is amplified. A simple booster with low gain will be less noisy when made from the same parts as a high gain fuzz because the noise that exists is amplified less.

A high gain fuzz is a particularly noisy thing because ordinary signal is amplified up and then clipped off, so the amount of signal it can put out is limited. But the noise at the input is small, but being small it can be amplified by the full gain of the circuit without being clipped, so the *relative* amount of noise is increased much more than the increase in signal.

The EE profession (at least those parts of it that still think about noise) holds that the quietest resistors are not metal film at all - they're wire wound. Less surface to have surface effects to generate noise for the same resistance.

And down at the bottom of all this is thermal noise. There is an irreduceable amount of noise from any conductor that depends on its resistance and its temperature. You can't get a conductor (hence, resistor) with lower noise than the thermal noise of that amount of resistance, no matter what it's made of. The only options are to use smaller resistances or keep it cooler.

So, IMHO, you're off in the tall grass worrying about where CMF55s would be most useful. Worry instead about getting quiet transistors, and using them in circuits which minimize the noise contribution of the biasing network, and prevent them from EVER having the base-emitter be reverse-broken, even once, because that permanently degrades the transistor's internal noise. It's one reason older electronics are often hissy, and why whipping in a fresh 2N5088 often makes them quieter.

[Snufkinoob]

Interesting read. Thanks.