Where to use metal films?

[jrfrond]

Well, we know there is a scientific basis to the carbon comp resistor mojo thing: they generate sweet-sounding 2nd harmonic distortion when hit with high AC voltages.
http://www.geofex.com/article_folder...carboncomp.htm

If you want some of that sweetening, you can sprinkle some carbon comps in places where they'll see lots of AC signal voltage. RG recommends the plate resistor of the last stage before the PI.

Metal films don't distort, so I suppose they might well sound bland in comparison, or the lack of sweetness could seem harsh. But they are quiet and stable.

Exactly Steve! In addition, due to their "block" rather than film construction, they are in fact much hardier and less prone to suddenly opening up under heavy transients. This is one reason why 1/2W. CC plate resistors have survived for so long in so many amps.

[tedmich]

yes, yes CC mojo, even order harmonics, last long enough to drift hugely...

[diagrammatiks]

I was reading some history on carbon comps the other day and it seems that the types of carbon comps that get the bad rep, the allen-bradleys had especially terrible quality for their consumer line.

probably why fender was able to get them so cheap.

I've got 5 band mil-spec rcr20/32s that are matched almost perfectly.

[Chuck H]

I know the CC mojo is linked to higher voltages, but I just can't bring myself to use them under these conditions. It's usually these conditions that are most prone to cause noise issues with CC's. The last project I did had a request for CC's attached. So I used CC's for most of the amp. For plate loads, slope resistor and any high series resistance I took the liberty of using MF. The amp sounds great and is quiet (enough).

It seems like most builders prefere CF resistors. I don't know why. Is it because they aren't metal film and they do have "carbon" in their moniker?!? Or is there some advantage to CF's over MF's for tone. I don't expect that CF's have any of the mojo of CC's simply by virtue of carbon being used.?. So why? Are they the best? Why?

[jpfamps]

I use carbon comps, metal oxide or MF resistors depending on the amp/ position in the amp.

Repairing vintage amps, I would generally use what's been used in the amp already, so that would be carbon comps in vintage Fenders, except in parts of the amp which in my view would benefit from a more reliable part, eg the power supply dropping resistors.

In amps I build, I will tend to use MF resistors in the early stages for lower noise and CCs later in the amp.

I don't use carbon film as I'm not sure that there is any benefit over using MF, although obviously if you are producing 10,000 units a month then there is a cost saving using CF.

I have always assumed that the "issues" with carbon comp resistors are due to their method of construction NOT the fact that they employ carbon.

[GregS]

It seems like most builders prefere CF resistors. I don't know why. Is it because they aren't metal film and they do have "carbon" in their moniker?!? Or is there some advantage to CF's over MF's for tone. I don't expect that CF's have any of the mojo of CC's simply by virtue of carbon being used.?. So why? Are they the best? Why?

I think that many believe that CF are a good compromise between the mojo of CC and the low noise, tight tolerance of MF.

I believe CF are quite good for low noise and tolerance. How much better MF are for this I'm not sure. The amps I built where I did the grounding right are very quiet with CF resistors so I'm not sure MF could be much of an improvement in this regard - maybe for really high gain amps.

I'm also curious to know if the CF are at all different sounding than MF simply because of the carbon - or if the construction method is the only reason for the difference in CC and has nothing to do with the carbon.

[Ronsonic]

ATMO, more important than the material is the voltage rating and physical size.

[exclamationmark]

Time for a necrobump

As I understand it resistor noise is primarily comprised of johnson (thermal) noise, and current noise (also known as excess/contact noise). Current noise only occurs when there is a DC voltage across a resistor. It then makes sense to make sure the plate/cathode resistors are metal film (they have lower current noise than carbon comp/film types) , but what about the rest? Does it simply not matter what type of resistor you use for the grid stopper and grid leak, since it has no DC bias, and hence thermal noise will dominate (thermal noise is the same regardless of resistor composition). It's actually a practical question, since I got a hold of some nice SMD thick film resistors - only problem is that their current noise levels are something like 10x higher than 1/2 watt carbon films, and probably 100x higher than metal films.

[R.G.]

Current noise only occurs when there is a DC voltage across a resistor.

I think I'd say that current noise only occurs when there is a *voltage* across a resistor, since current flows whenever there is a voltage, DC or not. Another way to look at it is that since noise has a wide frequency distribution, every signal is "DC" to some noise.

[nickb]

Well, we know there is a scientific basis to the carbon comp resistor mojo thing: they generate sweet-sounding 2nd harmonic distortion when hit with high AC voltages.
http://www.geofex.com/article_folder...carboncomp.htm

Working the numbers through for a typical 100K resistor with 100V across it and a 75v swing gives a THD of 0.09%. Like RG said so well "Urban legend is tough to kill.." and "The effect is real, though slight. ".

Indeed, very, very, slight.

[Gingertube]

yes, yes CC mojo, even order harmonics, last long enough to drift hugely...

Yes, just repaired a VOX AC30 (circa. 1965) Tremlo oscillator 270K anode load read 690K and the anode volts was down to 52V and the oscillator just would'nt run. Metal film 270K fitted, anode volts back to 150V and the tremelo/vibrato lurched back to life. Quick check of all the anode voltages, Phase splitter +52V and +210V, 100K anode loads reading 180K and 390K, replaced both anode loads with 1W Carbon Film 100K, anodes now read within a volt of each other and the amp suddenly sounds much more powerful (well it should too with both sides of the phase splitter now working). Note, I had both carbon film and metal film 100Ks in my parts bin. I put in the Carbon Films coz I knew the owner would be happier regardless of any "facts".n

220K anode loads on the NORMAL and BRIGHT input tubes read 260K and 300K but I've left them for now.

The Solid State HIFi DIY crowd know a bit about this (been there/done that) - add a few Riken Ohm resistors (designed to exhibit that voltage dependency) around the place to introduce a little 2nd Harmonic distortion - great for MASKING all sorts of warts and other problems. Of-course I eventually "went off them" when I found that these circuits always sound better when the Riken Ohms are taken out again and the real problems they masked are addressed instead. Riken Ohms are also a favourite for "Upgrade" kits for CD players etc.

I also do some Vintage Radio restoration work and it is quite strange that I don't seem to see the same degree of problem with huge carbon comp resitance value drift that I see in Guitar Amps, higher operating currents in guitar ampds maybe?

Cheers,
Ian

[shocki]

I have looked at many gushots from boutique amp builders. Most of the seem to use carbon film as this seems to be a good compromise. At few parts they seem to use metalloxid resistors.
Here is an Dr.Z. example.

DR.Z PRESCRIPTION JUNIOR AMP GUTSHOT... | Flickr - Photo Sharing!

Here you can see a gutshot where also carbon comp is used.
http://www.germinoamps.com/masonette25.htm#!prettyPhoto

[exclamationmark]

I think I'd say that current noise only occurs when there is a *voltage* across a resistor, since current flows whenever there is a voltage, DC or not. Another way to look at it is that since noise has a wide frequency distribution, every signal is "DC" to some noise.

You're probably right - Although I've heard several sources claim that noise is only proportional to DC current, but never explain why. Apparently in semiconductors, the spectra of noise produced is based upon the frequency of the of the voltage applied, though I dunno if it applies to resistors (there seems to be remarkably little information on this).

I've also wondered about noise from potentiometers... They typically use carbon tracks, or some sort of cermet material which I imagine to be incredibly noisy compared to your standard resistors. A 1Meg (log audio taper) gain pot up halfway would have something like a 'head' resistance of 900k, and a 'tail' resistance of 100k. IIRC shunt resistances (like the grid leak or the 'tail' section of a potentiometer) don't follow the standard conventions of noise addition (it's far less noisy), but the upper section of the pot is certainly in series with the grid of the tube. Given that it's still the second stage of the amplifier, wouldn't this be a large source of noise?

Anyway, I've just had a revelation about my original question...Hiss is predominantly noticeable when no input is present, so I guess it doesn't matter what resistors I use for the grid leak/grid stop if that's all I care about, since thermal noise is constant despite resistor composition (hey, I just answered my own question!).

[soundmasterg]

Read this if you haven't already:

Resistor Types--Does It Matter?

Greg

[Bluefinger]

Leaving all theory aside ... I have done a complete carbon comp/metal film comparison in three amps meanwhile. I can't hear a difference for the life of me. I'm always using 2watters throughout the amp simply because the legs are more sturdy than on the .6 ot 1 watt resisors. Metal films for me ... I never had one gone bad