When to increase wattage of resistors

[Lydian]

Hey guys,

I've been having a blast getting to know amps. It's a long journey ahead. I've started my first repair on my 68' Pro Reverb, it's the AA165 scheme. I bought this amp about 10 years ago. Since I've opened up the chassis I have tested the resistors. A lot of the original resistors are out of spec. The guy (local legend) I bought the amp from changed the electrolytics before I bought this amp. I went and bought the needed resistors and am currently starting to replace them where needed. What I noticed is the few resistors that were replaced from 10 years ago are the same resistance but are 1 watt/ not 1/2watt as suggested from AA-165 schematic.

Should I continue this process and increase that watt size of resistors as I change them for extra protection? Should I just go by scheme and use the 1/2 w resistors?

Thanks for any advise.
-Dalton

[Leo_Gnardo]

Won't make any difference 1 or half watt when replacing half watters in your Fender. Whichever you have on hand or more convenient to buy.

[Lydian]

Thanks Leo.

[Mick Bailey]

For a given rating many components have shrunk in size over the years. You don't say what types of resistors you bought, but modern carbon comps are a lot smaller than they used to be. To keep the same physical size sometimes you have to increase the rating. Just a visual thing, but sometimes it can mean an extra bit of lead length.

How much out of spec are the original resistors? Bear in mind the tolerance back then was usually 10%.

[Lydian]

I have some that should be 1500 reading 3000, like my 6800 mid resistor on the bass pot reading in the 8000's

Replacing with tolerance 5% or less. I would think accuracy would be desirable. Yes? Or is the magic in the 10% variations.

[tedmich]

These designs were (and are) very variation tolerant. Any "magic' would likely be the sum total of a whole lot of variation and hard to reproduce/maintain. I like stability, but CC resistors aren't great for that to begin with

[eschertron]

I'd have to crack open my copy of Merlin's preamp book to confirm, but IIRC resistor noise goes down as the size of the resistor goes up. If the resistors replaced are in a critical location, that might have been part of the previous owner's rationale.

[Mick Bailey]

With some resistors you also have to go up on wattage to get the voltage rating. Depends on reisstor value and construction technology.

[Tom Phillips]

All the information in the above posts is true but the sound, noise or reliability difference with 1/2W resistors changed to 1W may not be noticeable in a guitar amp. In the Silverface years Fender did use 1W CC plate resistors as stock parts in many builds. That was not documented by an update of the schematics. Attached is a gut photo of a stock 1969 Twin Reverb with the 1W resistors. If the 1W resistors in your amp are in the same locations then they may be stock parts. If you tried to change all the resistors to 1W there will be some eyelet positions where the leads will not physically fit. Overall, I don't think it is worthwhile making the change.
Cheers,
Tom

[R.G.]

Good place for another episode of "Pondering Resistors".

Resistors are neither the pure, innocent devices that simple theory says, nor the creatures of myth so beloved by the sellers on the internet.

(1) Resistors are thermal devices. Resistors impede (!) electricity by converting it into heat as a price of passage. Any device that converts something inside it to heat has an internal temperature that is **unlimited** until the temperature rises to the point that outgoing heat by conduction, convection and radiation balances the internally generated heat. A grain of wheat bulb may have about a watt of electricity shoved into it, but the tungsten filament is so insulated by the construction and it's tiny surface are that the tungsten glows yellow-white, about 2550C or 4600F, the temperature having risen until that 1W can be radiated away as light.

A 1W resistor is not even visible by its radiation, and 1W in a hair dryer won't even be detectable by baseline humans. In fact, resistors are rated by whether they will withstand the rated power without self-destructing over a reasonable lifetime - defined in the component biz as between 5000 and 20k power-on hours at rated dissipation. Modern smaller high power resistors get that way by being made from materials that can stand the heat. The surface temp of some power resistors is 200C at full power. That's branding temperature.

So putting in higher-wattage-rating resistors may or may not keep the resistor temperature down.

(2) Temperature makes noise. Every normal conductor creates thermal noise proportional to its absolute temperature. So keeping things cool makes them lower thermal-noise devices, which is why some microwave stuff is run while submerged in liquid nitrogen. Closer to home, using a higher-rated device at lower than rated power may keep it cooler, which will cut the thermal noise some. But remember that thermal noise is based on absolute temps, which start at -273C, so keeping it 10C cooler on top of a 300C base may not help much.

(3) Temperature makes things wear out. Resistors run at lower than rated temperature may last a very long time without drifting or dying.

(4) Some materials are nearly ideal resistors, some are sloppy. Metals are nearly ideal, so wirewound resistors are about as quiet and linear as you can get resistors. Bummer about that self inductance, but the inductance may or may not matter in the application. Metal doesn't drift much either. Carbon is not nearly as linear or stable. Carbon film is worse about linearity, drift, and noise than metal film or wire. Carbon comp is downright sloppy. CC is a nonlinear resistance, drifts with heating badly (for a resistor) and has "excess noise", another second helping of hiss that you get because the carbon is carbon and the granules in the composition goop are erratically connected.

(5) Resistors can be nonlinear as resistors. There is small print on resistor datasheets. It talks about the temperature and voltage coefficients of resistance as well as other stuff. Nearly all conductors have positive temperature coefficients of resistance, meaning that as they get hot, their resistance goes up. So a 1K resistor at 25C may be 1.03K at 200C. That's not so bad unless you're doing precision measurements. But the voltage coefficient is a bad deal. Carbon comp resistors have a relatively high voltage coefficient of resistance, which means that the resistance itself varies just because it has a voltage across it. Near as I can tell, this is the entire grain of truth hidden under the whole "carbon comp is the only resistor for good sound" myth on the internet. It is true that a carbon comp resistor with a largish DC voltage and a large AC voltage superimposed on it has a little (under 1% mostly) bit of even-order distortion, and the human ear loves this like kids love ice cream. The only conditions this is true for are for tube amps and then only when the signal is swinging well over 50V peak. I did an article on this at my web site, geofex.com. Using CC for the 1M grid pulldown resistors on guitar amp inputs contributes excess noise but no even-order-distortion goodie to the mix.

With that as background, swapping 1W for 1/2W might make good sense if the resistor dissipates any sensible fraction of its rating in the circuit. In practice, only plate resistors need to apply. Well, OK, power tube screen resistors, maybe. And be aware that changing the old drifted value for the new, "correct" value may change the sound of the amp. Maybe what you want is a new "drifted" value. Or something in between. The water gets deep about here.

For lowest noise, chuck out all the old supposedly magic carbon comp resistors. For tube sound mojo, leave them in the plates of the PI.

And about that wirewound resistor inductance - mostly don't sweat it. The inductance is very rarely enough to cause any significant change at audio. But if you're a self-appointed expert, you won't even make the calculations to find this out and will simply spout that wirewound is evil because of inductance.

[Justin Thomas]

About that wirewound resistor inductance - mostly don't sweat it. The inductance is very rarely enough to cause any significant change at audio. But if you're a self-appointed expert, you won't even make the calculations to find this out and will simply spout that wirewound is evil because of inductance.

Works decent replacing the dead chokes in the active EQ of a Super Twin... sorta. Hey, I'm not putting too much effort into a $100 amp! I just measured the other 4, figured out that the resistance kinda followed a trend of R:inductance, and slapped a 120R wirewound in the Bass spot. It passed...

Justin

[Dave H]

For tube sound mojo, leave them in the plates of the PI.

But will you be able to hear the mojo? It will be clipped by the power tube grids before the signal is swinging well above 50V peak (well before if you have EL84s).

[R.G.]

But will you be able to hear the mojo? It will be clipped by the power tube grids before the signal is swinging well above 50V peak (well before if you have EL84s).

Good question.

Resistor distortion is a small percent of the voltage across the resistor. The bigger the voltage, the bigger (or smaller, depends on the direction of the nonlinearity) the resistance is, and the bigger(smaller) the signal is for a quasi-constant-current feed like from a tube plate or transistor collector/drain. The distortion just goes down with smaller signals.

So yes, it gets covered up with power tube grid clipping if you drive the tube there. But just below grid clipping you get the most soft distortion from plate resistor nonlinearity. So it's a matter of degree. Depends on the resistor, how big the voltage coefficient is, and the size of the DC offset on the resistor. For biggest effect, you want the biggest swing possible and a DC offset about equal to Vpeak for the signal.

For a 40V peak signal, you get 80V pk-pk on the PI plates, and if you run them at maybe 120V on the plates, you get a swing from 80V to 160V across the plate resistors. That's enough to make the signal bigger on one output tube grid and smaller on the other grid.

Whether the effect is big is a matter of degrees. With small signals, triodes are the most linear amplifying device known. With big signals the triodes themselves start having nonlinear swings on the plate, so that gets lumped in too. A fair test would include listening to a triode plate signal with a switchable plate resistor of, say, CC and MF, matched in value and muted while the switch moved. It would be an interesting test. Sorting out which nonlinearities are important to the "tube sound" is a lot of work. A lot of the folks who claim to have discovered "the" secret discover one and quit. Then the next guy discovers something else.

Humans easily get confused with multivariate problems.

[tedmich]

Wonderful clear and concise writing as usual R.G.!

as a counterpoint...while bearing in mind that the typical Audiophile wouldn't touch a tantalum capacitor with a 10 foot pole, there are those who advocate...

The AUDIO NOTE™ Tantalum Resistor
by Peter Qvortrup
What makes a better sounding audio resistor?

Low or no magnetic properties, is a good start, and the A-N Tants are only slightly magnetic, as they use nickel-plated end caps. Tantalum film resistors have been generally regarded as the best sounding resistor material for many many years. The famous (and now sadly defunct) Shinkoh tantalum film resistor were the first to recognize this, and I remember using these in the late 1970’s.

The Audio Note™ Tantalum film resistor is made from a pure tantalum oxide. However, to allow their use in high voltage applications, we have chosen to use nickel-plated end caps, wherefore both the 1 and the 2 watt resistor’s max overload voltage is 1,500 volts, with a working voltage of 1,200 volts.The price for this is that they are slight magnetic, as I said at the beginning.

The A-N Tants are made to a very exacting specification, being 1% toleranced and able to withstand temperatures of over 150 degrees (C) without significant change in resistance. As well, we have specified a "particular" etching process which increases the surface area for a given resistance, as we have found this to sound better.

There are other processes involved, however, as with so many other aspects of specialist audio parts, we do not disclose these to prevent them from being copied!

from www.partsconnexion.com
numbers (other than price) are insufficient to describe the mojo of these...

[R.G.]

Wow. I should be immune to outrage at the excesses of the hifi tweeko world, but some of them sneak through anyway. I'd dearly love to do a double blind test on some of the hifi tweako **reviewers** to see if they could do better than raw guessing at hearing the substitution of one and only one resistor from metal film to tanalum oxide. I'm fairly sure that this is something they would never agree to, because all the examples of blind testing of "golden ears" have gone poorly for them.

I keep on deciding that my sig is appropriate.