Blumlein garter circuit, any good?

[mac]

Pros and cons of Blumlein garter circuit @ tubecad?

http://www.tubecad.com/2005/May/blog0046.htm
http://www.tubecad.com/2009/04/blog0163.htm

Maybe it makes more sense to use independent and adjustable cathode resistors to match idle current.
The BGC, supposedly, self adjusts as tubes age. Adjustable resistors must be checked from time to time.

I ask because I found an old hifi tube amp in a street trash that is going to be a 50w 2xEL34 guitar amp.

mac

[merlinb]

Yes it's an excellent circuit if you can spare all the extra voltage drop across the extra cathode resistors. For example, I'm in the process of converting an old hi-fi amp from tube recto to silicon, which will increase the supply voltage by 30V or so. I therefore intend to use this extra voltage to use cross-coupled cathode bias. This means the output tubes still have basically the same voltage across them as they originally did. I've also contemplated a guitar amp using cross coupling so you can plug in multiple different types of power valve without needing bias adjustment (KT66, EL34, 6L6, 6V6).

(BTW, it has nothing to do with Alan Blumlein, so don't call it that. It's just cross-coupled cathode bias. Broskie got the wrong end of the stick somewhere)

[mac]

Can the extra voltage drop help to get Push Pull Class A?

http://www.aikenamps.com/index.php/is-the-vox-ac-30-really-class-a

mac

[PRR]

The garter scheme MUST work class-A push-pull.

So is a simple common cathode resistor.

> 50w 2xEL34 guitar amp

Class A is 50% efficient. EL34 is nominally a 25 Watt Pdiss tube. Two can dissipate 50 Watts. 50% of 50W is 25 Watts output.

Yes, there is enough curvature in the EL34 to get bias-shift from idle to full roar. And the 25W Pdiss rating always was conservative. 30W-35W/pair self-bias is possible without bias-shift so extreme it burps the music.

But for guitar?? Why? Keep the original PT OT and bias scheme. In fact I'd generally keep the whole power section intact, only loosening the NFB for more raw tube-tone.

"Class A" is not magic. The cite you give "proves"(?) the AC-30 gets well out of A even before it clips. And a true-A un-clipped amp is so mild that you may as well get a chip-amp. Many of the much-loved guitar amps get well out of A, even into deep-AB. (The 5F6-a's gain is lower at idle than at mid-power, over-biased, and this may be part of its tone.)

[merlinb]

The garter scheme MUST work class-A push-pull.

Is that true? A shared cathode resistor (fully bypassed with a cap) is not restricted to class A?

[PRR]

Take a modestly extreme case. Pick a cathode resistor to bias to a decent current. Pick a load which requires cathode current to double at full wail. Double cathode current is double cathode voltage. When the full-wail passage ends, the tubes will be severely over-biased, nearly cut-off. Post-crescendo sound will be weak/thin.

In hi-fi you can try a cap so large that it "does not change" over the period of a musical crescendo. Since hi-fi operation rarely gets into high-high current, this often works. (However it will be embarrassed by bench-testing at high power.)

That does not work so well for guitar which is often over-driven for many seconds at a time. A many-many second time constant would be OTOO 50,000uFd at 30V-50V, a very large lump.

With an un-bypassed cathode resistor the gain changes radically as the amp output shifts from A to AB.

[merlinb]

Take a modestly extreme case. Pick a cathode resistor to bias to a decent current. Pick a load which requires cathode current to double at full wail. Double cathode current is double cathode voltage. When the full-wail passage ends, the tubes will be severely over-biased, nearly cut-off. Post-crescendo sound will be weak/thin.

But countless class AB amps like the Vox AC have been using shared cathode bias for donkeys' years, and everybody loves them. The bias shifts become part and parcel of the 'touch sensitivity' of the amp.

[mac]

Isn't the left tube taking some opposite signal from the right one through the left grid to right cathode resistor?

BTW, what about the current mirror circuit on the same pages?

mac

[merlinb]

Isn't the left tube taking some opposite signal from the right one through the left grid to right cathode resistor?

No, because the cathodes are both fully bypassed with big capacitors.

BTW, what about the current mirror circuit on the same pages?

What about it?
Unlike the cross-coupled circuits, the current mirror will try to force one (slave) tube to idle at the same current as the other (master) tube. Whether that current is 'too high' or 'not enough' depends entirely on the master tube -which might be defective! Cross-coupling, however, averages out any defects in each tube.