These Diodes.???

[trem]

Just wondering what the diodes from cat to ground (parallel with the bias resistors) on the power tubes are for.?
Thank You
MHZ40C_Haze.pdf

[Jazz P Bass]

I don't recall ever seeing that one.
Something to do with raising the ground reference?

[Merlinb]

I have wondered about that too. As far as I can tell they do nothing, and may be a relic of an earlier development phase. I will try emailing the designer.

[trem]

You guys have WAY MORE knowledge than I do.
I am not sure what would have to occur to turn the diodes on......
Lots of current, and then they save 40 cents worth of resistors, but burn up 50 dollars worth of tubes.?
None of you guys said that, so I must not really know what happens there.
What if those resistors were not there and it was just the diodes.? The circuit would not function under "normal" conditions would it.?
Thank You

[trem]

I have wondered about that too. As far as I can tell they do nothing, and may be a relic of an earlier development phase. I will try emailing the designer.

I looked in a few books I have by some character named The Valve Wizard, but he did not mention them......

[Steve Conner]

My best guess is they're there to save the cathode resistors from burning out if the tube shorts.

Why would you want to do that? Well, if the tube shorted, the resistor could burn and fail open. If the resistor failed open, then B+ could appear on the bias test jack. If that were accessible from the rear panel, the CE folks would not be happy.

[trem]

My best guess is they're there to save the cathode resistors from burning out if the tube shorts.

Why would you want to do that? Well, if the tube shorted, the resistor could burn and fail open. If the resistor failed open, then B+ could appear on the bias test jack. If that were accessible from the rear panel, the CE folks would not be happy.

That is a question I have been meaning to ask.
If the cat has NO connection to ground, does the cat (referenced to ground) rise to plate potential somehow.?
Thank You

[R.G.]

They are indeed there as a relic'ed bandaid to "protect" the 1/4W cathode resistors. It lets them use 1/4W 1 ohm resistors for bias sensing, but not have those resistors turn into Light-Emitting- and Smoke-Emitting-Resistors on current pulses. Mr. Connor has the right idea.

In general, it's sloppy, MBA-oriented (or dictated) engineering. A wirewound resistor there could not be all that expensive.

I think the issue is: what happens when a tube shorts? Today's tubes are not the best ever. If you use a power resistor, the plate current is limited only by the cathode resistor in that case. A 1 ohm resistor limits the current to ... um, maybe 200A? Well, no, the power transformer winding resistance is bigger than that. But big, in any case. Frankly, it might be less exposure to the power transformer, the single most costly part in the amp, to leave off the diodes and let the 1 ohm biasing resistors overdissipate and open, a kind of oddly specified fuse. But as Steve notes, CE and UL, and TUV and so on would think poorly of that. So the diodes go in to protect the user from external voltages and thereby expose the power transformer if the AC power line fuse doesn't clear quickly enough.

If the cathode opens, the cathode does indeed run up to nearly the plate voltage. It's possible to do a standby by putting a high voltage MOSFET power device in the cathodes and turning that MOSFET off. And it's possible to put a 1 R resistor in the source of the MOSFET(s) and get both bias sensing and standby switching with a single low-voltage switch.

If I didn't put this in, and probably even if I do, the next reply would be to tell me that the tubes would be harmed by leaving the cathodes hot and letting them float up to the plate voltage. That doesn't make a lot of sense, as the hot cathodes would just be ... hot pieces of metal with little or no current flow. Monodes are not very electronically or electrochemically active, but you can patch around this by putting a 10K-100K or so resistor from cathode to ground. Not much current flow in a cathode like that, but it's still got a place for excess current to go.

I once did a cathode current monitoring device which looked at the voltage across a couple of 1ohm sampling resistors like this and then did a comparator threshold and logic to pick out maximum normal/overdrive current and open the MOSFETs when the accumulated current-time was judged to be too much for the tube. Worked.

[Steve Conner]

In a properly working tube, if the cathode is left open it'll rise to the grid voltage plus the tube's cutoff voltage. But in a faulty tube all bets are off. The cathode could be shorted to the plate by some stray piece of metal.

[Merlinb]

Answer direct from Marshall- now that's service!



"Hi Merlin,
Good to hear from you.

The diode circuit is not my idea, it on some modern Fenders.
There are multiple small reasons for the resistors which added up to make them worth the board space.

The basic reason for them is to conduct fault current. I could have gone higher power, but as the amp was to be manufactured in a remote factory, I wanted to control some parts more closely and I could do that my spec-ing 1% MF for these. That meant that they were somewhat less susceptible to heat than a power carbon or ww and less susceptible to pulse damage than a metal oxide. It also removes the high resistance failure condition from the resistor which could have lead to multiple visits to the tech (valve goes short – resistor burns a little and goes high resistance – fuse blows – new valve – high resistance resistor sinks more power and blows open or amp tech does not check value and biases incorrectly).

I suppose the most important thing though is that it is a circuit that will be regularly (in a world-wide sense) be called upon to carry fault current and I am happier with two components carrying it, one of which is a nice beefy diode, to pull the fuse as quickly as possible.

Putting any extra circuitry in for reliability can be a double edged sword. You might notice that the Haze has flyback diodes on the anodes. These are 3kV rated, but have still blown on some occasions and I have been questioned as to why I put them in the circuit as they ‘reduced reliability’. Obviously I pointed out that the diodes, for about 1p, had saved a much more expensive output transformer which tends to end the conversation pretty quickly! To date, I think only one 40 watt output transformer and no 15 watt transformers have ever had to be replaced."

[J M Fahey]

Nice to know.
That objective might have been better served by adding an extra 500mA fuse in series with each plate.
It would be a more definite circuit opener but would also have the much more important benefit of instantly removing high voltage from that tube making it impossible to have dangerous voltages at the cathode, shorted tube or not, open resistor or not, and by extension, at the bias test points.
Of course, finding real estate at the PCB, plus buying and actually mounting a fuse there, is much more expensive than just fitting an 1N400x diode.
Not criticizing them, they might even have skipped the diode.

[R.G.]

The diode circuit is not my idea...

Putting any extra circuitry in for reliability can be a double edged sword.

Now that's a two-part design answer I can respect.

[trem]

It must just be MY lack of an electronics education (I was in the painters union for 27 years, not college), but that guy's answer is more indicative of a politician than an amp tech.
Was ALL that in answer to "What are those diodes for.?"
I imagine it was in respect to the Man that penned the email question to him.....
Anyway....
So is he saying, in event of trouble, they aid in pulling more current, faster through the fuse to insure it snaps and saves the amp.?
Thank You

[Jazz P Bass]

The Marshall tech may very well be non English.
The syntax of the reply is awfull.

[g1]

C'mon, give the guy a break, I doubt many manufacturers would bother to even reply. As far as the english, I found it to be pretty much perfect aside from 2 typo's ("it on some modern Fenders" and "I could do that my spec-ing"). The mention of "manufactured in a remote factory" leads me to believe he is an english speaker.
Anyway, I think it may have been used because it was on some Fender's, and he is just back-thinking the reasons for it. Red knob "The Twin" and '94 Twin come to mind.
Recently I believe there was also another thread dealing with something similar (diodes or resistors?) across cathode fuses, and there was speculation that it was to prevent excess cathode to heater voltages in the case of failure.
I'm with Steve Conner on this though, protection for the resistors . When the original red knob "The Twin" came out, it didn't have the diodes. They started coming in for warranty with badly burned cathode resistors, along with damage to the bias board they were mounted on and sometimes the bias jacks/test points.
The diodes were a factory mod, seemingly to limit warranty claims due to extra labour involved in changing resistors or bias boards just because a power tube failed.