I just posted a schematic for active protection of a tube amp,
and B+ turn on delay at ABSE.

At turn of the amp, the DC heater voltage from a silicon diode rectified

supply is quickly established.
This is applied to the circuit to give a rail at +12v.
The R2 and C2 form a time constant circuit with a slow rise in C2
voltage.
It takes 25 seconds for the C2 voltage to rise to about +10v, when the
ZD
begins to conduct current to the high impedance base input of Q2,
connected to Q1
to form a darlington pair.
After 25 seconds after turn on the heaters in the amp have all heated
the cathodes
to be able to conduct current.
When Q1/Q2 turn on, the relay closes, thus allowing the
HT winding to begin delivering power to the power capacitors, which
should be large as possible, to give a slow rise to full B+ potential.

If the amp is turned off for longer than 0.5 secs approx, the heater
voltage will fall very quickly,
and the 12v rail will fall.
D8 allows the fast discharge of C2, so that if the amp is turned on when
still hot, the
delay circuit is forced to re-commence a 25 second delay to the B+ turn
on.


There is active protection against excessive cathode current in
one or all output tubes.
If the normal idle voltage at any cathode is +20v,
and it rises to 27v, the idle current x B+ voltage will
make the tube dissipation excessive, but will not cause a fuse to blow.
Only the direct voltage changes are registered at the tops of C3-C6.
Adjustments for different amps may be made with variations to
R11 to R14.


So if the dissipation is 25 watts per tube with Ek = 20v,
then if Ek rises to 27v, the dissipation would be 33.75 watts,
which would be excessive in the case of a 6CA7 or EL34.

When the voltage at Q3 base rises to +6v, current flows through Q3
D3,4,5,
and over comes the forward voltage of the diodes of about 4.5v.
The voltage at R6 rises to +0.72v, latching on the SCR.
This turns on the red LED, D2, indicating a fault with the amp.

With the SCR turned on, the voltage in C2 is drained down through R3 and
D3,
thus cutting off current in the base of Q2, since ZD does not conduct
below its threshold voltage.
The relay is then allowed to spring open, and the HT is interupted, and
the B+
falls to 0V, but the amp is still turned on, but with only the heaters
on.

Its stays like this until the owner attends to the amp, and works out
what problem caused
excessive Ek.
Was it a shorted speaker lead?
Too much volume?

He can try to reset the amp by turning it off, then back on again
after a couple of seconds, and the SCR then unlatches, and the whole
delay
process restarts.
If the Ek rises yet again and turns off the HT, he knows he has a
problem
and he takes the amp to Mr Fixitkwik who will find out what's the cause
of the problem,
but who will not be faced with an amp which has melted down due to tubes
being allowed
over conduct for any length of time, possibly wrecking a precious OPT.

The rail voltage of the circuit isn't critical. I generally use a 6.3AV
winding in a doubler circuit
to make +15v, then C1 and R15 are not required, since the current in the
12v rated relay coil
won't be excessive if the rail is up to 16v.

Q2 and Q3 are common varieties of NPN signal bjts worth 10c each,
so almost anything will do.
Q1 is BD139, which will cope well with the current via the relay coil.

The SCR is a sensitive gate type that draws a tiny gate current, but
still requires
the gate voltage to rise to +0.7v to latch on the SCR.

The D3 to D5 could be replaced with a zener diode but I choes leds,
because
they are cheap, and give a good threshold effect for when voltage
exceeds about
1.4v in each.

The circuit only reacts to DC changes to the cathode voltages.
I have used the same circuit to monitor DV at cathodes where there is a
large cathode feedback signal
present. The C3-6 220 uF filter out any signal voltages, and the R11 to
14 are such high values as to not
waste any power from an active cathode circuit.

Patrick Turner.