That is correct some push pull amplifiers can be badly damaged by a short circuit on the output. Many of the old PA amplifier had a B+ fuse that would work occasionally to prevent damage. Shorted outputs were good for electronic repair shops.
I've had some further thoughts on a protection board, one could use a relay to switch the mains input, a push to start buttons and push to stop buttons would replace the normal on/off switch, in the event of a power blackout the amp would net restart when the power returned. This type of switching is now mandatory on many power tools.
An almost non invasive PCB could do a lot for protection with an octopus like structure of wires going out from the micro PCB to key nodes that require monitoring, on fault the mains relay would disconnect the power. When monitoring areas that have audio on them a suitable resistor often 10 megohms can be in series with the detect wire, the resistor and the wire end can be sleeved with heatshrink tubing.
There is an image of a latching relay circuit here. A simple over temperature "stop" circuit could be a normally closed thermal switch (mains rated) in series with the NC stop switch.
https://forum.digikey.com/t/push-on-push-off-control-example-circuit/633
kk
A 'shorted' OPT output could have a variety of causes, such as a short within the secondary winding, to a short in a speaker coil, to something in-between.
The short would likely change the loadline to a more 'vertical' line from the idle point, with each PP valve stressed alternately (and depending on signal level) to a current limited by the grid-drive voltage and anode curves, and with the power supply possibly sagging. A tube's dissipation may not increase much until say a square waveform forced the max current to more than twice the idle current. So the risk to the tubes and OPT (assuming the fault wasn't within the OPT) could increase within seconds depending on how much over-drive occurred and how robust the OPT primaries and the valves were to red-plating type levels.
Also a P-P primary could have a low inductance, due to the shorted secondary (ie. towards the leakage inductance), and if the amp used GNFB then without feedback the forward gain could be somewhat higher.
Given the need for a significant drive signal to cause substantial plate dissipation, some applications may notice a shorted secondary quite quickly.
Wrt a uP monitor type solution, it would seem that ADC type inputs would be more likely to avoid any vagueness of transferring output voltage and current signals into analog levels that were then sampled and tested across a range of fault conditions. However that likely moves the software into much more rarefied air, and as indicated, is more about avoiding long duration stress than initial transient stress.
The short would likely change the loadline to a more 'vertical' line from the idle point, with each PP valve stressed alternately (and depending on signal level) to a current limited by the grid-drive voltage and anode curves, and with the power supply possibly sagging. A tube's dissipation may not increase much until say a square waveform forced the max current to more than twice the idle current. So the risk to the tubes and OPT (assuming the fault wasn't within the OPT) could increase within seconds depending on how much over-drive occurred and how robust the OPT primaries and the valves were to red-plating type levels.
Also a P-P primary could have a low inductance, due to the shorted secondary (ie. towards the leakage inductance), and if the amp used GNFB then without feedback the forward gain could be somewhat higher.
Given the need for a significant drive signal to cause substantial plate dissipation, some applications may notice a shorted secondary quite quickly.
Wrt a uP monitor type solution, it would seem that ADC type inputs would be more likely to avoid any vagueness of transferring output voltage and current signals into analog levels that were then sampled and tested across a range of fault conditions. However that likely moves the software into much more rarefied air, and as indicated, is more about avoiding long duration stress than initial transient stress.
