Ampeg SVT3 Pro worries

[ldervish]

After several months hiatus I started rebuilding an SVT3 Pro that had at least 2 previous output failures before it came to me. Numerous parts were replaced and yesterday I hooked it to a lightbulb limiter and variac to do my initial checks powering up. With 50vac supplied I measured the 16v, 100v, and 65v rails respectively with the following results: TP11=3.95, TP12= -5.78, TP13= 35.18, TP14= -35.35, TP9= 23.73, TP10= -23.78 (see pp 2-3 of attached schem for location)

Is there reason to be concerned that the +16v rail (TP11) is not in balance with the negative? If so, where should I look?

[Jazz P Bass]

You have to go higher on the Vac in.

It's a solid state amp.

[ldervish]

You have to go higher on the Vac in.

It's a solid state amp.

I was concerned with the imbalance, not the level. The other two rails are pretty balanced, this one is not. I'll take it up higher and see if it resolves.

[Mick Bailey]

The way the two supplies are derived are not symmetrical - you have V1 series heaters that may affect the +16v rail until your voltage gets a little higher to allow the 16v zener to operate correctly.

Try removing V1 and see if the rails equalize any better. You may need to remove V2 as well.

[ldervish]

Interestingly removing V1 & V2 did not significantly improve the +16v side but did increase the neg reading, so I put them back in. Once I was fairly certain a catastrophic event wasn't in the offing, I removed the lightbulb limiter and plugged the amp directly into the variac. As Jazz P predicted, the voltages rose, with final readings at TP11= 17.02v and TP12= -17.36v. (Supply here runs 120vac to 123vac.)

Thanks to all for the usual great advice.

[ldervish]

Instead of starting a new thread I'll just keep this one going a bit longer since the reference schematic is already here.

The next step in my resurrection of this amp is to set the bias, if I am able to with the equipment I have, which is an mp3 player and a digital meter (no 'scope). AP1 was replaced with a multi-turn pot, currently fully CCW and reading 8.43 kohms. I know I won't be able to set the bias at the point the crossover notch disappears, for the simple reason that I can't see the notch without a scope. So the first question is, is this worth trying myself, or should I take it to a shop?

If I can get it close without the scope, what are the amp settings e.g. EQ: 'out'; tube gain: "max"; speaker connected; mp3 to "power amp in"; etc. as I bring the output into conduction?

[ldervish]

^^^

[Jazz P Bass]

To set the bias, do not use a signal or a load.
It is an 'idle current' adjustment.

The bias setting that I use is 25mv's across the 0.47 ohm ballast resistors.
This is an average value.
Some mosfet's will draw more, some less.
If you find any readings that are really out of wack, those are the fet's that will not be sharing the load very well.

[ldervish]

Bringing it up slowly, only one anomaly. At the present AP1 setting, readings across R32,34,36,38 (N channel) respectively 7.0mv, 1.8mv, 1.7mv, 2.1mv and across R33, 35,37, 39 (P channel) respectively 3.6mv, 3.1mv, 2.6mv, 4.0mv. Note the N channel readings are about 1/2 those of the P channel (excepting the R32 anomaly).

Again I hesitated to bring it up more. Should I continue?

Even with the multi-turn pot this is a sensitive adjustment, it must be nearly impossible with the OEM one.

[ldervish]

Got it! Replaced the one wildly different mosfet and the rest came into line (more or less). Setting the reading to 25mv (average) was still touchy and after several tries with long periods of 'burn-in' between, I left them about 1mv low. P-channel & N-channel averages were very close.

Much appreciation to Jazz P Bass

[Jazz P Bass]

Glad you got it.

Yeah, that one was a tad wild.

In a current sharing setup is is best to 'match' them as close as possible.