Annoying overwhelming brightness. Power amp problem?

[Dingleberry]

Hi.

I've been building a 6V6 p-p cathode bias amp and tried to do about averything to tame the high end. One 12AX7 in the preamp, one in the P.I.
I've rebuilt the preamp about ten times now. I have tried everything from single gain stage, parallel triode, cascade, with different tone stacks before and after the second gain stage and tried nearly every possible pre amp configuration what one can do with one double triode. Different speaker configurations and switching from 4 to 8 ohms didn't do much. Used snubbers in the pre amp plates and even went to 1nf in value. Wired also a Vox-type cut control but there is still annoying, not pleasing in anyway brightness. Harsh and "hard" in your face type tone. Far away from mellow and tubey. I'm coming to conclusion that the problem lies somewhere in the power amp stage.
Preamp voltages were good P.I voltages and balance (scoped that) are right on spec, and even 500pf cap across P.I plates. That led me thinking that how much will the Plate-Screen voltage relation affect to the sound?
Not much info I found after searching.

Little data for you:
Plates are now at 350V and screens are at 323V.
Plate supply node is 356V and screen supply node is 324V.
Screen resistors are 1k.
Cathode is at 21,65V with a shared 330ohm resistor,so they are biased about 11,5Watts each. Tubes are JJ 6V6's so I probably could bias them a bit hotter too. They are rated for 14W max.

I have a resistor+choke combination between plate node and screen node, very much like it's in the Mhuss Plexi 6V6 schematic.
I have now there 4k7 resistor and choke that has dcr 275 ohms so it makes 4975 ohms. So making that resistor smaller before the choke I could lower the voltage drop. And get the screens higher.
With only the choke the screens were 2 Volts higher than plates so that's why I installed the resistor and used the Plexi 6V6 value.


Could it be the remarkably low screen voltage that makes the amp sound kind of hard, cold and earpiercing bright?
How about the power supply filtering? I used what I had on hand and filtering arrangement is: Plates:110uf -4k975- Screens:56uf -9k1- P.I:33uf -22k- Pre amp:22uf.
Could that little stiffer (for example comparing tweed deluxe) filtering affect amp tone in a bad way?

Anyone other have had same kind of problems and how did you solved it?

-T

[Firestorm]

What kind of PI? How is it configured? Are you using feedback? Global or local? Grid stops on the 6V6s? I can't see how your screen voltage would cause that problem; it doesn't seem especially low. There is always the possibility of some kind of parasitic coupling even though the amp is clearly not high gain.

[Dingleberry]

Long tail P.I. with
470 ohm bias, 22k tail and 1M grid resistors. 10nf input cap and 100nf from grid2 to ground.
No feedback neither global nor local.
5k6 grid stoppers right now. Tried values up to 47k with no noticeable difference. Very simple and nice layout, shielded cable in the necessary places. I did not notice any oscillations when scoping.

-T

[xtian]

I had luck adding a conjunctive filter across the OT primary on my SE build--don't know if this is helpful for a PP design.

[Gaz]

A conjunctive filter is a good idea because it is even after the OT, that way you can see if the problem is in the power amp. Maybe the amp just doesnt have enough bass?

[David Root]

500pF across the plates surely takes care of any possible parasitics in that area. After all the standard Fender value was 47pF.

What are your preamp cathode bypass cap values?

I would try the conjunctive filter too at this point.

[Dingleberry]

First cathode: 1k5/4u7 (poly), second cathode: 2k7/1uf (poly).

-T

[David Root]

Gaz might be right. 1uF is pretty lean. Can you post a full schematic so we can see the whole picture?

[rp]

Are you copying a known circuit or is this an original design? If a vintage build then on the magic mojo side rather than the technical side if it's all metal film and modern caps (715s, hifi types, etc) then try all carbon comps and old school style caps like 150s, 6PS (still bright but not cold/sterile), or maybe the Sozos, Jupiters that I've yet to try. If you're using a high spec OT like a Hammond 1600 then maybe something like a Heyboer.

Anyone other have had same kind of problems and how did you solved it?

I built a few clones that were too bright and cold and shotgunning to the above changes did the trick.

[Andy Le Blanc]

Is the cathode resistor bypassed?

[Dingleberry]

Thanks for replies guys.
It's a sort of mixture of existing designs.
12AX7 in the preamp, 12AX7 LTPI and 6V6 p-p cathode bias power stage, no neg feedback.

Actually now when I think of it, nearly all CATHODE BIASED 6V6 designs what I've seen has not LTPI.

Can someone point me to direction of "PROVED to sound good" amp which has 12AX7 preamp (in what ever configuration), 12AX7 LTPI and 6V6 p-p cathode bias amp with no feedback.

I sorta copied Tweed Deluxe power amp, Blackface Deluxe reverb PI (without feedback) and been trying different preamp designs.
This has been kind of platform for trying different preamps and listening the differences and just to see what will different preamps sound and what kind of will be the one I settle on.

Could it be that the LTPI is not the best PI to drive cathode bias 6V6's?
Is there some impedance issues that I should be concerned about?

Iron is recycled from old "HI-FI" tube amp/integrated radio system. There is a cocktail of CC's, MF's and some CF's in the amp.
MO and WW in the PSU.
Mostly 150-type caps, and polystyrene for small values.

Today I have the time to get back to it so I'll report what I find out.

-T

[tubeswell]

Hi Dingleberry

Have you got a complete schematic you can post?

[surfsup]

What are your preamp stages biased at? Maybe you've got some cold-bias stages there that are causing the harsh brightness?

[Firestorm]

Just another thought: with the low screen and plate voltage you have, your output tubes may be idling too cold. Maybe try the "traditional" value of 250R (or even 220R). Cathode biased amps running in Class AB can develop weird distortion artifacts because the current through the cathode resistor varies so much with signal. The fact that you're using a "hi-fi" OT might exacerbate that.

[Dingleberry]

Thanks guys. I'll have to check the bias issue too, both pre amps and power amps.
Probably try also convert it to fixed bias... It starts to seem like one of those never ending projects...

PT does not have a bias winding.
Do I have to use an external bias voltage transformer or can I tap the other lead of the secondary? PT has full wave bridge, no center tap.
I assume that you can tap the other leg only with PT's with center tap.

I think I might have some small transformers lying around that I can use for creating a bias voltage supply.

Whats a good range for 6V6's or should I just go with Deluxe reverb topology?

-T