OT protection??

[mook]

if you install a fuse (.5 amp?) between cathode and ground of the power tubes (KT88, 6V6, etc.), will it protect the OT if the fuse blows?

Mook

[pdf64]

Where practicable, I put a quick blow fuse in series with the cathode of each power tube, following a tube short that damaged an OT primary a couple of years ago.
I got the idea from a marshall schematic, it also allows for a blown fuse led to illuminate if put across the fuse (with a series resistor).
315mA has worked for me so far, for 6L6GC and EL34, with no spurious blown fuses.

[R.G.]

if you install a fuse (.5 amp?) between cathode and ground of the power tubes (KT88, 6V6, etc.), will it protect the OT if the fuse blows?

Maybe. Yes for some faults, not for others.

OTs die from a few different scenarios. These are the biggies:

1. Voltage punch through by transient spikes.
2. Overcurrent on the output tubes
3. General overheating.

In all cases, what actually kills the OT is insulation failure between the turns inside the windings. The differences above only vary in how the insulation fails.

You're protecting from item 2 with a fuse on the output tubes. Many amps have a B+ fuse. This does the same thing.

I once designed an electronic monitoring circuit which looked at cathode current for the output tubes and turned off the tube when it was bigger than X for longer than Y. Worked fine. Didn't have to replace fuses.

Running without a load is the most common way to get 1. Fuses won't help you here.

[pdf64]

The T0.5A fuse that was in the B+ happily carried on supplying current for that tube short. The tube's 1 ohm 0.5 watt cathode resistor eventually popped, but by that time the OT primary insulation had been damaged.
Perhaps I had a bad one, but slow blow (T) fuses can tolerate non-catastrophic overcurrents for longer than I'm happy with.
Maybe I was just unlucky, but as tube shorts aren't unknown (!), I figured that it would be beneficial to increase the mitigation against another one killing an amp that passed through my hands.

[Hardtailed]

The T0.5A fuse that was in the B+ happily carried on supplying current for that tube short. The tube's 1 ohm 0.5 watt cathode resistor eventually popped, but by that time the OT primary insulation had been damaged.
Perhaps I had a bad one, but slow blow (T) fuses can tolerate non-catastrophic overcurrents for longer than I'm happy with.
Maybe I was just unlucky, but as tube shorts aren't unknown (!), I figured that it would be beneficial to increase the mitigation against another one killing an amp that passed through my hands.

Shouldn't the B+ fuse a quick blow?
That's what I always use! For that precise reason.

I have yet to see one blow for no reason.

[pdf64]

Quick blow fuse in the B+ - could you double check?
How can a quick blow fuse cope with the switch on current surge from charging the power supply caps up?
The only way I can think of that might be feasible would be with a slow warm up rectifier like GZ34, and no standby switch.
What value are you using, and where is it positioned in the circuit?

[R.G.]

Maybe I was just unlucky, but as tube shorts aren't unknown (!), I figured that it would be beneficial to increase the mitigation against another one killing an amp that passed through my hands.

This is why I came up with the circuit I mentioned. I distrust fuses.

A comparator, some resistors and caps, and a big MOSFET used as a switch can do a much more predictable job of cutting off the juice when something goes nuts. It's one of the few techniques I know of that will save the tube(s) if you suddenly lose the bias supply, perhaps from a bias cap short.

Note for purists: No, you can't hear the MOSFET. It's used as a switch. It's either off, and the amp is silent, or it's on and looks like a sub-one-ohm resistor.

[Alf]

Seems like a very useful circuit ! Do you have the schem somewhere on your site R.G. ?

I would be much obliged to have a closer look at it !

Alf

[Hardtailed]

Quick blow fuse in the B+ - could you double check?
How can a quick blow fuse cope with the switch on current surge from charging the power supply caps up?
The only way I can think of that might be feasible would be with a slow warm up rectifier like GZ34, and no standby switch.
What value are you using, and where is it positioned in the circuit?

I can guarantee that's what I use. My main amp is a 50W Marshall type (2XEL34), SS rectifying, the 500mA quick blo never failed me.

I do not know if this has an impact, but I have the first filter cap before the Standby switch, therefore when you turn it on, only this cap will draw current, which means that when you switch from Standby, this one is already charged (and the rest is behind a choke and decoupling resistors, so they don't draw as much)

[pdf64]

Thanks for checking. I can't understand that. The initial switch on current surge should blow it.
I'll go and take some measurements to try and get my head around this!

[R.G.]

Seems like a very useful circuit ! Do you have the schem somewhere on your site R.G. ? I would be much obliged to have a closer look at it !

I don't have it up on my site. It was developed about ten years ago, and I intended to commercialize it. Life got in the way.

[pdf64]

I'd also be interested in seeing how to implement mosfet based protection, if you wouldn't mind putting it into the public domain (for non commercial use).
The amp on the bench at the moment is a Traynor YB1A, full wave bridge rectifier, 6L6 tubes, 100uF reservoir cap on the rectifier.
I've got the fuse and standby switch between the B+ secondary winding and rectifier, so as to protect the transformer from diode/cap/tube shorts.
After 1min warm up, switching standby to operate, fluke set to 'max' detect across the empty fuse socket, gives a switch on surge current of 1.07A.
No signal is about 0.13A, max sine is 0.49, max square is 0.68!
An anti surge 0.5A ceramic B+ fuse seems fine.
I was considering monitoring the current between the reservoir cap and rest of amp (as per Hardtailed's B+ fusing arrangement), but those eyelets in the Traynor were really tight, and once I'd got the wires and components in, I didn't care to remove some for such a test.
I'll repost with those results from another amp.
As I understand the theory, the dc should be 0.7071 x ac on a full wave bridge, which puts the 0.68A max continuous ac as being just on 0.5A dc, and without the 100uF reservoir cap to charge, the standby switch on surge should be much less.
So perhaps fusing the B+ on the dc side with a quick blow would be better, accepting the loss of protection from cap/diode shorts that fusing the winding provides.

[daz]

Sorry if this is a stupid question, but could you use a circuit breaker for any of these applications?

[R.G.]

It developed out of looking into a MOSFET based standby switch.

One way to do standby without floating it on B+ is to interrupt the low side, the cathode to ground. There are some hard-switched circuits like that.

So you put a high voltage MOSFET in the cathode-to-ground. Now you can make the actual standby switch be trivially small, because all it does is switch the MOSFET gate between ground (no current flows) and a zener-derived low voltage of less than 20V. A 12V to 15V zener is ideal. Now you can standby by using a low voltage/low current toggle switch if you like, and there is never more than your zener voltage on the wires to the standby switch.

You can still use a current-sampling resistor on the tubes, by connecting the source of the MOSFET through a 1 ohm resistor to ground. Any current in the resistor makes a voltage trivial compared to the gate switching voltage, so you still have solid control of the MOSFET switching at the gate. And now you can use the 1 ohm resistor's voltage to generate a signal that's proportional to the instantaneous current. Or you can low pass filter that with simple RC filters and generate a long term average of the current with any time constant you want.

If you do that, a simple comparator can detect when the voltage goes over some reference, and trigger a flipflop to un-set the voltage to the MOSFET gate, turning it off. The MOSFET will have zero trouble turning off the tube current. Most high voltage MOSFETs in TO-220 packages will hardly notice the tube currents.

98% of the complexity in doing this is figuring out the filtering time constants and trip levels to prevent undesired triggering. It's not false triggering, just unwanted, because we all expect tube amps to take huge transients without whimpering. The trick is to sort out the huge transients from the deadly ones, and still keep the player happy.

Another 1% of the complexity is setting up the turn-on/off behavior and deciding whether the thing unlatches on its own after some cooling-off delay or if the player has to do something like turning power off/on or pressing a reset button.

I'll see if I can draw something up. I may even have the old PCB layouts. I do a good archiving of techie stuff.

[Alf]

I'll see if I can draw something up. I may even have the old PCB layouts. I do a good archiving of techie stuff.

That would be great !!

tia, Alf