Pre-Amp Plate Load Resistors

[ajeffcote]

The circuit is an AB763. Running two 6L6GCs. Can I use 1/2 watt plate load resistors? And about the reverb driver cathode resistor, is a two watt OK?

[Tonewood]

The circuit is an AB763. Running two 6L6GCs. Can I use 1/2 watt plate load resistors? And about the reverb driver cathode resistor, is a two watt OK?

My super reverb uses 1/2 watt but many people say larger is quieter.

[Chuck H]

My super reverb uses 1/2 watt but many people say larger is quieter.

Resistors generate thermal noise. I have assumed this to mean that the hotter a resistor gets, the more thermal noise it generates. But now I'm wondering if the thermal noise isn't just an analogous byproduct of work performed?!? My simple mind is telling me that since a higher wattage resistor doesn't get as hot the thermal noise will be lower. But my logical mind says that the same work generates the same noise (and the same heat) and the actual surface temperature is more a matter of surface area and how a given resistors inefficiency is distributoed over it's mass. This based on the ideal that wattage rating doesn't alter the actual efficiency of a resistor. I've heard this reasoning before. That is, a higher wattage resistor generates less thermal noise. But I'd be interested to hear from some of the heavy hitters here. R.G.? Steve?

[Tonewood]

I have replaced several plate resistors in early preamp stages over time with 1W metal resistors & had a reduction in hiss. Not sure if it is the metal, the size or just needed a new resistor. Here is some info:

Resistor Types--Does It Matter?

[Chuck H]

Right. I'm familiar with that artical (and have even attached it in other threads). But a repeated theme is that resistance is analogous to thermal noise. Yet the use of higher wattage resitors is suggested, but not substantiated!?! As I said, it makes some sense to me that a cooler resistor should have less thermal noise. But... Does a bigger tire have less friction braking a given mobile weight, or just the same friction though dispersed over a greater mass? The answer to the latter is obvious to me. The former, not so much. Not being schooled in such things I refuse to assume and instead digress. As mentioned, in the artical the use of higher wattage resistors is suggested, but not substantiated. Since the sheer resistance being analogous to thermal noise is a repeated theme. Does the actual temperature matter in this instance? Or is it pedantic to the equation? The attached article doesn't seem to address this. And I still don't know.

[madkatb]

I googled resistor thermal noise and came up with this:
Thermal Noise Calculator
Since thermal noise is affected by temperature it makes sense that higher wattage resistors would have less noise. Their mass dissipates heat so the resistor would never get to as high a temperature, resulting in less noise.
It also indicates that this is unrelated to resistor composition. Your linked article indicates that shot noise, not thermal noise, is affected by resistor composition.

[diagrammatiks]

it's a factor of resistor construction.

a 3watt resistor dissipating 90 percent of its max will be quieter then a 1/8th watt resistor dissipating 90 percent of its max.

[R.G.]

Resistive materials have thermal noise, excess noise and flicker/shot noise. Thermal noise is unavoidable, and a function of the material, the resistance, and the absolute temperature. For a given resistance, the thermal noise can be calculated easily enough from those factors. Flicker and shot noise tend to be influenced heavily by surface conduction conditions, and so are heavily a function of construction details. Construction and passivation (or lack thereof) is one reason early germanium transistors are noisy: the surfaces are contaminated and conducting.

The power rating gets into it because room temperature is nominally 25C, which is very close to 298K. A resistor usually is rated for a power which would increase its surface temp to 100C to 200C, depending on how stable the materials are. So a resistor being run at 100% of its dissipation rating heats up to 398 to 498K, which is a substantial increase in temperature, so increases the thermal noise. But it's the temperature of the resistive material, rather than the resistor rating that makes the thermal noise smaller. If the resistor's dissipation is less than 10% of its rating, going to ten times the power rating won't help you. If the resistor has substantial power in it, increasing the rating will lower temperature and lower noise.

Excess noise is rated in uV of noise per volt of applied voltage per decade of frequency. High voltages across a resistor make its excess noise worse. Carbon comp may have 40db more excess noise than a low noise type such as metal film or bulk metal foil. Wirewound are usually very low noise.

The first step in getting low noise in any amplifier is to worry about the first stage. Any noise in the first stage is amplified by all other stages. Especially worry about the input noise on the first stage, because it's being amplified by the gain of even the first stage. In tube amps, the input resistors really ought to be metal film or bulk metal foil if you can get them. Carbon comp's "tone mojo" is most likely a result of the voltage coefficient of resistance, and the signals are too small to give you mojo in the first stage. While the low voltages may also help with low excess noise because the applied voltage is small, why take the chance?

For collector resistors, I'd go with metal film, wirewound, or bulk metal foil if I could get the values right up to the high signal stages. The noise importance is less as you go through the amp stages, but since carbon comp won't help with tone mojo until quite high signal voltages, why take the chance?

I'd never knowingly use a resistor where it was dissipating more than 50% of its rating on a continuous basis, and for low signal input stage resistors, I'd cut back to about 10% to keep them cool. This matters less as the signal level rises through the amp.

[Alan0354]

One thing about carbon resistor, it has flicker noise that is so called 1/f noise. The lower the frequency, the more prominent it is. It over take the thermal at lower frequency. That's the reason people never use carbon resistor in low noise designs unless the working range is in the MHz or higher where flicker noise die down.

The formula of thermal noise is sqrt[4ktRΔf] where Δf is the bandwidth and temperature t is in Kelvin where 25 deg C equal to 298 deg K. So if you talk about difference of 20 deg between the 1/2 W and 1W ( I don't think is that big a difference), in Kelvin it is say difference of 300 to 320. If you take sqrt[320/300] = 1.0328. The thermal noise difference is only 3.28% for difference of 20 deg C. I doubted you can see the difference.

Shot noise is sqrt[2qIΔf] should be the same as the current is the same in both case. Also the resistance are supposed to be the same for comparison. So shot noise is not an issue here.

Noise (electronics) - Wikipedia, the free encyclopedia

Shot noise - Wikipedia, the free encyclopedia

Flicker noise is complicated, people just avoid using Carbon resistor or hand pick if needed!!! I don't know much about flicker noise, matter of fact, I don't know much about carbon comp resistors as I just avoid if possible. Only time I use carbon comp at work was when I need resistors that can take large surge current that easily blow the MF resistors, that's when carbon comp resistors shine. FET and MOSFET have the same problem, the weak point is the flicker noise in low ( lots of times not so low) frequency.

[Gingertube]

Useful stuff above BUT the main reason for using a larger wattage resistor is usually VOLTAGE rating.
If I could I would build everything using 0.6W Metal Film resistors, but I can't, why? Because they are rated for 300 to 350V maximum and an "over voltaged" metal film has a habit of going open circuit on you.
Carbon resistors have fantastic fixed voltage and pulse voltage withstand but are noisey.
So my default resistor for anode loads is a 2W Metal Film and in higher voltage circuits I often use 2 resistors of half the required value in series.
Cheers,
Ian

[R.G.]

If I could I would build everything using 0.6W Metal Film resistors, but I can't, why? Because they are rated for 300 to 350V maximum and an "over voltaged" metal film has a habit of going open circuit on you.

An interesting way to look at this is that (for instance) metal film resistors have a voltage rating of 300 to 350V for a single resistor.

Two equal-value resistors in series automatically equalize the voltage across them to within the resistor tolerance, and therefore the series combination has a voltage rating twice the rating for one of them. They also share the power equally, so they have a total power rating equal to twice the power of either one of them. So if you like metal film otherwise, use two of them in series of half the resistance. The combination is then rated for 600V. Three of them gets you 900V. I don't want to play inside amps with more than 900V B+.

The noise performance is the same as regards thermal noise based on resistance, and probably unchanged based on self-heating.

[defaced]

Useful stuff above BUT the main reason for using a larger wattage resistor is usually VOLTAGE rating.
If I could I would build everything using 0.6W Metal Film resistors, but I can't, why? Because they are rated for 300 to 350V maximum and an "over voltaged" metal film has a habit of going open circuit on you.
Carbon resistors have fantastic fixed voltage and pulse voltage withstand but are noisey.
So my default resistor for anode loads is a 2W Metal Film and in higher voltage circuits I often use 2 resistors of half the required value in series.
Cheers,
Ian

You may want to look at the Vishay CCF60 or CMF60 resistors. 500v rating, 1/2 to 1W power rating (depending on ambient temperature).

[diagrammatiks]

You may want to look at the Vishay CCF60 or CMF60 resistors. 500v rating, 1/2 to 1W power rating (depending on ambient temperature).

cmf70s are even better.

I'd hate to have to build everything with .6watt resistors.

tiny little things. leave those to robots.

[exclamationmark]

Most 2 watt resistors are literally about 1mm larger than cmf60's. Though probably not as good in noise, stability and a myriad of other things, though. I think they're perhaps a little too conservative with their ratings, as generic 2 watt resistors I use de-rate in the same manner, despite being almost the same size.

[diagrammatiks]

the cmf/ccf/rn60 is only a half watt commercial rated resistor.

the 2watt resistor is the rn70.