diyAudio (
-   Instruments and Amps (
-   -   musing on guitar amp power supply filters (

cyclecamper 11th December 2012 07:51 PM

musing on guitar amp power supply filters
More musings from a tubey-newby still learning about tube guitar amps as opposed to transistor amps. Looking at some schematics and doing some reading and making some inferences and guesses, I know just enough to be dangerous and came up with some ideas (about altering filter cap values and using chokes) which I want to confirm, and questions to pose.

Instead of multiple seperate supplies at various voltages, or regulating the various voltages required, tube guitar amps tend to make a long set of resistor and/or choke and cap filters, with various voltages tapped off as the voltage drops farther down the filter line.

Of course, these voltages vary a lot with changing signal and related load.

It may be important that the relative votage taps retain similar proportions or at least ordinal hierarchy as the voltage sags, and the voltage characteristics over time under sag should be considered in the design unless the supply is radically overbuilt to never sag significantly.

It's good for each supply section to be somewhat isolated from the others. You don't want the output load to impose signal on the supply to the preamp, worst case causing oscillation or weird feedback characteristics ruining frequency response. Filtering to an even lower sub-audio frequency reduces the effect of compression being compounded by reduced preamp voltage and reduced preamp gain whenever the power section is working hard, which effect might be harder to avoid, or even enjoyed (or not) as a component of "sustain" by us crazy eccentric musician types.

On the other hand, I can't always just arbitrarily greatly increase the value of some filter caps way down the filter line. I can envision situations where earlier higher voltage taps could have much more load and radically sag momentarily until those first voltage taps reach lower B+ than the successive voltage taps which require very little current and continue to be supplied by their erroneously oversized filter caps. Might not be good, especially if it results in any grid becoming more + than the relevant plate.

So some right-sizing and proportionality of the stored energy (caps) in the various stages brings several things to my mind.

If I add caps to the power supply, I should add them to the first raw tap before I add any to subsequent taps down the filter line. That would be an attempt to keep the first taps from sagging below subsequent later taps. In an earlier life I used to do this to my rack-mount tube power amps to get a punchier sound with more dynamics, just increasing the size of the first filter stage caps.

And this might give some insight into why chokes may be preferred between filter stages. A choke makes the filter a 2nd order instead of the shallow first order of just caps and resistors. So it may be a steeper filter, and work well at keeping power-section signal from being imposed on preamp supply, yet not neessarily have a much larger total reserve of stored current or have a significantly different frequency knee to the filter. So the various voltage taps are well isolated from each other in the audio range and below that to handle peaks and even crescendos, but once it does begin to sag significantly the rates of the various stages sagging is somewhat controlled. The sharper filter behaves better before it drops, then drops more steeply. They engineer guitar amps to remain safe and controlled well into abuse, and the characteristics of how they handle abuse must be considered. And again, the choke and later caps must be right-sized to the first caps if I want to be sure to avoid possilble problems.

Just looking at schematics without watching voltages during use yet, I keep expecting the voltages in later stages in the filter network to slowly rise up and eventuallly get really high in very low-signal ultra-low-load situations. I keep expecting some resistors across those later taps, wasting a little power just to reduce the high voltage at no load and reduce the voltage swings. I guess I'm forgetting that the bias keeps the tubes conducting some, and this voltage rise isn't really significant. I keep forgetting the preamp tubes are biased class a and always on...

sreten 11th December 2012 08:43 PM


As they say, a little knowledge is the most dangerous. You could double
all the capacitor values all along a RC voltage dropping chain, but after
that be probably wasting your time, but how do you know that the
original designer has not already taken a belt and braces approach ?

The load on such a supply for class A preamp stages is essentially
constant currents, that is the only way it can work well, and the
situations you are suggesting simply don't happen in reality, the
load on the cascaded networks is constant, with basically
constant voltage and constant ripple for each supply.

A little known trick is if you have enough cascaded stages the
the phase of the ripple in the final stage is out of phase with
the first stage. A high value resistor from the first stage
can inject current into the last stage to reduce its ripple.

Inductors are simply not needed, why they are not used,
except of course for the main power amplifier power rail.

rgds, sreten.

Keriwena 11th December 2012 10:27 PM

All I know is, many years ago, I "upgraded" the power supply in a friend's Blackface, and he hated the way it sounded. Too clean. :(

Enzo 11th December 2012 10:53 PM

And now you are left to wonder if the problem was the value of the new caps or just the newness of them.

Keriwena 12th December 2012 12:12 AM

Oh, no, it was many years ago. The Blackface was probably newer than the caps! ;)

I know, I should have been baking cakes and sewing dresses, but I liked the smell of solder. That might explain.... :warped:


Enzo 12th December 2012 03:47 AM

OMG, you're a...a...GIRL???

cyclecamper 12th December 2012 05:16 AM

I don't really get the out of phase thing, I'll have to think about that. Out of phase at 60 cycles, or some signal freuency, or the duration of a crescendo?

Marshall and Fender usually used chokes between the first caps and taps for the raw rails and the next filter stage, instead of a resistor, isolating the high-power tap from all the rest. That's the reason I was interested.

The more I think about it, the more I like the swap of the Peavey 100 power transformer into the Peavey 50. I don't think I'm likely to regret upgrading the power supply...

cyclecamper 12th December 2012 05:33 AM

...but in general I'm still thinking in solid-state hi-fi amp terms, where the power supply can't have too much capacitance. Gotta shift perspectives...

Keriwena 12th December 2012 07:39 AM


Originally Posted by cyclecamper (
The more I think about it, the more I like the swap of the Peavey 100 power transformer into the Peavey 50. I don't think I'm likely to regret upgrading the power supply...

A pedal steel player probably would have loved the Blackface I modded. But the guy that owned it wanted to sound like Eric Clapton, and I took it totally in the wrong direction. Now, I love the Classic 50, but partly because it has no headroom and a really soft crunch. You very well might like it with a bigger tranny.

Still, I suggest making the change in a way that allows you to reverse it. ;)

Keriwena 12th December 2012 07:42 AM


Originally Posted by Enzo (
OMG, you're a...a...GIRL???

Not any longer. Now I'm a grandmother. :angel:

All times are GMT. The time now is 06:35 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2