Another PSU question

[daz]

Reading that other thread reminded me of a question i've always had but never asked. Hum and other PSU related noise isuues aside, because my amp has always been quiet, what exactly are the possible tonal benefits of parallel nodes instead of the typical series rail? I noticed the amp i based my output section on has this, the matchless chieftain. Mark Sampson obviously knows his stuff, so there must be some reasoning behind it.

[Chuck H]

There are at least three reasons that a designer might do parallel instead of series that I can think of.

One would be to get higher preamp voltages than you could with a series design. Some of the designs discussed here have used low voltage power tubes but the builder wanted higher preamp voltages. So if you wanted to build, for example, a Dr. Z type el84 power amp with 330Vp but wanted a BF Super Reverb type preamp with 250Vp at V1 (which is still low for a SR) the only way to get there would be a parallel supply.

Another reason I can think of would be to reduce resistance in the rail that feeds the preamp to reduce sag or otherwise keep whats happening voltage wise between the power and preamp circuits apart so that changes in current in the power amp don't effect the preamp as much.

And the last reason I can think of would be to give the preamp and power amp more independant filtering. As long as the power amp and preamp are on the same rail there is more risk of coupling of any sort. Including noise. Electrolytic caps have comparably high impedance (called ESR for equivalent series resistance not long ago) next to film caps. So there is usually some noise hiding in the B+ rail. If that noise gets onto the preamp it can be greatly amplified. So the parallel arrangement would minimize this.

That's all I can come up with. If Im full of $h1t I'm sure someone will let me know.

Chuck

[daz]

So then i guess the only tonal reason would be to get higher preamp voltages. But since i actually have rather large dropping resistors to get the lower voltages i prefer, no sense in it for me i guess.

i'm trying to figure out what the best PSU for my amp would be because i've found big tonal differences can be had. So i must assume my PSU is anything but optimal since it's just a mutt i threw together looking at several schematics. But i suppose i can stop thinking about parallel nodes.

[defaced]

In the design you're looking at, where do the nodes start to parallel? In Chuck's post, it sounds like we're talking about a completely parallel design where the preamp and power amp node both come off the rectifier. I've seen designs, namely the Soldano SLO 100, that simply parallels the last two gain stages. I've not played with the designs, but something tells me they're not going to result in the same sound/feel.

I also wouldn't be too harsh on your PSU design. Optimum means different things depending on what you're after. To me, if I don't hear noise from the PSU, I'm close enough to optimum that it's just getting that last 5% to me.

[Chuck H]

But since i actually have rather large dropping resistors to get the lower voltages i prefer, no sense in it for me i guess.

i'm trying to figure out what the best PSU for my amp would be because i've found big tonal differences can be had... i suppose i can stop thinking about parallel nodes.

Maybe, but look at reasons two and three also. Maybe your amp would sound better with less sag in the preamp voltage supply. Maybe you would like the lower noise.

You could do a parallel supply with a big a$$ Pi filter that would sag very little and supply the correct voltages to your preamp. It would sound very differen't from what you have now. Remember the "bolders in a stream" analogy? Those large dropping resistors create sag in your preamp response. Also, the power tubes current draw affects the preamp voltage with a series arrangement. So there is alot more to it than just mean voltage figures. There is the behavior of the supply.

I haven't found a reason to veer from the standard series rail supply myself. But if I ever saw one I wouldn't hesitate to try a parallel curcuit.

Chuck

[daz]

You know, i just spotted a mistake in my PSU that has been there since day one !!! The PI has no cap! Look at bthe schematic. It's in need up updating, but look closely. I don't know why I never saw that, nor did anyone else who's looked at it. But do you see a filter for the PI? there isn't one ! this could explain some things.

http://i24.photobucket.com/albums/c3.../myEL34new.jpg

EDIT: maybe it's because i can't play it past bedroom volume right now. Maybe it's placebo effect. i don't know. But damned if i didn't add a cap there and the freakin amp just got much cleaner and fuller and just better. i cannot wait til tomorrow to really test this. (darn neighbors and thier desire for peace and quiet ) This could be the magic bullet i've been looking for, and turns out the magic bullet is just fixing an idiotic mistake by yours truly. geez.....maybe i should take up knitting or something.

[Chuck H]

Well the fact that no one else saw it is a point in your favor. I just made a similar error on one of my designs. Steve Connor spotted it. The funny thing is that on paper the schem was flawless. But Steve's familiarization and repetitive implementation allowed him to see the fault. You can read about it on the "attenuator build problems" thread. Which you have already posted on.

The truth is that most of the power supply issues are hammered out by the first two filter nodes. But what you get by putting the cap in for the PI is a bit more stabilization and probably less voltage drop on V1 and V2 on power chords. Let us know how the testing works out.

Chuck

[daz]

Well, NOT in my favor is the fact i went all this time not noticing it ! In any case, the difference i heard while A/B'ing the cap in and out of circuit seemed very obvious. the difference was that the tone was more round, full, and lacking the distortion that was there without it that I DIDNT EVEN NOTICE ! Not distortion as in overdrive, but a general rattiness that was keeping the tone from sounding pure and solid and clear. Thats the best i can explain it. But i have had an issue from day one that i think this may well have been responsible for. That would be that when i tried a post PI master in it, the tone would just become ratty thin and completely useless past about 2 or 3 oclock on the master. It also manifested itself, or so i think, as a loss of smoothness and complexity in the tone at louder levels. It just seemed like at loud levels i wasn'[t getting any EL34 goodness, just rattier harder tone that was not even close to as nice as it was at say 9 or 10:00. This has GOT to be the reason. All this was centered around the PI and it's the PI that had no filter. granted, it's filtered before it, but thats apparently not enough or we wouldn't see rails with caps at every dropping resistor. Apparently the PI filter is important from what i've read tonite and seen in schematics i've been looking at, which always have a PI cap yet sometimes share one for V1 and 2. anyway, i'll update you tomorrow.

[daz]

Well cr@p. Brought it to work and cranked it up. I think it's actually better w/o the filter on the PI ! I swear, if i live to be 1000 i will never learn the lesson i actually know all too well but never heed.....when something sounds great, chances are it just does right then and there under those circumstances and probably won't anywhere else. I coulda sworn this was a big improvment.

[guitician]

Ahh, it's like Chuck H said, the whole amp is a dynamic circuit and one thing effect the other. In this case your added filter changed the preamps sound.

[daz]

Another thing i'm now beginning to see is that the more filtering the more high end, and it gets harder feeling. With less filtering it gets smoother tho less articulate. So i guess put simply, more filtering=more high end.

And check this out....i DID have PI filtering all along. I just realized that when i wired this thing about a year ago, i wired the B+ under the board so i can't see it. So i forgot exactly what went where and was looking at the orientation of the dropping resistors and thinking the PI was on the resistor side of the 2 nodes it's in between, but it wasn't. HOWEVER, a week or so ago, and i posted about it, i changed the way the first filter is wired and since then the screens has no filter. So i just added one. It made the lows tighter, so that helps.

EDIT: i was a bit off base...it's more than just tighter...it's a lot better in pretty much every way. In fact, it sounds better than it ever has. I then took it home and it passed the most crucial test of all, the one it never really has at least to this degree....it sounded equally good there and at work. Thats the test it rarely passes to this degree. I guess the screens were never happy. at first i had a 33uf/33uf with one side for the plates and one for the screens. then a week ago i paralleled the cap leaving no filter on the screens. It actually sounded better like that, and i a/b'd it numerous times to be sure. So now i have added back 80uf to the screens and it's just fabulous. I got a 50/50 last week and that is paralleled for 100uf on th plate and i have a seperate 80uf on the screens. I just ordered some more stuff, and among it is a 100uf/100uf that will replace the 50/50 and one side will be for plates and one for screens. All preamp nodes have 80uff too. Seemed like the more UF i had on the screens the better. And now i look back and wonder if the lack of filtering is why i love sag yet never liked the way the amp felt with bigger screen resistors. I may have to try that one again now. anyways, now i don't want to stop playing the thing ! yeah, i know....boy who cried wolf. And sure enough i'll probably be effing with this amp forever. but honestly, this is easily the best it has sounded. And Chuck.......maybe tonite i can go to sleep w/o the iron.