OT protection

[lowell]

Just had an idea stemmed from my last thread. It may be a great way to protect an OT if someone has nothing plugged into the speaker jack and/or is plugged into the wrong jack. Use a stereo speaker jack. Ground the power tube cathodes via the stereo lug. I'm aware that this will not protect against a bad tube or faulty speaker cable, or faulty speaker. However it is one step towards protecting the OT is it not?

[black_labb]

while it will work, it would be difficult to keep a proper grounding scheme using this method.

what has been mentioned by someone before was to use gas discharge tubes between the plates on a pp design. works well if you can get the right value

[JoeM]

Rivera amps use a similar scheme, but it still wont protect if the other end of the speaker cable isnt plugged in.

[R.G.]

Protect an OT?

OTs die two ways. One is from overtemp softening the insulation and letting two windings short. The other is from high voltage punctures.

The subset of overtemp itself happens two ways: just too much current going through and overheating the wires, and by excessive high frequencies running the core dissipation through the roof by eddy current losses.

Case 1 (too much current) can't happen in otherwise normal operation because the output tubes won't do it. They themselves won't provide enough current in normal operation. But they can if they short. So a fuse in the B+ prevents overheating by tube failure. Or you can use something like the Tube-Saver (r) that I designed to monitor tube current with 1 ohm resistors in the cathodes and shut down the tubes with a power MOSFET switch in the cathodes if it gets out of hand. Fuses are easy but MOSFETs do what you tell them to. Take your pick. 8-)

Case 2 (excessive high frequencies) happens when your amp oscillates with too high a load. A high resistance, like maybe 240R, 2W, on the output jack will usually stop that. Several historic amps do this or have been modded to. You have to test the amp to make sure it does not oscillate with a 240R load. You can also put a passive highpass filter on the output and only let through stuff over maybe 100kHz. That's beyond anyone's ability to hear, and it for sure should not be there. So you look for it and shut down the power to the output stage if you find it.

Insulation punctures is what happens when the current in the primary is suddenly interrupted. The current through the transformer's leakage inductance gives you a flyback kick that will go to literally any voltage at all to keep the current flowing. This is what those strings of diodes from ground to primary are intended to prevent. They clamp the transient voltage at each plate to be less than the punchthrough voltage of the wire insulation by reverse breaking over. Sadly, they don't do it all that well, although something - anything! - is better than nothing for this problem.

I used AC powerline protection MOVs in the Workhorse line of amps. These are strung from plate to plate across the OT, and are sized so that the normal peak voltage from plate to plate, which is two times the B +, can't set off the MOVs, plus a little bit to remove nuisance trips. The MOVs cost about $0.50 each, and will happily eat any transients at voltages much closer to normal operation than betting on the reverse voltage of a diode which is rated for a minimum - not maximum! - breakover voltage.

And you can leave your grounding scheme where it should be.

[defaced]

MOV = Metal Oxide Varistor?

[R.G.]

MOV = Metal Oxide Varistor?

Yeah - the modern, snap-action ones, intended for protection against transient voltages on the AC power line and other places. You have to order high-voltage rated ones to withstand 2x B+. I used two in series to get the necessary 800-900V rating.

[Steve Conner]

Oh, that's interesting RG! Did the MOVs have any effect on the tone?

I ask because when I've played with MOVs, they always seemed to have a fair amount of capacitance. I'm thinking of one batch in particular, that I got in a lucky dip at a hamfest, and actually spent several months thinking they were ceramic disc caps. They measured 25nF, if I remember right. I got some really weird results in my circuits before I realised what they actually were, 75 volt MOVs.

Silicon devices like MOVs are available now, called TVS or Transzorbs. These might make good "upgrades" for the little stacks of OT protection diodes found in most modern amps. Again they have a fair bit of capacitance, though this time it's non-linear and any significant reverse voltage reduces it greatly. I've been bitten using them to protect MOSFET gates: they had more capacitance than the gate!

Your protection network can go plate-to-plate, or one from each plate to ground. The required clamping voltage is as RG says in both cases: 2x B+ plus a bit. Plate-to-plate won't short your B+ to ground if the clamp fails short. But neither will it clamp common-mode transients where both plate voltages shoot off in the same direction at once.

I put a 220 ohm resistor on the speaker output of every amp I build or restore, and test them for stability with an open speaker.

[defaced]

and test them for stability with an open speaker

As in not mounted in a closed back cab, or do you just mean a speaker as opposed to a resistor?

[Steve Conner]

I mean open as in no load at all, the speaker wires dangling in the breeze.

I started doing it because I was interested in investigating the myths behind tube amps. Every amp I've built has been quite happy in this condition. They don't oscillate (if they did, I'd fix it) and the NFB keeps the output under control.

With no drive or low volume, that is. The damage starts when you really crank the amp without a load. Either the screen grids in the tubes melt, or the screen resistors blow, or those protection diodes fail short, or the OT breaks down and turns into a doorstop. The 220 ohm resistor and diodes protect the OT, but they won't save your tubes or screen resistors.

[R.G.]

Oh, that's interesting RG! Did the MOVs have any effect on the tone?

They did, but not a lot. I didn't test the actual capacitance on the ones we used, but I did do A/B tests with them in and out. On most actual guitar sounds there was no difference I could hear, but it did drop the measured response a bit up in the top octave. Exactly where in the top octave varied from device to device and on output level. That's not too surprising since at zero signal, there is zero signal across them. It's only when you swing the full signal range that the quirks become more apparent.

I sometimes chewed on myself for deciding to go ahead with them. I could not make a Workhorse oscillate with no speaker load either, but I figured that for the $1 in product cost, it was worth protecting the very few amps which would ever oscillate under edge conditions. Then too, the Workhorse amps do not use feedback. They are full-on tube-y, nothing to hide the tube sound.

Every amp I've built has been quite happy in this condition. They don't oscillate (if they did, I'd fix it) and the NFB keeps the output under control.

I have seen a few oscillate, primarily the metalhead gain monsters. But I did not catch them in act of dying. And I've seen a few which were reported as dying suddenly when the speaker cord was pulled.

With no drive or low volume, that is. The damage starts when you really crank the amp without a load. Either the screen grids in the tubes melt, or the screen resistors blow, or those protection diodes fail short, or the OT breaks down and turns into a doorstop. The 220 ohm resistor and diodes protect the OT, but they won't save your tubes or screen resistors.

We seem to have studied in the same school.

People forget that the PT and OT, in that order, are the single most expensive parts in a tube amp, possibly excepting the speaker. If you gotta kill a tube to protect an OT, it's probably a good tradeoff to do so.

That's why I liked that tube saver thing. It would interrupt even a failed-shorted tube's current. Cost a MOSFET and a couple of analog ICs, though.

[hasserl]

I put a 220 ohm resistor on the speaker output of every amp I build or restore, and test them for stability with an open speaker.

What wattage rating? I've seen this posted many times before, but I've not done it myself. I'm finishing up a new build right now, this may be something to incorporate.

[Steve Conner]

I think 5 watts or thereabouts, a decent sized wirewound thing. Maybe RG has a way of calculating the wattage, but I just copied what was in my old Selmer. I think that was 500 ohms, but it was across a 16 ohm speaker, so I figured 220 for an 8 ohm tap.

I sometimes chewed on myself for deciding to go ahead with them.

Well, if they didn't make any audible difference in A/B tests and they didn't catch fire when you abused the amp with a Fuzz Face, which is probably as bad as any oscillations... What's to worry about

We seem to have studied in the same school.

The University of Magic Smoke (motto: Garde Fumus In Componens)

[Diablo]

How do you test or measure oscillation to know you have it?
Are some type of amp designs immune from oscillation (Tweed style?), and others prone to oscillate (modern high gain designs?). What are the design features that promote oscillation? Does the 220 ohm resistor on the output jack tend to promote oscillation?

[Gaz]

RG, I picked up the MOV trick from your "Immortal Amp Mods" article, and, with a few unanswered questions, ordered this one:

ROV07-621K-S Tyco Electronics / Raychem Varistors - MOV & MLV

I didn't end up using them because I was too unsure about voltage rating vs. clamping voltage: which one was to be double-plus-some higher than B+?

They are physically small, and were only $.10 each. Would they have worked in an amp with a 300vdc B+? How 'bout that capacitance rating?

A clarification would be wonderful considering this is a cheap and easy mod to protect an expensive part.

Thanks.

[otto pärt]

Does the 220 ohm resistor on the output jack tend to promote oscillation?

Yeah, I'm not totally clear on this either. But if I understand correctly, the resistor is supposed to prevent the oscillation that can occur when no speaker is connected, and one should test for such oscillations simply to see whether it works. But adding the resistor isn't going to cause oscillations. Is that about right, Steve/RG?

Also, any suggestions for how to estimate the necessary wattage rating of the resistor?