Voltages - other than expected

[txstrat]

I have this one amp that is not behaving like I would want it to do. It sounds a little too harsh and brittle for my ears. I tried to troubleshoot it with different bias settings and was thinking about cathode bias. It's a Vibrolux (6G11) circuit. It has a PT with 0---300v sec., diode recto and 50v bias tap, 90ma fender choke, 35w OT at 4ohms.
At first it was running at 387v at the plates (386v at the grids before the 470 ohms resistor, 385v after the resistor) and I biased it to around 56ma.
First thing I found was that one of the tubes had drifted. When the 1st tube was at 52ma the 2nd was at 61ma (9ma difference).
I thought it would be better to turn down the 2nd tube and checked the voltages at the plates before. Now the plates are at 397v. Thatfore I turned down to 52ma for the 2nd tube and found that the other was at 47.5ma (only 4.5ma difference).
The grids are now running at 399v before and 398v after the grid resistor (470 ohms 2w). Both higher than the plates. As far as I understand that should be avoided.
What am I missing?
Any help appreciated.

Matt

[MWJB]

Firstly you don't need anything like the currents you are talking about, 30-35mA should work fine, a little more won't hurt if you like the tone but at 50mA-ish I'd be concerned about the current load on the OT.

Screens don't draw much current, your plates are drawing an awful lot, this difference alone might explain why the screens are a volt over the plates...even so a volt more at the screens over the plates is not going to damage anything. Rebias to <40mA and see how they settle down.

[txstrat]

MWJB
Thanks for your quick reply.
I was using the weber bias calculator.
According to that a 6L6GC (didn't mention that before) running at 375v can be biased at 48ma (60%) or 56ma (70%).
At 400v that would be 45ma to 53ma.
I thought running the tubes cooler would result in a harsher sound due to disharmonic distortions.
Anyway. I'll try what you proposed, run the tubes a little cooler and see what happens.

[MWJB]

Most 6L6GC on the market today are not true 30W tubes, the JJ is. Unless you are using JJ/quality NOS I think it's safer to assume 25W.

Even so, a tremolux would never have left the factory with 40mA+ (too much risk of warranty return, blown OT especially with the bias trem), it should still sound as factory intended @ at least 30mA. If the amp sounds harsh & brittle, it is for a reason other than plate current...how do the other voltages check out?

[txstrat]

The plate voltages are around 200v on the PI and 160v on the preamp plates. Haven't measured other voltages like cathodes etc..
And they are JJ 6L6GC.
Ooops, WRONG, they are TungSol 6L6GC. There are also TungSol 5881 (25w) on the market so I assume the 6L6GC are 30w, right?

[MWJB]

Voltages are a little down on the PI & preamp, check the value of the power supply resistors?

[MWJB]

Any current ratings for that PT?

[txstrat]

0-300v @ 200ma
50v @ 60ma
6.3v @ 4A
14.5v @ 2A (for relays etc.)

[MWJB]

Cool, no worries there.

[tubeswell]

At first it was running at 387v at the plates (386v at the grids before the 470 ohms resistor, 385v after the resistor) and I biased it to around 56ma.
First thing I found was that one of the tubes had drifted. When the 1st tube was at 52ma the 2nd was at 61ma (9ma difference).
I thought it would be better to turn down the 2nd tube and checked the voltages at the plates before. Now the plates are at 397v. Thatfore I turned down to 52ma for the 2nd tube and found that the other was at 47.5ma (only 4.5ma difference).
The grids are now running at 399v before and 398v after the grid resistor (470 ohms 2w). Both higher than the plates. As far as I understand that should be avoided.
What am I missing?
Any help appreciated.

Matt

Hi Matt

As I understand it, when you reduce the current flowing through the tube, the potential difference at the Plate increases because the Plate is not getting 'neutralised' by as many electrons as before. So increased plate voltage as a resulted of reduced tube current sounds normal (Was that part of your question?). However the screen current is determined by separate resistors, therefore as you reduce the tube current, the screen voltage will also rise, but at a differential rate to the plate voltage. If you're worried about the screen voltages being higher, you can always put in 1k or 1k5 screen grid resistors (just make sure their power rating is sufficient). Ideally screen voltage should be at least a volt or two below the plate voltage to ensure that you get optimum gain out of the tube, and to prevent the screen dissipating too much. Having said that, screen voltages are not always below plate voltages. (On the 5E1 for example, the screen is hooked up to the B+, and yet the plate is going through the OT primary. The effect is reduced power output from the amp)

Also are you sure the tubes have 'drifted'? It is not unusal to have similar tubes measuring up differently on different sides of the OT primary. DC resistance can differ on different sides of a primary winding due to one side being wound 'on top of' the other, and the total lengths of wire on each side of the winding's CT therefore being different (which affects the current being drawn through each side of the winding).

[txstrat]

Hi TW

I wasn't sure if the higher grid voltages were somehow responsible for the quite harsh sound of the amp due to less plate current caused by too high voltage at the grid.
I think one of the tubes has drifted. When I first put the tubes in, the difference between them was about 2ma. I'm not sure at what difference the tubes are to be considered NOT matched. (that was the reason why I started the thread about twin bias in the first place, remember?)

Also I'm quite unsure about the ma bias setting for the tubes and if the weber calculator states the right setting for the specified voltage.

Think I'm gonna test it out some more and see how the amp (especially the voltages and current) reacts with different settings and therefore if the grid resistors are too small. Unfortunately I have the amp in the rehearsal room right now. Can't get to it until next wednesday .

[Robert M. Martinelli]

Hi Matt, Hi all!
Well, I won't worry too much about tube matching ( unless they're severely mismatched ). I'm saying this because you complain about the amp sounding "brittle" and "harsh".

Push-pulls tend to eliminate 2nd order harmonics, and this tendency is maximized when the output tubes are matched, this means that the more you worry about installing matched tubes, the more your amp will sound harsh and sterile, because odd harmonics ( "harsh" ) will prevail over even harmonics ( "sweet" ) and the harmonic distortion will be lower ( after all sometimes we DON'T want fidelity.... ).

Your idea of using a cathode bias is a good one IMHO, cathode bias tends to enhance even order harmonics, so it could sweeten the sound somehow.

You can also choose to add some "cathode bias sag" by not bypassing the cathodes, or bypass cathodes to keep gain constant and use a Tube Rectifier ( or a TR simulator ) to add some sag to the amp.

Another area to play with would be the NFB network, you could try to play with different amounts of NFB and add a cap to limit the HF response and sweeten the amp even more.

As to the bias amount, as a rule of thumb I bias amps a just a little over the minimum amount required to eliminate cross over distortion ( which gives a very unpleasant sound indeed ), but, as always, it's a matter of taste....some like it hot )

Hope this helps

Best regards

Bob

[Enzo]

The voltage at the plate is determined by the current through the plate load. Remove teh tube, and the voltage at the plate pin on the socket rises to the same voltage as the B+ supply. Why? Ohm's Law. There is zero current through the resistor, so there is zero voltage drop. Whatever voltage is on one end will also be on the other.

With the tube in place, the current through the tube passes through the plate resistor. The voltage drop across the tube will be determined by Ohm's Law as V = I x R. So is an example triode circuit, 1ma through the common 100k resistor will drop 100v. So the plate voltage will be 100v lower than the B+ at the other end of the resistor. I wouldn;t think in terms of electrons.

[txstrat]

What makes me wonder is that the grid are at a higher level than the plates.
I use a choke (90ma, 1.5 K) in the B+ line. Shouldn't the voltage drop there at least so much to keep the grids under plate potential?

[Enzo]

Figure it out. There is some DC resistance to the OT winding, the plate voltage will drop as current through that resistance rises. Measure the resistance and that current - ought to be close to what you find there with a volt meter. The plate voltage will thus be lower than the B+ voltage at the plate node.

The drop across the choke will relate to the current through it to the screens and everything else farther downstream.