What would you do?

[daz]

I'm pretty sure this will be met with "thats not a good idea". So if thats the case what would you do that WOULD be more correct and would give the same result. In my many attempts to get the tone "right" at the PI and beyond, because i just know it never had been quite right due to levels i believe. I decided to lower the plate resistor values. In order to keep the 100k/82k ratio correct i decided to parallel the 82k with the same and the 100k with a 100k to get 50k/41k. It worked well to get me closer to my goal. But not as much as i hoped, so I then i paralleled each one with a 3rd resistor. So my plate resistors are now the equivalent of about 33k/27k, and the tonal improvement is now great, and the closest thing i have yet found to my goals for the PI/PA. For the first time it sounds like i feel it should, it sounds much more "right" than it ever has. But I need to be sure this isn't a bad thing to do for some reason.

As i said i'm sure it's not going to find favor here for some technical reason, so have 3 questions...

1-is this indeed a bad thing for some technical reason or is it ok to do?
2-the plate V is now about 300-309VDC, so if the answer to #1 is that it's ok, should i drop the voltage to the PI in the rail?
3-if this is just a bad configuration for some reason, what else could be done to lower the gain within the PI that should give the same result?

[Gingertube]

daz,
This worries me for a few reasons.

By reducing the anode loads on the PI then you are doing 3 things:
1) reducing the gain significantly
2) increasing the distortion significantly
3) reducing the headroom before clipping in the PI significantly

Are you using global feedback?

I ask this because i'M wondering if you have shifted the first clipping from the output tubes to the PI. Global feedback would agravate this.


Don't worry too much about that 82K 100K ratio. That ratio applies to a particular Fender design. With a 10K "tail" resistor. The 82K was calculated to give near perfect balance.
As soon as you use another value of the "tail" resistor (and many designs use 15K or 22K) then the 82K / 100K is no longer valid anyway but guys just copy it it "ignorance".
http://www.valvewizard.co.uk/acltp.html - see the para on balance

Cheers,

Ian

[daz]

daz,
This worries me for a few reasons.

By reducing the anode loads on the PI then you are doing 3 things:
1) reducing the gain significantly
2) increasing the distortion significantly
3) reducing the headroom before clipping in the PI significantly

Are you using global feedback?

I ask this because i'M wondering if you have shifted the first clipping from the output tubes to the PI. Global feedback would agravate this.


Don't worry too much about that 82K 100K ratio. That ratio applies to a particular Fender design. With a 10K "tail" resistor. The 82K was calculated to give near perfect balance.
As soon as you use another value of the "tail" resistor (and many designs use 15K or 22K) then the 82K / 100K is no longer valid anyway but guys just copy it it "ignorance".
http://www.valvewizard.co.uk/acltp.html - see the para on balance

Cheers,

Ian

Well, as long as there are no technical issues that could eat tubes or cause some other problems, I'm not too worried about the tonal implications you mentioned at lower volumes because it sounds better than ever. If you are right about those things then i will likely find out when i crank it up which i will hopefully do tomorrow after work. Theres way more gain going into the PI than i want so #1 is not a worry, and in fact is one of the things i was trying to accomplish. #2, well, not at the lower volumes i tested it at but again i will have to see what happens when i turn it up. I'm a bit curious tho....i thought higher plate voltage and smaller plate resistors=more headroom, cleaner operation. No? As to PI balance, i just wanted to retain that ratio because i tried a PI balance trimmer and found it sounded best in the middle, IE: 100/82k. And with no intention of changing the 10k tail i figured i should keep that ratio.

But assuming what you listed turns out not to be a problem, theres no reason not to use such small plate resistance? Or do you know of a better way to accomplish this?

[Enzo]

Have you tried the stock circuit but with other tubes? For example stick a 5751 in there instead of the 12AX7.

[daz]

Have you tried the stock circuit but with other tubes? For example stick a 5751 in there instead of the 12AX7.

As a matter of fact, yes....5751, 12AT7, and 12AU7. They reduce gain, (tho the 5751 didn't really seem to noticeably) but they don't do what this did. This changed the tone for the better while lower gain tubes changed it for the worse. Not horribly, just not quite as good let alone much better as this did.

[ReadyTeddy]

daz,
This worries me for a few reasons.

By reducing the anode loads on the PI then you are doing 3 things:
1) reducing the gain significantly
2) increasing the distortion significantly
3) reducing the headroom before clipping in the PI significantly

Are you using global feedback?

I ask this because i'M wondering if you have shifted the first clipping from the output tubes to the PI. Global feedback would agravate this.


Don't worry too much about that 82K 100K ratio. That ratio applies to a particular Fender design. With a 10K "tail" resistor. The 82K was calculated to give near perfect balance.
As soon as you use another value of the "tail" resistor (and many designs use 15K or 22K) then the 82K / 100K is no longer valid anyway but guys just copy it it "ignorance".
The Valve Wizard -Long Tail Pair - see the para on balance

Cheers,

Ian

Reducing the plate resistors, if anything, should reduce the distortion in the PI stage, because it reduces changes in Vag. It will reduce gain within the output-stage feedback loop, and that will increase the overall output distortion, but perhaps not by much, and perhaps in a way the OP likes. The reduced open-loop gain will also reduce the damping factor of the output, which the OP may like. Additionally, by reducing gain in the PI, the output tubes will not get driven so hard, and maybe that eliminates a source of distortion that bothered the OP. All this is speculation on my part.

[potatofarmer]

Fender in the CBS era used 47k plate resistors, 470R bias resistor, and a 22k tail resistor for their later LTP designs. Granted, they used a 12AT7 for the inverter - have you tried a 12AT7 with the lowered plate resistor values?

[daz]

Reducing the plate resistors, if anything, should reduce the distortion in the PI stage, because it reduces changes in Vag. It will reduce gain within the output-stage feedback loop, and that will increase the overall output distortion, but perhaps not by much, and perhaps in a way the OP likes. The reduced open-loop gain will also reduce the damping factor of the output, which the OP may like. Additionally, by reducing gain in the PI, the output tubes will not get driven so hard, and maybe that eliminates a source of distortion that bothered the OP. All this is speculation on my part.

Thats all more or less what i like, yes. My ultimate goal if i thought it were possible (i know it's not) would be for the PI on back to sound the same at any volume and simply amplify the preamp tone which i spent forever getting where it is now. Even if it did or course i realize my ears wouldn't hear it that way. But my main goal is to get it as close to that as possible so that the tone i get low where i mostly use it wouldn't change much as i turn up, and extra distortion/compression would be acceptable and even welcome as long as it's the right kind and only compliments the preamp OD. I also have always felt the signal being to hot into the PI was causing other issues and thought by curing that i would see a overall improvement in tone in certain ways, and this seems to do that. I've tried a number of things over the last few weeks or even months and many seemed to do the trick to one degree or another but in the end failed by creating other issues. Hopefully this one will be all good, and at low volumes it certainly seems to be nothing but goodness. There it's the best the amp has sounded to date and certain things that have always bothered me seem to be gone. But said the same several times only to find when i turn it up thats where it reveals the bad. So far all the others did.

If it DOES pass all tests, can anyone answer the question i posed about the plate voltage? That being that with small plate resistors the voltage went from about 240 to 310, so should i lower that via the PSU rail ? Or is that part and parcel of what i like about it and is it not too high to where it might kill some tubes? This is one of those many things i have tried that i don't see in other designs so of course that always worries me a bit.

Fender in the CBS era used 47k plate resistors, 470R bias resistor, and a 22k tail resistor for their later LTP designs. Granted, they used a 12AT7 for the inverter - have you tried a 12AT7 with the lowered plate resistor values?

No, in fact i just tried this last nite so i haven't tried anything else. Any particular reason i should?

[potatofarmer]

If it DOES pass all tests, can anyone answer the question i posed about the plate voltage? That being that with small plate resistors the voltage went from about 240 to 310, so should i lower that via the PSU rail ? Or is that part and parcel of what i like about it and is it not too high to where it might kill some tubes? This is one of those many things i have tried that i don't see in other designs so of course that always worries me a bit.

Why would you? The published maximum plate voltage spec for a 12AX7 is 300V, sure, but that's with respect to the cathode. Your cathode voltage should be noticeably higher than 10V, so if you like the sound and the tube's biased to a safe point, leave it alone.

No, in fact i just tried this last nite so i haven't tried anything else. Any particular reason i should?

I was just curious. Those conditions look like they'd better fit a 12AT7, but again if you like the tone...

[daz]

I was just curious. Those conditions look like they'd better fit a 12AT7, but again if you like the tone...

I'll give it a try then, what the heck.

[printer2]

Is the bias resistor right for the new plate loads?

[daz]

Is the bias resistor right for the new plate loads?

Don't know. I didn't change it but i wouldn't know what to change it to.

[daz]

Cranked her up just now and no go. Still pretty bright and hard. But heres an idea i never thought of and it really has a major effect on smoothing out high end even at higher volume. I saw this in a electroplex schematic and tried it. A .0047uf (they use .005) across the cathode follower's 100k cathode resistor. Effectively cutting highs at the input to the tone stack. I have lots of highs in the pre which lead to the output getting quite bright as i get loud, but those must stay because they shape the OD. I have always tried to cut after the preamp but never found a good way and never thought of this. Seems like it could be a bit too smooth after a while but if it turns out that way a resistor in series might do the trick. It'll require some more testing as always. The plate resistors may go back to stock with this mod.

EDIT: upon further experimentation, with a largish cap (say .02uf-.0047uf) this mod, a cut control, and a cap bypassing the NFB resistor all sound like the same effect to me. I use caps in the .02-.0047uf range mostly. And unlike using a small picofard size which cuts above a pretty high frequency, with a large cap like that it give more of a gradual slope all the way from a pretty low frequency on up so it's not an abrupt slope at some point in the mids to highs which can be hard to make very useful. This is too much by itself, but in series with a resistor, preferably a pot or a pot till you know the value you want, you can really tune it to perfection.

The difference between this and a cut control or a cap across the NFB resistor is one very important one.....it's pre power section so what you get doesn't change with master settings, which is what makes cut controls and that cap across the NFB R sorta worthless to me at least. A cut for example works great till you turn the volume up, then you have to cut more and at some point it either runs out of cut ability or if it can cut more (depending on cap value) the power amp gets to it's volume limit a lot lower and the tone becomes rather wooly and loses focus. Another advantage is it supplies the tone stack with a darker tone so that the treble and mid controls are more useful.It's in the perfect place to set up the tone stack and PA with the right EQ curve if it's not already receiving that.

That said, this may not be real usable unless like me you have a very bright preamp, tho i may be wrong since this can me switchable and tweakable with a series pot. What really has me scratching my head is that the rocket 50 has a .05uf there and no resistor in series with it which would make the tone way soft and dark in any amp who's preamp isn't stupid bright, even much more than mine. Even more surprising to the point of my jaw hitting the floor is that the rocket 100 had the same thing but with a cap value of 1uf ! Granted, it's switchable in that model, but a 1uf would render my amp unusable. I tried a .02 and that almost did. I cannot for the life of me understand whats going on there unless the preamp is insanely bright, which it doesn't look to be. Even then i think that would be a very odd circuit if even usable. Those amps were some of the first on the boutique market and i remember many years ago reading a review in GP that praised the rocket 50 like it was the second coming. If you're interested, here are the schematics. (available til schematic heaven dies in about 2 weeks) I'd love to hear opinions on this, as it seems to work great and this is the first i have seen such a tweak in any schematic if memory serves...

R50: http://www.webphix.com/schematic%20h..._rocket_50.pdf

R100: http://www.webphix.com/schematic%20h...rocket_100.pdf

[Gingertube]

The 1uF on the Rocket_100 schematic is a drawing error - No way will there be a 1uF in the circuit. 1uF across 100K would form a 1.6Hz low pass filter. Maybe 1nF not 1 uF.

Cheers,
Ian