Green Light for Exceeding Anode and G2 Voltage Limiting Values?

[pdf64]

Obviously, it's a long established practice for some AB amp designs to idle their anodes and screen grids above the limiting values noted on the manufacturer's info.
I've hypothesised that the anode voltage limit must be breakable provided that dissipation is below its limit. Because with inductive / transformer loading, the anode swings up to 2HT-V sat.
So it seems reasonable that there are sliding scales for anode and g2 voltages, such that if all metrics are simultaneously at their limit, as dissipation is reduced down from its limit, it is acceptable for the electrode voltage to be increased above its limit.
Anyway, after long searching for something to back my hypothesis up, I found this from a 1967 Mullard Technical Handbook


The chart is great, but the text has so many qualifications that it might be difficult for a real world application to fully comply.
What do others think?
https://frank.pocnet.net/other/Mullard/Mullard_TechnicalHandbookVolume1_ReceivingSqCrt_1967-09.pdf

[tubeswell]

That chart is a steady-state picture I’d say. Under dynamic conditions, increase in Plate voltage above the idle point is accompanied by a decrease in plate current, so Plate dissipation kinda evens out, and eventually drops away completely at the point where cut-off occurs in Class B (where the ‘on tube’ is pulling more current on the downward voltage swing)

[stratomaster]

I think the key caveat is that this only applies when the anode load is a resistor. In our case it's the OT with all its associated inductance plus what is reflected back by the speaker on the secondary. 

I don't think there's a textbook example or theory for the way we typically (ab)use tubes.  Don't under bias, carry spares, and prepare to be occasionally disappointed by unplanned failure.  In exchange we get compression, equalization, and saturation characteristics that 50 years of solid state and digital advancement has yet to fully replicate. 

[HotBluePlates]

Obviously, it's a long established practice for some AB amp designs to idle their anodes and screen grids above the limiting values noted on the manufacturer's info.
I've hypothesised that the anode voltage limit must be breakable provided that dissipation is below its limit. ...

"... provided that other limiting values are not exceeded ..."

While I acknowledge we might see both idle plate & screen voltages higher than rated maximums, the same tube likely doesn't simultaneously exceed rated maximums for plate and/or screen dissipation, or cathode current.

The RCA tube manuals point out that in Class B the maximum plate dissipation theoretically occurs at 63% of maximum drive signal, while in Class A it occurs at zero drive signal.

The thinking here is that in Class B (or deep Class AB) has rising plate & screen current at the same moments that plate & screen volts are falling (due to anode load and/or a series screen resistor) such that excessive voltages don't occur simultaneously with high current, constraining electrode dissipation.  For the plate, Power Input from the supply rises from zero drive signal to maximum drive signal, but the Power Output to the speaker subtracts from this and the remainder is Plate Dissipation in the tube.

Having done the tedious work of calculating average electrode dissipation from a large series of electrode voltages/currents over the signal cycle, I've noticed brief periods of very high dissipation, but these were brief enough that the average dissipation was not excessive, and that the moments of high-dissipation did not coincide with peak-voltage or peak-current.

All this is a long way to say there is some unknown amount of safety margin present in tube ratings.  It seems like tubes will tolerate rougher usage conditions than we might surmise if maximum cathode current is egregiously exceeded, and if multiple maximum ratings aren't exceeded simultaneously.