Big high uf!

[Stevem]

Over the weekend I was working on a 65 vibrochamp with many issues and the last one I had to deal with was too much bottom end.
When I got around to checking the first gain stage bypass cap that little 25 uf bugger had shot up to 100 uf!
That's the highest I have ever seen one drift up to in my 45 years of doing this stuff!

[R.G.]

Interesting!

It makes sense, though. Electrolytic decay involves the oxide layer gradually getting thinner as it dissolves or decays into the electrolyte. Thinner = higher capacitance, but lower voltage.

It's really unusual to see one go to 4X without some small place punching through, though.

[R.G.]

Hit "send" too fast.

It's a maxim in EE that one should never use electrolytics for any timing or filter/frequency response purposes unless there is simply no other way; it's for this reason.

[Stevem]

And to that end R G is not bypassing power supply filter caps with both a .22 and .022 cap a nice tonal improvement?

[R.G.]

Maybe. I'd like to see frequency spectra of a sine wave being played through an amp with just electro filters, then with succeedingly smaller parallel caps switched in as the only change.

[Leo_Gnardo]

And to that end R G is not bypassing power supply filter caps with both a .22 and .022 cap a nice tonal improvement?

That's a mod suggested for wide band hi fi amps, wouldn't have much relevance in a guitar amp. The idea is, supposedly electrolytic caps don't filter too well at high frequencies, so add parallel film caps to handle that task. I first saw it proposed in Audio Amateur in the late 70's, when electrolytics weren't so good after all, intended to handle AC line frequencies, plus double line frequency from rectification, and upper harmonics from that rectification up to a couple KHz. I've seen some hi fi manufacturers factory install such helper caps. And I've tried it a couple times in an attempt to stabilize guitar amps with ultrasonic oscillations - didn't help much.

[pompeiisneaks]

And to that end R G is not bypassing power supply filter caps with both a .22 and .022 cap a nice tonal improvement?

That's a mod suggested for wide band hi fi amps, wouldn't have much relevance in a guitar amp. The idea is, supposedly electrolytic caps don't filter too well at high frequencies, so add parallel film caps to handle that task. I first saw it proposed in Audio Amateur in the late 70's, when electrolytics weren't so good after all, intended to handle AC line frequencies, plus double line frequency from rectification, and upper harmonics from that rectification up to a couple KHz. I've seen some hi fi manufacturers factory install such helper caps. And I've tried it a couple times in an attempt to stabilize guitar amps with ultrasonic oscillations - didn't help much.

It's not just a problem in HiFi, any situation where high frequency oscillations or interference can occur, it's a common part of the circuit design, anywhere they need a high uF e-lytic they parallel it with a low level cap for that same reason. It damps or removes the high frequency that the e-lytic can't. Agreed that it's not needed for guitar amps.

~Phil

[Reeltarded]

If you are using a bunch of gain, I think hi-fi ideas apply more often than not.. things that happen over 6kHz are still affecting response. It's nice to have a black background.

[pompeiisneaks]

If you are using a bunch of gain, I think hi-fi ideas apply more often than not.. things that happen over 6kHz are still affecting response. It's nice to have a black background.

Oh wow, now that you say that, that actually makes sense, hi gain is very sensitive to almost any noise. Good point! Has anyone tried putting a .01 cap across the filter caps to see how that impacts noise?

~Phil

[Reeltarded]

If you have ever driven an amp and smiled until you got 10th 11th fret and higher on EBG strings and had the amp eat your sustain from phase beat you might want to try it

[didit]

phase beat

This ... seems the case that a few degrees at even 8KHz and the ears perk up or miss all the joy.

Best .. Ian

[R.G.]

I urge caution here. It's easy to talk yourself into Teflon pcb stock and gold plated circuit boards. Not saying there isn't any effect, but it would be good to measure what's happening and figure out what's real.

The whole mess about bypassing electrolytics came at least partially by people poring over old charts that plotted the impedance of a cap versus frequency. All caps have some Equivalent Series Inductance (ESL), and bigger caps have more of it. The impedance of the capacitance part of the impedance declines linearly with frequency. The impedance of the inductive part increases linearly with frequency, and there is some frequency where the two are equal. After declining for a few decades of frequency, every capacitor's impedance bottoms out and starts increasing again. The only way to keep bypassing all the way through the MHz and GHz is to put smaller ESL caps in parallel.

Smaller ESL caps are also smaller >capacitance< as well. A 47uF cap has 470 times the capacitance of a 0.1uF cap, and at any given frequency up to the 47uF cap's inevitable run in with its ESL, it will have 1/470th of the impedance of the 0.1uF. The 0.1 then decreases from there as frequency continues to rise. But it will never be a lower impedance than the 47uF would have been, and it will meet its own ESL Waterloo at some higher frequency too.

RF guys are very sensitive to this reasoning. But they also only have to get low impedances at a few frequencies, not a 1000 to 1 range in most cases.

The question then becomes - how low a capacitive impedance for power filtering is low enough? Reasonably, we want these capacitive reactances to help filter out noise, hiss, grunt, grunge, and grack. Filtering depends on the balance of a series element and a shunt element, with the limited case of power filters making the shunt element always being a capacitor and the series element an inductor or resistor.

In most guitar amps you can find 50V of ripple on B+. That's shoveled wholesale into the center cap of the OT, and we sure hope it gets mostly cancelled. The next step in most B+ chains is an inductor leading to the screens. That inductor has a high inductance, Henries. The ESL of an electro, even a big one, is going to be a few orders of magnitude less, and so the reduction of ALL frequencies by that inductor and the followup capacitor is going to be in the same few orders of magnitude. The next few stages in B+ are likely to be in the 10K-ish reason, and electrolytics that follow them are usually in the 10uF to 20uF region.

I suspect that given how high the series impedances are, even ESL strapped electros reduce noise and such pretty well. We actually know this from the equipment of the "Golden Age", where they use poor (by today's standards) electros and managed good results.

All of which is what led me to say "maybe". And that I'd like to see some real numbers, not a ton more speculation.

[didit]

I'd like to see some real numbers, not a ton more speculation.

Yes. Would someday look to oblige, or at least help out. Measuring this, particularly fully out to the acoustic end of the chain is challenging. Spacial perception of phase is somewhat less speculative than it used to be but how that plays when playing an amp ... still it fits with some less than empirical observations.

Ian

[Leo_Gnardo]

If you have ever driven an amp and smiled until you got 10th 11th fret and higher on EBG strings and had the amp eat your sustain from phase beat you might want to try it

I sure have noted a gargle tone on some higher single notes played loud on amps with poorly filtered / failing power supplies. Tried the small parallel cap trick. No improvement noted. Replace failing or insufficient filters, definitely an improvement.

"Phase beat" gonna have to add that one to my vocabulary of tone descriptive lingo. Makes me wanna hear some Blondie, "Eat to the Beat." Deb beat the rest of the crowd to rap, Imma lovin' it! Hand my my dentures...

[Tony Bones]

If you mentally switch gears and think of the PS caps not as filters but as bypass caps it starts to become apparent how some strange interactions between stages can take place. The obvious result is motorboating, but it's not inconceivable that more complicated interactions at higher frequencies can take place.

Really, I think power supplies are WAY more complicated and subtle than most of us give them credit for. In particular, chokes seem like a good idea at first, but if you're not careful you can end up with a PS that rings at some audible frequency. Since the ringing is not related to the music, it comes through as, well, non-musical. And, in most cases, nobody ever sees the ringing when they hook up a scope because we look at the B+ while the amp is idling, or maybe with a steady input signal, but we don't excite the ringing while we're looking, so we don't even know that it's there.

Anyway, the choke leads to a supply that does not respond quickly to changes in load and actually leads to an unclean sound. A resistor works better. A resistor and choke in series can work, but you need to tune the series resistance to get the damping right. When have you ever heard of anyone doing that?