• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Where low noise resistors matter

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have found a lot of discussion of what resistors have the lowest noise (apparently wire wound and metal film) but not much about where it matters most.

It would seem that the lower the signal voltage the more important it becomes so early preamp stages more important than power stages. So grid pull down and grid stoppers would seem to be more important than plates but how about cathodes? Would screens be more important than plates?

Any general light to shed on the question?

mike
 
Noise in a resistor is caused by current flow. Plate resistors have more current flow commensurate with more signal then grids in most cases. Therefore, plate resistors are the most critical to producing noise and traditionally given the most attention. Cathode bias resistors have current flow but with less signal level, so I'd put them next with grid resistors after that. For cathode followers, obviously it's the cathode resistor. Screen resistors, while not directly in the signal path, could produce noise too I suppose if you had a really bad one.

Victor
 
mashaffer said:

It would seem that the lower the signal voltage the more important it becomes so early preamp stages more important than power stages. So grid pull down and grid stoppers would seem to be more important than plates but how about cathodes? Would screens be more important than plates?

The grid stoppers rarely carry much current in normal audio applications, so it is better to reduce inductance rather than noise. Arguably best to use the noisy carbon composition resistors there! For grid leak resistors, I typically use standard metal film, as the inductance is benign.

For noise, you want to look for resistors that carry current and dissipate power, since they will be the warmest and thus the noisiest. So, focus on plate resistors, and any other resistors that are dissipating power / carrying current. For example, plate loads on driver stages, cathode resistors in phase inverters, screen resistors that carry current, etc. Better to overrate any resistor that will dissipate power if it can contribute to the noise signal. An exception: never overrate cathode Rs in an output stage (they should be metal film and preferably fusible!) -- you want them to pop before the output transformer :hot: And pay attention to voltage ratings. Carbon comps are great at high voltage AC and pulses, but some film resistors will go "pop" under those conditions.

If I were to have a simple formula for low-noise and overkill:
Power supply dropping R > 1W = Conventional wire-wound, appropriately sized
Power supply dropping R < 1W = Metal film 3W
Plate or cathode R dissipating > 1W = Wire-wound Non-inductive 5W+
Plate or cathode R dissipating < 1W = Metal film 3W
Feedback R with some dissipation <1W = Premium 3W metal film or bulk foil
Feedback R with minimal dissipation <<1/2 W = Premium 1/2W metal film or bulk foil
Grid leak with minimal dissipation <<1/2 W = Premium 1/2W metal film
Grid stop with no grid current = 1W Carbon Comp
Screen stopper with minimal screen current = Appropriately-sized carbon comp
Cathode resistor for a power tube = size so it fails at excess current!

I'm not keen on $$$ resistors -- military spec 3W and 1/2W are exceptionally useful, and are available with numeric labels vs. color codes. They are non-magnetic and consistently affordable. GS3 (3W) and RN60 parts (1/2W), for example, along with new production carbon comp resistors, are great options. Same with KOA Speer carbon films, which were recently discovered to be the Kiwame R's true identity ;) Anyways... just my guide to Rs, I'm sure others will have much to add/disagree with. :smash:

(just a note, I also like to oversize Rs also for the sake of the fatter leads -- a 3W metal film is much easier to work with in a point-to-point circuit, and is no bigger than a 1W carbon comp. I drove myself nuts trying to install a 1/4W film feedback R the other day when I couldn't find a 1/2W).
 
Good starting point, Jon. It is most helpful to understand WHY we choose the technologies we do.

I recall the time when I finally understood proper capacitor application (when to use lytic, tantalum, film). Although still a debated subject, the basic understanding is important.

I'm not sure I like the idea of using resistors as fuses, though. Wouldn't a flameproof resistor, at a minimum, be recommended in this application? Better yet, design the power fuse rating to be sensitive to a failed tube?
 
But keep a watch not only on power ratings but also on voltage ratings. Even if a resistor can carry the power (W), the voltage may be to high for that kind of resistor.
A 1/4W or 1/2W resistor i physically small and applying a high voltage across it will case a very strong electric field that progressively will ruin the resistor. This progress is slow and will propably begin with increased noise levels and finally in total collaps.
I try my hardest to use 1W all over, except where it calls higher wattages.
 
zigzagflux said:
I'm not sure I like the idea of using resistors as fuses, though. Wouldn't a flameproof resistor, at a minimum, be recommended in this application? Better yet, design the power fuse rating to be sensitive to a failed tube?


You can buy Vishay Fusible metal film resistors, which are flame proof and fail promptly at their power rating. These are good for cathode Rs in fixed bias amps. The reason I find them appealing is that a fuse adds complexity, and you need an R regardless for convenient measurement of cathode current. Chances are, the R will never get the chance to blow (hopefully). ;) I do wonder if they smell less than standard film resistors, as I've never popped one... sounds like a useless experiment waiting to happen :smash:
 
Source 100K wire wound

Ok. So I can easily get metal film and metal oxide in 100K 0.5-3 watts but I am having trouble finding the same in wire wound. Mouser does have some really big 11 to 50 watt ones. They also have some 50K (I think they were 3 watt) of which I could use two in series.

The reason that I am looking for the lowest noise resistors that are practical is that the application has a gain of about 1200 (guitar amp) and I really want to be sure the first stages are quiet.

Would WW have a significant advantage over 1/2W MF or 1W MO?

mike
 
Re: Source 100K wire wound

mashaffer said:
Ok. So I can easily get metal film and metal oxide in 100K 0.5-3 watts but I am having trouble finding the same in wire wound. Mouser does have some really big 11 to 50 watt ones. They also have some 50K (I think they were 3 watt) of which I could use two in series.

The reason that I am looking for the lowest noise resistors that are practical is that the application has a gain of about 1200 (guitar amp) and I really want to be sure the first stages are quiet.

Would WW have a significant advantage over 1/2W MF or 1W MO?

mike

A good choice for audio-grade wire wounds are Mills non-inductive. But, I wouldn't bother unless you need a premium-grade plate load in a high voltage / high current driver stage.

Metal film is absolutely acceptable as long as you aren't pushing close to the max power rating. Wire-wound is great for high power plate loads on a big driver stage, but if this is on a first stage in a guitar amp, the current is probably fairly low. So, stick with film unless you need to dissipate more than ~2W. If you are in the 2W territory, try putting two 3W film resistors in parallel for a huge cost savings over wire-wound -- this is a fine solution. Frankly, I find "noisy" carbon comp resistors to be sufficiently-quiet for high-gain stages (phono), just because the current is so low that they can stay cool and behaved.

BTW, I would skip the metal oxide. They are slightly noisier than film. Great 3W resistors are the IRC GS3 metal glaze Rs available for $0.45/each from Mouser or Handmade Electronics. I cannot personally read color codes (I'm a lousy EE), so I always buy military-style parts (like these) that have numeric labels :xeye: Standard color coded film Rs are really a-okay if you can read and sort them.
 
mashaffer said:
Ah, thanks Jon. Yes indeed I did not consider that the current would be low (about 1mA) so 140ish volts times .001A = .14W so 1/2W MF should be fine huh? The MO I have are 1W so potentially noisier but farther from rated power capacity. Might be a wash???

mike


If it's a plate resistor, which will contribute to the signal, I'd go with the metal film. The metal oxide Rs are probably more useful as screen stoppers (non-inductive) or power supply dropping resistors (small but decent power handling).

.14W on a 1/2W R sounds a-okay to me! Running 25% of rating is pretty "safe" IMHO.
 
Ex-Moderator
Joined 2003
Vishay do themselves a disservice...

Gary P said:
If you want some noise numbers for the various resistor types go to Vishay's discrete metal foil web page.

Hmmm. I'd expect a manufacturer's document to be rather better than that. They can't spell Boltzmann's name correctly ("Boltsman") and they provide no evidence, merely assertions and wild comments like, "The major objection to wirewounds, however, is the inductance that chops the peaks and fails to replicate the higher frequencies of the second and third harmonics." Absolute drivel.
 
All stages of an amplifier are important so why compromise , if you know its going to sound better and give you trouble free service than don't skimp particularly if it is going to be the final product . It does not matter what a component costs if it takes longer to build because money is tight (as in my case) building an amplifier that is second rate is just a waste of time . Build the best you can your time and effort is worth more.
nkg
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.