Bias issues with new 6G2 build... WAY low current draw from power tubes - huh?

[Britt Rossman]

I am working on an amp that is a (not quite) conversion of an old Conn Organ amp. We have reused the chassis and iron, while essentially retrofitting all guts (stripped and rebuilt). The base schematic is a Fender 6G2 Princeton, with some tonestack mods, and substituting 5881's for the 6v6's of the 6G2 to get a bit more power/headroom...

http://music-electronics-forum.com/a...&thumb=1&stc=1

At this point, I'm getting a good bit of sound, but I thought it seemed way quiet for what it should be... my B+ at the plates is ~455 and Bias voltage is sitting around -47v, but a 10R resistor to connect the cathodes to ground gives me 0.22v, which equates to ~2.2mA of current... How am I at 20x lower than where I need to be?

I came across this thread here about a similar kind of issue... http://music-electronics-forum.com/t37497/

5U4GB recto is good... I checked my heaters (as was the issue in the linked thread), and I've got 6.7vAC, which is high but still reasonable. Is there someplace I'm not looking that could be the culprit for not having enough current draw? Is the OPT dead? I've noticed that I can feel some vibration coming from the PT and OPT... Boy, I hope they aren't fried...

[Enzo]

Is there someplace I'm not looking that could be the culprit for not having enough current draw?

perhaps in your calculator. I = V/R = 0.22/10 = 0.022a = 22ma, not 2.2ma.


Why on earth do we suspect all the transformers?

Knock that bias back a bit if you want it running hotter than 22.

other folks here know more about it that I, but I suspect the change you made to the tone stack is sucking the volume out of it. Looks to me like it is a good 10db down compared to the standard Fender stack with controls centered. And the Princeton didn't even have a whole stack.

Just for science, try connecting the volume control wiper directly to the next grid and disconnect the tone circuit entirely, see if that restores the missing volume. If so, we can work in that, if not, we then look elsewhere.

[Britt Rossman]

No, the error was with my brain... long days, late nights, and no breaks. Anyway, that still gives me ~half of my expected current and only 10W of dissipation.

Lemme see if I have this... I wanna "knock down the bias". I am looking for lowering that voltage (to like... -60V) to drive up the current draw from the cathodes? OR I want to raise the voltage (say to -35V) to make the grids more positive, which would... draw more cathode current...? This is the part that confuses me...

So to do this, I will change the 100k resistor in the bias arrangement, yes?

[Randall]

It can be confusing. What Enzo means by "knock that bias back a bit" is to apply less negative voltage to the grids. Higher bias voltage results in lower cathode current. Lower bias voltage means higher cathode current. Think of the grid as a screen, with electrons from the cathode flowing through it. Higher bias voltage makes the screen holes smaller, resulting in less electrons getting through. Lower bias voltage has the opposite effect.

Perhaps the bias voltage is higher due to modern line voltages? I'd try a 120K ohm in the bias supply to see where that put things. Also, this amp makes tremolo by modulating the bias voltage, so have you made sure the trem is off, and intensity pot is all the way down?

[The Dude]

The more negative the bias, the colder the tubes run (less current). That is why if you lose bias, the tubes red plate.

<simulpost>

Edit: To avoid confusion, I like to use "more negative" or "more positive". In the case of bias, which is negative, it can be confusing to use "higher bias voltage". To some this means a bigger number. To others it means closer to 0V. Not nitpicking, just trying to make it easier to understand.

[g1]

And by the way, what is your screen voltage at pin4 of the power tubes?
Low screen voltage will reduce the idle current.

[Britt Rossman]

@Randall - Currently, I have bypassed the trem controls (it is not yet wired at all) such that the bias tap goes directly to the pre-grid voltage divider. I figured I would ensure proper bias (center point) before making it oscillate...

@g1 - I *strangely* seem to have a plate voltage that is almost always 3vDC lower than my screen... That is not right... Currently, the amp has some 6L6GB's which are rated at 400v MAX for plate and screen, but I have a pair of GC's coming. I'd like to be able to use these tubes since they are a matched set of (damn near) NOS with matching date codes to each other and the rectifier, which are all CONN-labeled OEM parts. Given yellow stamps, I'd guess re-branded Sylvania?

While we're at it, it seems that according to my calculations, I should have ~15W as it stands right now and ~30W after I properly adjust the bias. Regardless, I *feel* what I hear is only about half of what I expect... Could I have reversed the leads on my OPT? It is old enough to have followed the old conventions: SEC has black yellow and green leads; current hook up is to Green into an 8 ohm load... I suspect that I may have the 4 ohm tap hooked up, thus losing ~half of my power that way... Also may explain some of the heat from the OPT when I played it for ~10 minutes.

[Britt Rossman]

I changed the bias resistor to a 120k and got the following readings:

Mains: 124 VAC
Plate: 443 VDC
Screen: 448 VDC
Grids: -43 VDC
Cath, L: 40.7 mA
Cath, R: 42.7 mA

So, Power out of left side should be 17W, and right side should be 18W, thus ~35W before expected losses from some less-well-matched-than-I-thought tubies... I'm quite confident in these tubes' ability to take some abuse at those settings, but I don't have enough experience to feel comfortable with that going out the door just yet... I expect the pair of JJ-GC's I have on the way should fare better at those kinds of voltages.

Still getting low output for what I'd imagine from 35W... I have a 12AT7 in the first slot, which may have something to do with it, but I wanted the amp to be a bit cleaner... that may explain a bit as well (@Enzo). I'll try rolling...

[The Dude]

Just an IMO, but that's way more idle current than I would like to see. If it works for you,...........

[Justin Thomas]

Just so you know, PLATE dissipation is NOT the same as AUDIO power output. IOW, that 35W you're getting is how much heat the power tubes dissipate just sitting there. Depending on the actual circuit and how you have it biased, two 5881s could theoretically put out anywhere from 13.3W up to 47W of SIGNAL power. But, the plates can't dissipate any more than 23W each. The circuit itself from input to speaker has more to do with the signal power than just the idle dissipation that we measure. I've built amps where the power tubes are throwing as much heat as possible yet still sound weak due to bad preamp design...

That idle plate dissipation is a safety check for our amp - so it does'nt melt down. But to measure the actual output power takes a signal generator and a scope. There are a HOST of simpler solutions to your low volume - wrong/mislabeled resistor, wrong value pot, weak preamp tubes, bad speaker, etc...

Justin

Edit: agree with The Dude. 18W from a tube rated for 22W is 80+% dissipation. Not enough margin of error for me, either.

[Steve A.]

Did you already try what Enzo suggested? Tone control circuits can soak up a lot of dB which is why Fender kept theirs simple in the small tweed amps. (You could use an EQ pedal if you need to shape the tone.)

"Just for science, try connecting the volume control wiper directly to the next grid and disconnect the tone circuit entirely, see if that restores the missing volume. If so, we can work in that, if not, we then look elsewhere." Enzo

Steve Ahola

[Tom Phillips]

You have posted various voltage measurements but you have not actually measured the amp's output power. Do you know how to do that? Do you have other test equipment besides your meter. I.e. Signal generator, scope, dummy load?

[Britt Rossman]

Thanks man. I agree, but I didn't factor that, honestly. I think I'll reconsider the bias point. Working through it, let's shoot for 15W... two 1/2W 220k resistors would theoretically give a value in between the two I've used (~22mA w/ 100k, and ~42mA with 120k) to give ~32mA with 110k for ~14W giving 65% of rated max at idle... sounds like a sweet spot to last for years.

@Enzo and Steve - Maybe I'll add in a tone bypass switch? I honestly don't see how I could have such considerable insertion losses... a BMT stack supposedly drops a crapton by comparison. I understand that one control is probably the best, just trying to work a little versatility into the amp. I also have to add in NFB, which should tighten that "squishy as a February jack'o'lantern" bass (in this vein, I'll try a cathode bypass cap too)... Just trying to take things one step at a time and not blow things up. It is a learning experience, so I'm trying to get the most out of it...

[Britt Rossman]

@Tom P. - I have a generator that may or may not work. I have a couple unreliable scopes. No dummy loads. Honestly, I didnt think it'd be this tough...

To answer your question, no I do not know how to do that. Hazarding a guess, I'd figure I could read the voltage drop across and current flow through a dummy load then calculate I*V=P ....? I'd also guess the scope is to monitor the waveform for noise and distortion..?

[g1]

It sounds to me like you are trying to adjust the bias by altering the grid leak resistors. This is not the way to do it, you are shunting part of your drive signal to ground. Stick with the 220K's there.
To alter the bias, adjust the 100K and 30K on either side of the bias diode (shown above AC power switch on schematic).