Regarding carbon composition resistors in tube amps, ilimzn quoted below: "If you see a low value carbon comp resistor..........in an amp where everything else is more 'modern' - leave it there. The key is that CCs have almost no seies inductance and this is critical in some places such as grids of triodes and gates of MOSFETs."
Yeah, ummm, not so much. Let's apply some physics and electrical engineering to debunk that.
Here's what I previous wrote on the subject, slightly edited.
The purported inductance of the metal-film resistor
does not matter at audio frequencies. If one is going to claim it matters, then some supporting data is required.
Here is the counter argument.
The metal-film resistors have nH of inductance—
nano Henries—which is essentially zero inductance for our purposes, and won't matter at any frequency encountered in audio. Those nH of inductance do matter at hundreds of MHz or GHz, but at those frequencies even the inductance of PCB traces becomes a design factor. Never at 20 kHz, which is essentially DC as far as the resistor is concerned. The inductance of component leads (resistors and capacitors), sockets, and chassis wiring dominates and is greater than the metal-film resistor.
The problem of parasitic inductance of components only arises at
hundreds of MHz or low GHz where tank circuits are formed from the tiny parasitic inductance and capacitance from the wiring and circuit-board traces. Even the inductance of decoupling capacitors matter at those frequencies, as different value capacitors start forming tank circuits with each other. This is why the ground planes can oscillate in computer motherboards and cellphones, and why special high ESR capacitors are used to damp such oscillations. But at audio frequencies? Naaah. Basic electrical engineering and physics explain this.
If money were no object, then wirewound resistors, with greater trivial inductance per unit than metal film, would be the best choice as these are the lowest noise. Wirewound resistors cost a great deal more than metal film, which costs a tiny amount more than carbon film, which costs a tiny amount more than awful carbon-composite resistors. See:
This price disparity was greater when the amplifier was made, although metal film did not yet exist. Carbon film, aka "cracked carbon", was the premium component, and back in the day these cost ten times what ordinary resistors cost.
Even a conventional wire-wound resistor, with inductance typically under a µH—
micro Henries—sometimes as much as tens of µH, doesn't matter for audio
provided the amplifier does not oscillate. That is key. I must qualify this statement because many audio amplifiers oscillate > 100 kHz, usually > 1 MHz and the designer (or owner) is blissfully unaware of this. A similar problem exists with high slew-rate opamps which oscillate in the same region. So, yes, a wirewound will affect that oscillation, but using a component in a poorly designed circuited and blaming the component isn't fair. (Low value resistors, like the grid and screen stopper, ruin the Q of the circuit and damp that oscillation. Zobels can also be used for this purpose. Removes a lot of lost power and noise.)
The resistors in question have inductance of nH to uH, yet the filters in the circuits we use (such as crossovers) have mH inductors—
milli Henries—which is a
thousand times (10^3) greater than µH and a
million times (10^6) greater than the nH:
milli (m) = 0.001 = 10^-3
micro (µ) = 0.000001 = 10^-6
nano (n) = 0.000000001 = 10^-9
How could such a trivial inductance have any effect? (Provided, again, that the circuit (a) is not being used at high frequency and (b) does not oscillate at high frequency, of course.)
Audiophiles, who generally are not electrical engineers or physicists, talk a lot about nonsense read on the interwebs and then endlessly regurgitated until it becomes unquestionable dogma, without any understanding of the underlying engineering or that the original claim was not just unfounded but contrary to basic principles. People instead just nod and say, of course, when the issue is not grounded in reality.
The fact remains that if the trivial inductance of a metal-film resistor mattered at the frequencies in question, then the inductance of the wiring itself would also matter. It does not.
But believe it or not, greater knowledge than yours does exist (and/or has been forgotten), and you are not infallible. By all means, please have the last word (or last 10,000 words as the case may be). I will continue to have confidence in the wisdom of the post I quoted initially, and if that post helps one person somewhere down the road, then the effort will have not been in vain.