guitar speaker/tube challenge redux

[R.G. Keen]

>Yes--the Mesa/Boogie company puts fuses in the cathode circuit of the power 
>tubes in their big rackmount power amps (Strategy 500, etc.).  It's much safer 
>to fuse the cathodes than the plates, especially when the fuses are on the front
>panel.  It would probably be a good idea to fuse the cathode circuit of any amp,
>however.  And with a big tube power amp, fuses are pretty much essential to 
>prevent a really catastrophic meltdown.  Things can get out of hand real fast.
>From my earlier life designing power circuits, I know that specifying a fuse to 
protect electronics is tricky business.

A circuit I like is to put a small value resistor in the cathode line to 
monitor the current in the cathode of either each tube individually or all tubes
together. A comparator watches the voltage across this, and controls the state
of a power MOSFET which is also in the cathode circuit. Power MOSFETs with ratings
to over 1kV are available, and they typically have ratings in the several ampere
range and saturate to a few ohms. The comparator/resistor lets you set any current
level, and response time is the propagation delay through the circuits. I use a 
4013 to latch the devices off on a fault.

>FWIW, I use low-watt (1/2W) 10 ohm resistors, matched to 1%, in the cathode 
>circuit of my tube power amps as current-sense resistors.  This makes it really 
>easy to measure idle current when I'm setting the bias; 1 mA = 10 mV.  If 
>something really goes haywire, these resistors burn out and open the circuit. 
>Never had a problem so far.  But I think I'm going to start using fuses as well.
>(Years ago, Ampeg used this resistor-fuse trick in their mammoth SVT bass amp).
The fuse-resistor trick is OK as long as you're sure that faults will be massive
overloads. Modest overloads will just cook things without killing the resistor. 

Speaking of which... how do you replace it? The exact brand and production method 
matters to the burn-through point.