Noise Reduction in high Gain circuits - Metal Film?

[wizard333]

How much of a difference can you expect to hear in noise/hiss reduction with all metal film vs carbon film in a high gain circuit? Worth it, or "audible, but not enough to care"?

[tubeswell]

Yes IMHO the difference between hiss level using metal film and carbon film is audible all other things being equal (like resistor power rating and gain). Increasing the power rating also reduces hiss, and so does reducing gain.

[guitician]

That all depends on the circuit. If your gain is very high like a phono input stage you'll hear more of a difference than a cascade of stages like are used for guitar. The higher the resistance the more noise there will be. But remember that the gain devices themselves have resistance and can bring down the noise level, or add to it. All factors have to be weighed and a choice made based on them.

[mooreamps]

Carbon comp resistors will hiss like a snake. Apparently the most popular approach is to go with the metal film. Myself, I don't use any resistor in the cathode circuit, instead using fixed bias.


-g

[Hardtailed]

I'm guessing, considering the little different in price, going with metal film is cheap insurance when building high-gain amps?
I'm talking mostly about preamp stage, mainly the plate load resistor, and cathode resistor.

I'm planning a high-gain build and I've factored 1W metal film resistors everywhere in the preamp.

[wizard333]

Do you get more noise reduction from going from say, a 1/2W Carbon Film to a 1W carbon Film, or going from a 1/2W Carbon Film to a 1/2W Metal Film?

i.e., Does the composition of the resistor make more/less difference than the power rating?

Also, do you get more bang/buck in series dropping resistors (i.e. stuff directly in the audio path) or still get a big effect from things not directly in the audio path like RKs or Plate resistors?

Thanks!

[Hardtailed]

Do you get more noise reduction from going from say, a 1/2W Carbon Film to a 1W carbon Film, or going from a 1/2W Carbon Film to a 1/2W Metal Film?

i.e., Does the composition of the resistor make more/less difference than the power rating?

Also, do you get more bang/buck in series dropping resistors (i.e. stuff directly in the audio path) or still get a big effect from things not directly in the audio path like RKs or Plate resistors?

Thanks!

You can get 1W 5% tolerance metal film resistors for less than 0.20$ each at Mouser, why bother with anything else if the lowest noise floor is what you're after?
That said, some will say the carbon has more "mojo" (actually, it's because it's less linear, but that's another debate), I haven't done enough testing to confirm. But I did build my Marshall 50W clone with carbon film all around (didn't go as far as to bother with carbon composition) and it's quiet enough, but it's not exactly high-gain.

Note that the plate resistors are technically in the audio path. Remember that a tube modulates current, not voltage. The drop accross the load resistor is what becomes the voltage signal you use to drive the next stage. So yes it makes a difference. Actually, I believe that's where high-gain builders will switch to metal film first (the cathode resistor being a comparatively low value, not to mention it is bypassed by a cap, will have less effect)

[Joey Voltage]

Increasing the power rating also reduces hiss, and so does reducing gain.

This only makes a difference if the power rating dictates the physical size of the component (it often doesn't these days), otherwise it does you no good.

[Steve Conner]

I think new production metal film and carbon film resistors are all decently quiet.

All resistors generate thermal noise (aka Johnson noise) in proportion to their resistance, and the quality of a resistor is measured by how much excess noise it generates, over and above this.

It's just the old carbon comps you have to watch for, that "hiss like a snake" as Mooreamps said. They have a lot of excess noise. I don't know why people pay good money for NOS carbon comp resistors.

If you've ever whiled away a Friday afternoon at work pondering whether the fridge light really goes off when the door is closed, you might like to try pondering whether resistors still generate Johnson noise as they're sitting in a drawer.

[R.G.]

I think new production metal film and carbon film resistors are all decently quiet.

+1

All resistors generate thermal noise (aka Johnson noise) in proportion to their resistance, and the quality of a resistor is measured by how much excess noise it generates, over and above this.

+1

It's just the old carbon comps you have to watch for, that "hiss like a snake" as Mooreamps said. They have a lot of excess noise. I don't know why people pay good money for NOS carbon comp resistors.

Tee-hee. Because they're told to... These aren't the droids you're looking for...

If you've ever whiled away a Friday afternoon at work pondering whether the fridge light really goes off when the door is closed, you might like to try pondering whether resistors still generate Johnson noise as they're sitting in a drawer.

Actually, that has been studied. Yes, they do!

On a more serious note, re the whole issue addressed in this thread:
1. As a general principle, for any amplifier of significant gain (i.e. over about 10) the first stage noise dominates the noise of the whole unit unless there is a serious flaw in the design or fault in later components in the chain. This is because the noise at the very front end is amplified by all the following gain, and the noise from later stages is amplified by less gain.
2. There *is* a difference in tone quality with carbon comp resistors as well as a difference in noise. However, understanding that the tone difference is because of the voltage coefficient of resistance lets you notice that this is only apparent where the signal excursion across the resistor in question is upwards of 75-200V peak to peak.

Taken together this leads to the understanding that
(a) Carbon comp resistors in the first stage of an amplifier add noise, not tonal quality. You get all the noise, none of the "good" distortion. Mother Nature says that this is a losing proposition.
(b) Carbon comp resistors in later stages add less noise, but not much tonal advantage until the signals get *big*. So it doesn't make sense to pay extra for carbon comp resistors to add some noise, even if less, if they don't get you more distortion-goodie.
(c) Carbon comp resistors work best where the signal swings are biggest and where they are latest in the amplifier. Guess where that one position is...

[mooreamps]

On a more serious note, re the whole issue addressed in this thread:
1. As a general principle, for any amplifier of significant gain (i.e. over about 10) the first stage noise dominates the noise of the whole unit

RG makes a very good point here.. It is even more true when applied to radio frequency receivers. The noise floor of the first gain stage in a radio receiver sets the noise level for the entire unit. The same seems to hold true for a tube amp.


-g

[tubeswell]

The same seems to hold true for a tube amp. -g

Of course it does. One of the fundamental principles in recording music is to set the gain up so that you minimise the noise floor at the beginning of the signal chain. No good having the signal gain-to-noise ratio really low, because when it gets amplified later in the circuit, the noise portion is merely increased. That's why you pick a mic with a really low noise floor and you get as much of that signal as you can as early as you can.

[mooreamps]

Of course it does. One of the fundamental principles in recording music is to set the gain up so that you minimize the noise floor at the beginning of the signal chain.

Precisely why I pull all the cathode resistors out, and use fixed bias. Then, the preamp gain can be set at any level since there is no noise floor to contend with.


-g

[wizard333]

Uh, what?

[Gaz]

Myself, I don't use any resistor in the cathode circuit, instead using fixed bias.

Precisely why I pull all the cathode resistors out, and use fixed bias.

Yeah, I think I read that somewhere on the internet.