OT protection

[ThermionicScott]

To protect against the case where nothing is plugged into the jack, you can use a shorting jack with a power resistor wired between the inner tab (the one that contacts the hot lug when nothing is plugged in) and ground. Might be tough to find a spot for the appropriate size resistor in a big amp, but if you can pull it off, you've always got a dummy load ready for testing!

- Scott

[R.G.]

To protect against the case where nothing is plugged into the jack, you can use a shorting jack with a power resistor wired between the inner tab (the one that contacts the hot lug when nothing is plugged in) and ground. Might be tough to find a spot for the appropriate size resistor in a big amp, but if you can pull it off, you've always got a dummy load ready for testing!

- Scott

What happens when the cable is plugged in but your bass player pulls the cord out of the speaker cab, not the amplifier?

[R.G.]

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?

That is correct. Some amps can oscillate if the speaker is unloaded. What that does is present the plates with a load composed of the leakage inductances and the parasitic capacitances, since the real load on the secondaries is open circuit. If the feedback network lets this back into the amp from the open secondaries and the phase is (in)correct, it will oscillate, and the inductances let the signal get big. This can punch through the insulation of the OT, or burn it up from eddy current losses in the core; not to mention melting the output tubes and overloading the power supply. But apparently it's not that prevalent to get everything lined up.

The resistor helps by (a) making the real loading not be an infinite resistance and (b) damping the Q of the leakage inductances and distributed capacitances.

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

I've seen 2W used, but a 5W noninductive or MOX would be dandy. The problem is that it's hard to get the thing into disaster mode to test how big it needs to be. And the resistor doesn't need to be so power capable if it's not in disaster mode.

[otto pärt]

... If the feedback network lets this back into the amp from the open secondaries and the phase is (in)correct, it will oscillate, and the inductances let the signal get big. ...

Does that mean there's no danger if there's no NFB?

[bob p]

What happens when the cable is plugged in but your bass player pulls the cord out of the speaker cab, not the amplifier?

sounds more like something a DRUMMER would do!

[Bruce / Mission Amps]

What happens when the cable is plugged in but your bass player pulls the cord out of the speaker cab, not the amplifier?

RG knows this but,
I'd just do it like the old AC50s and AC100s had.... with a 5 watt 470 ohm resistor right across the 15 ohm tap on the OT terminal block to ground... for those that don't know what that means... the 15/16 ohm tap wire is shorted to ground by a 5 watt 470 ohm resistor before the lead out wires go to a speaker jack.
This way there is always a 470 ohm load on the OT.

[Steve Conner]

Does that mean there's no danger if there's no NFB?

There's no danger from oscillation if there's no NFB. Actually, maybe there is, because there are different kinds of oscillation with different causes.

You might not get the oscillation of the power amp's negative feedback loop, what I call Nyquist oscillation, but you can still get oscillation caused by early stages picking up the signal from later ones through capacitive coupling. This is extremely common in high-gain Boogie-style homebrews, builders of these have been known to go insane trying to cure it. Sticking pins in an effigy of Randall Smith doesn't work, but it can make you feel better.

And, NFB, oscillation or not, you can still blow the OT with excessive audio signal voltage by cranking it unloaded. This is actually easier in an amp without NFB, because pentode and beam tubes are current sources: they'll try to push as much voltage as they can if the load goes open, and NFB helps to tame that voltage down, the same way that it damps speaker resonances.

5 watt 470 ohm resistor right across the 15 ohm tap on the OT terminal block to ground

So like my old Selmer T'n'B 50 too :-) This is what I meant earlier. But maybe 220 ohms if your OT only has an 8 ohm tap. The resistor won't cause oscillations, in fact it can't cause any harm that I know of.

[R.G.]

Does that mean there's no danger if there's no NFB?

Well, to be rigorously correct, there is no danger if there is no feedback of any kind, positive or negative.

As Steve points out correctly, there is always feedback, even if only parasitic.

The necessary elements for oscillation are
(1) gain; no gain, no oscillation because the losses in the signal path are never made up
(2) feedback; positive feedback always causes a lockup solidly in one direction unless it is arranged to "disengage", perhaps because the signal is through a capacitor or transformer to cause the feedback to be non-DC. If the feedback is DC, and the circuit locks in whatever state it is in. Some external signal must cause it to change. This whole class of circuits includes flipflops, one shots, Schmitt triggers, and relaxation oscillators. Negative feedback can be made into positive feedback by circuit phase shifts
(3) a phase shift to make the feedback into positive feedback at the input of the circuit. Positive feedback circuits do this automatically. Negative feedback circuits do it when reactive components (caps and inductances) cause the negative feedback to shift to positive at some frequency. The caps and inductances may be in the circuit or in the power supply, and most confusingly, the ground path. The power supply path usually causes so-called motorboating. The change of high frequencies to positive at some frequency is what most people mean by the Nyquist Criterion, although strictly it's either high or low frequencies. Sometimes it's just called gain-phase.

To get oscillation, you have to have (1) a gain of one or more (2) through some feedback network (3) at a frequency where the phase shift is N times 360 degrees, where N is 0, 1, 2, ... Removing any one of these conditions kills the oscillation.

[mooreamps]

RG knows this but,
I'd just do it like the old AC50s and AC100s had.... with a 5 watt 470 ohm resistor right across the 15 ohm tap on the OT terminal block to ground... for those that don't know what that means... the 15/16 ohm tap wire is shorted to ground by a 5 watt 470 ohm resistor before the lead out wires go to a speaker jack.
This way there is always a 470 ohm load on the OT.

Folks. Listen to Bruce..... He knows what he's saying.... I do it the same way ; and I "no not" use a shorting plug for the speaker jacks.....

[JoeM]

Sticking pins in an effigy of Randall Smith doesn't work, but it can make you feel better....

Unless he's already patented that...

[d95err]

RG knows this but,
I'd just do it like the old AC50s and AC100s had.... with a 5 watt 470 ohm resistor right across the 15 ohm tap on the OT terminal block to ground... for those that don't know what that means... the 15/16 ohm tap wire is shorted to ground by a 5 watt 470 ohm resistor before the lead out wires go to a speaker jack.
This way there is always a 470 ohm load on the OT.

Kevin O'Connor strongly recommends this practice. He says he test-drives all his new amps with no speaker plugged in, just a 270 ohm resistor over the secondaries. His take on OT blowouts is that they are always due to tube failure or intermittent load (i.e. glichy speaker jacks or plug in/out of speaker while playing, never due to impedance mismatch.

Using the jack to switch in a load is not safe. Jacks eventually get dirty, so you can't rely on the switch actually working.

[otto pärt]

I've seen 2W used, but a 5W noninductive or MOX would be dandy.

Is noninductive important?

[R.G.]

Is noninductive important?

Not critical, for most amps. Speakers, are, after all, inductive. But with an amp you're worried about, why not just get noninductive and not worry about that at all. It will cost you a quarter more per resistor. OK, quality costs money.

[Enzo]

Switching a guard resistor in and out isn't necessary. Bruce's 470 ohm resistor winds up in parallel with the speaker. As an exercise, calculate what effect a 470 ohm resistor has in parallel with something like an 8 ohm speaker.

Hint: not much. Your speaker won;t know it is there.

[crazytooguy]

OK, I really like these protection schemes, since I'm playing around with output section topologies using 2 EL34s and 2 6L6s. The OT is from a 30 yr old Music Man amp, so I'm not interested in killing it.

So, to be clear about the placement of the MOVs - is it best to use a MOV from plate to ground on each leg of the OT, or across the plate connections (not grounded)? There's MOVs available with 600V ratings so there's no need to gang them up (I'm operating with a 485V B+) and I'm thinking that a 1.2kV clamp voltage is just about right. What do you think? Thanks!!