Low Power Tweed Twin Build: Sounds great but low bias

[JoeTele]

Hi,

I just finished a mojotone low powered tweed twin build. I did a careful power-up with a variac, starting with no tubes, taking all measurements, etc. The amp sounds great, is plenty loud, and has no noise or overheating. Voltages are a bit higher than the fender schematic but I understand that's normal. (Funny, they're very close when you variac down to 110v!) Thing is, I'm only getting 10 millivolts of bias current on the 6L6s. I incorporated bias test points via the 1ohm resistors, which I measured myself before installing them. The 10 millivolts is there on both tubes. I took the same meter and measured my super reverb, which is fine, so I think my measuring method is solid. I've got about -52 volts on the grids of the twin and about 438 on the plates. Anything wrong, or just very low rated tubes? They're tube amp doctor with a plate current of 23 and an amplification rating of 682.

Thanks!

Joe

[martin manning]

No idea what those TAD numbers mean without knowing the test conditions. By themselves they are only good for matching to other tubes tested at those same conditions. Based on your voltages, the set you have is low in current as compared to the data sheet specs. I'd just bring the bias voltage up until you get 35-40mA.

[JoeTele]

Thanks Martin!

This one does not have adjustable bias, but I may add it. In the mean time, will playing it under these conditions do any damage? The friend I built it for really likes the sound right now.


Joe

[martin manning]

I surely won't hurt it, but I'd guess it will sound much better with hotter bias. Right now you are at 0.01A*438V/30W = 15% of max dissipation. Most would say get it up to at least 50%. You can probably swap out a resistor to get up there; just need to see what schematic you are using.

Edit: Mojo's low-power Twin schematic shows 3k3 and 56k resistors in the bias circuit. If that's what you have, I'd try replacing the 56k with a 36k, or tack a 100k across the 56k to make 36k.

[MHProd]

Yes, what Martin said. My suggestion is bias them around 75%, a sweet spot for these tubes from my experience. And don't forget they are rated for 30w!

[JoeTele]

Thanks again!

Before I built the amp, I had my friend punch a hole in the chassis for an adjustable bias pot, just in case. I think in the not too distant future I'll just go ahead and add that in.

Joe

[Stevem]

With that General amount of plate voltage ( about a 10 volt range up or down ) you will have the outputs idling at 70% with a individual current draw of 49 ma, 60% with 42 ma, and a cool 50% with 35 ma.

[JoeTele]

Martin or others, could you just tell me which numbers you considered to arrive at the value of 36k for the resistance? I'd like to use this issue to finally get some amp math under my belt. If I'm understanding correctly, the 2 resistors in the bias circuit constitute a voltage divider, with the (stock) 56k or replacement 36k functioning as "R2" and sending more or less of the voltage to ground. I assume that if I knew what output bias voltage was sought, I could use a voltage divider calculator arrive at the 36k by plugging in input voltage, R1 (3.3k) and said output voltage. Question is, how did you determine that output voltage? Or is this the wrong path?

Thanks again!

Joe

[martin manning]

Well I kind of made an educated guess by looking at a 6L6 data sheet and estimating the increase in grid voltage needed to get the plate current up in the right range, and then figured the required change to the voltage divider. Have you made a change to the circuit? What happened to the grid voltage and plate current? Hows the tone now?

[JoeTele]

Amp is at my friend's place now but hope to modify this weekend. Will let you know.

Joe

[JoeTele]

OK, so I added adjustable bias and brought the amp up to 75% dissipation. Truth be told, neither my friend nor I felt we heard a much of a sonic difference but it's not like it sounded bad before. I did find it interesting that to get the 75%, I had to bring it up to 61 milliamps, which dropped the quiescent plate voltage down to 373 volts (the schematic reads 400, but we obviously know that today those values are typically low). The one thing that concerned me was the sweep of the adjustment pot. I followed the Weber schematic, which uses a 4.7k input resistor and replaces the stock 56k resistor to ground with a 50k linear pot grounded via a 10 k resistor. That pot is able to dump 150+ milliamps if you turn it to far. I guess the answer is not to set it too high

Newbie question: Would a different set of tubes potentially give a different current/plate voltage relationship or would the voltage at the first filter cap have to be higher for the tubes to reach a given percent dissipation at a higher plate voltage?

Thanks!

Joe

[martin manning]

For given plate and screen voltages, the grid bias voltage determines idle plate current, and different tubes of the same type will require slightly different bias voltages to reach the same plate current. Matched tubes show the same plate current at the same plate, screen, and grid voltages. As you dial up more current, the plate voltage will drop, but that is a characteristic of the power supply, not the tubes.

75% of max dissipation is kind of hot for a Fender, but 15% (where you started) is very cool, and likely to have significant crossover distortion a high volume. I would shoot for something like 65%.

For the adjustable bias, I think the 50k pot/10k resistor is not such a good plan. It has too much range and it's probably hard to adjust precisely. A smaller pot in combination with a larger fixed resistor would be a better solution.

[Reeltarded]

Actually, a schematic from the 50s that says 400v ends up looking more like 435vdc if the transformer isn't wound for 120v in.

As far as adjusting bias, that's quite a range you got there. I would have split 33k/25kl probably.

damnit Martin...

[martin manning]

damnit Martin...

What... ?

[Reeltarded]

You posted slighty more than me and maybe seconds earlier.