Supply Filtering

[Structo]

I was reading some info at The Valve Wizard site yesterday about power supply filtering.

Something that interested me was the concept of piggybacking a cap that is 10 X smaller than the first reservoir filter cap on the first cap and possibly the second large filter cap as well.
This is what he said:

The reservoir capacitor needs to have a low reactance at frequencies up to 40kHz, because the HT at this point will contain high frequency harmonics caused by rectification. Its performance at these frequencies can be greatly improved by adding a capacitor in parallel with it, ten times smaller in value. (Another can also be added, ten times smaller than the second, but this would be usually be considered an extravogance.)
Do you guys implement this into your amps?
Does it tighten the bass or have other desirable effects?


Also he talks about using snubber caps in parallel with the silicon diode rectifiers to filter switching noise.
I wonder how many of you do this.
<snip>
If diodes are used in series, a capacitor should be placed in parallel with each one to ensure equal voltage sharing between each diode. Values of 10nF to 47nF (1kV or better) are common, and they will also act as snubbing capacitors to suppress the voltage transients produced by the diodes switching.

[img:258:177]http://www.freewebs.com/valvewizard/Fullwave2.jpg[/img]

[chief mushroom cloud]

This is common w/ high speed electronics. In a guitar amp, you aren't gaining anything worth putting in the extra caps.

Electrolytic caps ESR and ac reactance aren't linear....so adding adj film cap (or better yet, monolithic ceramic) adds a high freq reservoir to keep switching currents at a minimum....

putting 'snubber' caps across rectifier diodes is a trick as old as diodes themselves. it minimizes switching transients that can cause minimal amounts of 120Hz buzz (although rarely)

put some FREDs in there (UF4007)

IT's ALL ABOUT THE RISE TIME

[chromaticdeth87]

If you have properly designed LCR network after rectification and like the above poster states, the use of FREDs, then you should have no use for those caps "upstream"

[Structo]

So the FREDS are quieter than the 1N4007's?

[drhulsey]

...IT's ALL ABOUT THE RISE TIME

That's what my wife said

[chief mushroom cloud]

So the FREDS are quieter than the 1N4007's?

in a sense, yes
less switching transient spike...but they tend to ring longer, but this is not an issue w/ guitar amps...or even any audio amps

there is no sonic voodoo about FREDs...they don't 'sweeten' or add detail or any other cork sniffing bull$hit....they are just a tad quieter....but I've heard a lot of amps w/ 1N4007s that are pretty damn quiet....it's just a bootstrap for noise

I use them because they're cheep enuf and same size as 1N parts, for insurance

you could do a LOT better w/ a filtered IEC inlet, ESPECIALLY w/ amp amp you're going to gig w/

[Structo]

Well to be fair, the reason the diode snubbers attracted me is because Gil Ayan had mentioned that he used them on all his amps.

As far as switching noise, I may be experiencing that, but it also could be my Stereo Memory Man in the loop that is generating some of that.

And the piggyback filter caps was something I noticed on a modded Ceriatone OTS that had 50uf caps piggybacked on the 100uf caps.
Like in this layout.

[chief mushroom cloud]

hmmm, that looks like he's trying to get 75uF HT filtering at 700+ V..nothing more.

But I could be wrong. There might be some voodoo going on in there.

Ooga Booga

[Andy Le Blanc]

their purpose is to improve the transient responce of the amp.
Its usually say a .1uf parallel with say a 100uf. with a srpp, or mu-follower
and the rest of that ilk, theyll be right on the plate and or in the supply and
reciprocated in the cathode circuit. They work. The transient presentation
and sound stage are cleaned up a bit, less blurr in the sonic image.

You need to careful with a guitar amp. All that increased detail and presence
can make for a very brite and over-tone-ish amp.

[Structo]

hmmm, that looks like he's trying to get 75uF HT filtering at 700+ V..nothing more.

But I could be wrong. There might be some voodoo going on in there.

Ooga Booga

Correction, when capacitors are placed in parallel, the values are added, unlike resistance.
So the total there is 150uf.

Yeah, I don't know who made that layout and if the theory from the Valve Wizard is valid the piggyback caps should be 10uf right?

I also agree that if you implement too many Hi Fi tricks in a guitar amp, it can make it sound sterile or lack luster.

Guitars amps are primitive amplifiers and some of their "m0jo" comes from the harmonic overtones and other artifacts that would make an audiofool cringe.

[chief mushroom cloud]

their purpose is to improve the transient responce of the amp.
Its usually say a .1uf parallel with say a 100uf. with a srpp, or mu-follower
and the rest of that ilk, theyll be right on the plate and or in the supply and
reciprocated in the cathode circuit. They work. The transient presentation
and sound stage are cleaned up a bit, less blurr in the sonic image.

You need to careful with a guitar amp. All that increased detail and presence
can make for a very brite and over-tone-ish amp.

ooga booga

edit:
goddamnit, andy, do you really believe that? does that make sense when you have so many other bandwidth limiting components down the signal chain...INCL yer guitar?
why not put a 3uF film cap on EVERY plate resistor? this way you decrease inductance of the wire and decrease esr. does this make sense? this way you can DECREASE filtering. what about multiple e-caps in parallel. this way you decrease inductance AND decrease esr...why not do that? it will DEFINITALY improve transient reponse capability at that node. jesus h christ, I remember when dan boul was raving about barker caps, because they were 'fast'. fuck...somebody just fucking shoot me right now.........

OOGA BOOGA!!!!!!

[chief mushroom cloud]

Correction, when capacitors are placed in parallel, the values are added, unlike resistance.
So the total there is 150uf.

you're right....I'm an idiot....wtf do I know
I'll just shut up

edit:
NO I'M NOT GOING TO SHUT UP

Structo, in that layout, there are two sets of paralleled caps, each added to make 150uF...BUT...they are stacked in series w/ each other...w/ balancing resistors, to make 75uF total for that B+ node.

[Andy Le Blanc]

OOga-BOOga-BOO, its the appilcation. Whats that Floyd album, two discs,
with the fly and the bird takeing off, and the grooving pict, umaguma.
how about yes on 25.......

its the illusion of audio reality without the drugs. despite all the pompus B.S..
anything to make more real.
Whats the magical mix of caps (no pun) in a trainwreck that puts it over the top.

[chief mushroom cloud]

trainwreck....bad example

illusion is the right word

bla....

I'm not knocking anything that might add coloration to an amp's sonic characteristic (incl adding .1uF in parallel)...but I'm hoping that one would really think about if what they are doing makes any sense......

i'll shut up now.....

[Andy Le Blanc]

your not wrong,
for a guitar amp it doesnt go so well, but for a relativly simple tube pre for Hi-FI
its one approach that really does affect the subjective quality of the amp.
its also fairly common, build a tube pre or a home stereo amp and
explore it yourself. Ive built a small stereo for the TV, DVD's end up being what gets
listen'd to the most, and sound tracks are very demanding.
You really do end up chasing after the detail, Its another way to follow your ears.
you might be suprised. Its just another amp after all.