Power rating for "safety resistor" across OT

[Gaz]

I was wondering how to properly rate the safety resistor, famously used in Ampegs. I currently have a 10W/220R from the 8R tap to ground, and measured 45 VAC across it with max signal, which comes out to 9.2W dissipation. Is this correct, or am I figuring it wrong because it's in parallel with the speaker?

[trobbins]

You're correct with your calculation - which would exist if you didn't have a speaker connected, and maxed the pots with signal input.

However, average signal level is likely to be less, and hopefully you wouldn't play for too long before you realised there was no sound through your speakers! So resistor power rating seems fine.

Unless you have a 250W amp, then your protection resistor wouldn't see that max level during normal operation with a speaker.

There are other techniques that can also be added to help protect your amplifier output transformer than just a light load.

Ciao, Tim

[Gaz]

Thanks, Tim, it's a 100 watter, so it'll be okay, but maybe a 25 W would be a fail safe I suppose - some folks turn it up when they don't hear anything, of course

You seem to be hinting at a more classy technique, care to share?

[trobbins]

The other technique I often use is a series MOV plus resistor across each primary half of a PP output transformer. The MOV has to be DC voltage rated (1mA level) at above the HT level. The resistor damps the MOV if it conducts, so I usually use a 2W rated at the nominal PP impedance (eg. 2k7 or similar). Smaller MOVs are preferred as they have almost negligible capacitance. Two MOVs can be used in series for a higher rating. Most of the movs used on 240VAC equipment have a DC rating pretty well aligned for 500-600V HT levels.

The aim is to catch any over-voltage swing on the cut-off winding side of the OT. The advantage of primary side protection is that is where the overvoltage occurs and where any consequential insulation breakdown occurs, rather than loading a reflected voltage on the secondary where the R loads at all signal levels.

[Gaz]

Thanks, Tim. I've read about that approach in one of R.G.s "Immortal amp mods" articles, but he does not mention adding the series resistor. Interesting, and thanks again.

[mooreamps]

I was wondering how to properly rate the safety resistor, famously used in Ampegs. I currently have a 10W/220R from the 8R tap to ground, and measured 45 VAC across it with max signal, which comes out to 9.2W dissipation. Is this correct, or am I figuring it wrong because it's in parallel with the speaker?

I have a question...... You did measure this with the loud speaker plugged in, yes ??????

-g

[Gaz]

No, with a resistive 100W/8R dummy load. I think hearing that sine wave would've made my head explode.

[mooreamps]

OK, so it is a big amp... One thought comes to my mind.. One could put the "safety resistor" across the 4 ohm tap... I would think it would still put the constant load on the secondary winding, but at almost 1/2 the voltage, but not quite 1/2 the voltage....

-g

[trobbins]

Thanks, Tim. I've read about that approach in one of R.G.s "Immortal amp mods" articles, but he does not mention adding the series resistor. Interesting, and thanks again.

The series resistor makes sense to me, as it effectively loads the winding section with the nominal impedance presented by a matched output for that portion of the over-voltage waveform where the MOV is conducting. A low resistance clamp from just a mov would also be ok but would certainly be an abrupt form of clamping. The OT insulation would be ok for reasonable levels of 2x HT voltage, so I can't see the benefit of a hard clamp. The MOV-R also becomes an RC filter with a corner frequency that is typically at a high enough level for even hi-fi afficiando's, and probably quite effective as a zobel network for evening out impedance variation above 50-100kHz.

[Bruce / Mission Amps]

45vac RMS? With an 8 ohm speaker that is a solid 260 watts of power. That's some serious amp power there.
Is this the stock resistor that has been in there for ... years?
Because the two loads are in parallel, the total approximate impedance with the 8 ohm load is 7.72 ohms.
However, with respect to a real 8 ohm speaker, that depends on the frequency you are testing at too.
Most of the current is going into the speaker load since it is so low compared to the 220 R anyhow.
However... like you, I would be nervous with a 10 watter there and an open speaker wire connection.
I'd get one of those chassis mountable 25-50 watt aluminum Dale type resistors... with the small cooling fins.