Power Scaling

[gearhead]

Heh all,

I've got a Rockytop Electronics-built JTM-45ish build from a while back that has some probs. Jesse has since retired from amp building and am on my own.

It has powerscaling (where I think the prob is), but have never worked on such a unit.

When it's cranked up halfway or over, there is this sputtering sound. It's as if the signal is getting randomly chopped, with a warbling sound to it. It also seems to break up earlier.

Sent it back to him for another reason, but believe he modded the PS circuit to a newer version.

Anyone know of schematics/instructions and/or advice on all PS units out there?

[John_P_WI]

Contact:

Kevin O'Connor

at:

http://londonpower.com/

Good luck,

John

[gearhead]

Thanks for the advice.

Kevin was quite willing to discuss early on. We emailed back and forth, but the stream of info has ended. Don't know if he's busy or stumped.

Basically I have a sitution where I can easily set the bias at idle and there's no probs. However, cranking it halfway and playing results in ONE tube redplating. Swapped other tubes in and it's unique to that socket. Have an Alessandro bias meter (analog panels) and the needle is completely pegged for that tube when playing.

At idle, the screen and screen resistor voltages are equal (465 and 2.5-3V). Plate voltages were equal, but didn't write them down.

Anybody here know how to de-Powerscale an amp? PS setup is 3 SB-1 units.

Thanks all.

[FunkyE9th]

Is there a way to bypass the power scaling circuit just to see if the problem is still there?

Also, per your post, it only red plates when there is a signal. I wonder what is going on with Vgk...could it be shifting to a more positive voltage (the one that is red plating) and then not returning to its idle value? ...maybe something with the PI instead?

[gearhead]

Unfortunately, the PS is core to the install so it's kinda hard to bypass it. Two of the bias pot wires go directly to one of the PS boards. Would really have to de-PS it to really have a go.

It redplates only on playing. Stop, and everything returns to normal (idle) values.

Hadn't traced back to the PI, but will start poking around there and at Vgk, providing the PS isn't in the path.

Thanks

[paulster]

It could quite easily be parasitic oscillation that often shows up only at full power output on Marshalls rather than anything actually wrong with the power scaling.

Is this the same one that was discussed a year or two back at TGP with this problem, or is a common trait to these amps?

What value grid stoppers do you have on the power tubes? If you don't have access to a scope to look at the output waveforms then you could add a higher value (33K or 47K) grid stopper temporarily to act as a low-pass filter and see whether that gets rid of the problem. If it does then you either need to look at the lead dress and power supply routing and try to resolve it that way, or just go for a grid stopper value that's high enough not to give you the red-plating but low enough not to kill your top-end. 5K6 or 10K should be a safe bet.

[FYL]

If one tube is OK, the Power Scaling modules work properly.

Check the coupling caps on the problematic channel, one or more could be leaky. Ditto for grid R's.

(Edit: typo corrected)

[gearhead]

The schemo I have is the Marshall JTM45 reissue, which has no grid stoppers. Now, that in no way means this one doesn't, lol, so will have a look see.

FYL - It is endemic to both normal and bright channels (if that's what you mean). It's a problem manifested at one of the power tube sockets; but yet it could be (I hope) independent of the PS circuit.

[FYL]

It's a problem manifested at one of the power tube sockets; but yet it could be (I hope) independent of the PS circuit.

I meant tube, not channel (typo now corrected).

Check the coupling caps, one of them could be leaky; check tube sockets, wires, etc. Reflow solder joints if necessary.

[gearhead]

Thanks for the tips. How do you check for a leaky cap?

Just had a chance to do a quick runthrough on camparisons, not writing down yet. Most VDC points seem to be the same all the way from EL-34 grids to PI plates. Three things do standout:

Idle (DC) voltage at PI Plate that leads to the redplating socket is -lower- by 5% than the other PI Plate.

Applied input (old Eico SigGen) cranking about halfway, and all along the the chain, PI Plate up to EL-34 grids, the VAC readings for the redplating side were 20+% higher than the other. El-34 grids at 45 VAC and 55+ VAC (IIRC).

With Normal channel volume control all the way down, any signal in results in audible output to speaker (albeit low volume).


Checked out the PI tube (have a VTV 9 pin tester) and although not ideally balanced, checks out otherwise (110/120 gain and not too noisy).

Thanks, will keep at it.

[jaysg]

Thanks for the tips. How do you check for a leaky cap?

If you have a scope, there's no DC after the cap for a good cap. I always hate using a meter to check things like this. The scope is unambiguous.

I had a similar problem recently, and it was a couple tubes (in a quad). I think they simply have very different bias requirements...means somethings off with them, but not dead yet.