Cathode biased PP 6V6 max plate dissipation.

[alchemy]

I have searched around a bunch and could not find a definitive answer. Is it normal for cathode biased push pull amps to be biased closer to the max dissipation of the tube (like 90-100%) than a fixed bias counterpart? I keep running into older amps that are running way hotter than I am used to.

[scole]

yes, cathode bias will be at almost max dissipation at idle, it will cool down when played. fixed bias is set to about 65-75 percent at idle. check out this vid
YouTube - Bias Part 2 - Steve Carr Gets Technical

[TransLucid]

I think you can set the idle current anywhere you want, whether cathode or grid biased. The same tradeoffs will be present (tube life, maximum output power, distortion characteristics, etc).

[pdf64]

'I think you can set the idle current anywhere you want, whether cathode or grid biased. The same tradeoffs will be present (tube life, maximum output power, distortion characteristics, etc). '
No, the tradeoffs are rather different, as scole describes.
Try increasing the value of the cathode resistor - by 50%, 100%.
As it goes up you'll see that the output power decreases, due to the cathode voltage taking an increasing share of available B+, and a sine wave gets strangely mis-shaped well before clipping, due to the bias increasing with signal level.
So cathode biased amps are best biased so that there's not much increase in avarage current between 'no signal' and 'full signal', which generally results in the output tubes being run close to max plate dissipation.

[TransLucid]

If the cathode resistor is bypassed the AC current does not (in theory at least) affect the bias voltage. I would suggest that the distortion is due to the load line being in a non-linear part of the characteristic curves. This will happen no matter what the biasing method.

Your point about the output power decreasing due to lower plate-cathode voltage is a good one. The voltage across the tube is used to determine the load line, not the voltage from plate to ground.

I guess what I'm trying to say is that in a proper design, supply voltage, output transformer primary impedance, and cathode resistance/bypassing must all be considered together. Varying one without thinking about the other two can result in a poorly-placed load line and resulting distortion.

[pdf64]

For resistively cathode biased class B, AB amps, the bias (ie cathode) voltage has to respond to increasing average signal level by increasing (that's inherent in non-class A amps I think?).
To avoid the dreaded amp class wars, can we limit this to non-clipping amps? Pete.

[TransLucid]

Pete -

Here's how I see it - in a non-class A output stage, the total plate current varies with signal (because only half the waveform is being fully amplified - the other half is clipped). This current flows to the cathode. With cathode biasing, the DC component (the quiescent current) flows through the resistor, resulting in the bias voltage. The AC signal current flows through the bypass capacitor - it does not affect the DC voltage at the cathode.

This is different from a class A stage, where the total (DC + signal) average current remains fixed, because the average AC current is zero.

Bob

[pdf64]

'the total plate current varies with signal '
Yes, I see what you mean, the ac at the cathodes gets cancelled / bypassed to ground.
But it's the grid's effect on the plate current, eg as average signal level increases, the average plate current increases, that causes the cathode voltage to increase, thereby changing the effective bias.

[TransLucid]

the total plate current increases, as does the total cathode current. However, the cathode current is divided into two paths, one through the resistor (the DC component) and one through the capacitor (the AC component). The AC component is the one that increases, so the current through the capacitor increases. The DC current through the resistor remains the same, as does the average cathode to grid voltage (the bias voltage).

[pdf64]

It doesn't work like that, I'm not sure where your operational rationale is wrong, but with a few moments on the test bench you can check it for yourself.
Or see ac30 schematic
The EL84 cathode voltage is noted as being '12.5v at 30 watts, quiescent 10v'.
http://www.webphix.com/schematic%20h...s/ac301960.pdf