Cooked power tranny? (Vibrolux Reverb)

[MWJB]

plate voltage at V3, this will be the same voltage as you get after the choke, less what is dropped accross the reverbtrany primary, so the Q is do you have 440v at the downstream side of the choke? If you have less here, then you will have less at the V3 plate.

Don't assume the 79 schem is "right", balance it against the voltages that you actually have.

Was the vibrato footswitch RCA shorted (on) when you took the readings at V5?

Trace the problem methodically, power down & remove ALL tubes (except rectifier). Now reinstall the power tubes, do the ammmeter accross the standby trick again, you should only have plate & screen current showing (90mA-ish?). If so, carry on & install the PI tube...by how much does the current rise? If all is good, fit the vibrato tube, keep going until you see what upsets the applecart. Use known good (new) tubes, unplug from the wall between installing new tubes, as the standby is shorted and you'll have full dc on the tubes from the moment the power switch is thrown.

153mA is probably around the max current that the PT was designed for.

[slidincharlie (Carlo P)]

plate voltage at V3, this will be the same voltage as you get after the choke, less what is dropped accross the reverbtrany primary, so the Q is do you have 440v at the downstream side of the choke? If you have less here, then you will have less at the V3 plate.

I have 412V at the OT ct end and 407V at the other end.

Don't assume the 79 schem is "right", balance it against the voltages that you actually have.
Was the vibrato footswitch RCA shorted (on) when you took the readings at V5?

No. When the vibrato is engaged I only get a flashing display on my DMM and no reading.
Shouldn't I care about the 0V on the cathodes of V5?

Trace the problem methodically, power down & remove ALL tubes (except rectifier). Now reinstall the power tubes, do the ammmeter accross the standby trick again, you should only have plate & screen current showing (90mA-ish?). If so, carry on & install the PI tube...by how much does the current rise? If all is good, fit the vibrato tube, keep going until you see what upsets the applecart. Use known good (new) tubes, unplug from the wall between installing new tubes, as the standby is shorted and you'll have full dc on the tubes from the moment the power switch is thrown.

Current draw test:
- only rectifier and OLD power tubes: 130mA (plate current: 80mA)
- like above with NEW tubes: 144mA (plate current: 90mA).
I stopped putting the other tubes because it is clear that the problem is not related to preamp or PI tubes, right?
So where's the problem?

153mA is probably around the max current that the PT was designed for.

I have kept the amp on for about two hours (no guitar plugged in) and the PT got very hot...

[MWJB]

"Current draw test:
- only rectifier and OLD power tubes: 130mA (plate current: 80mA)
- like above with NEW tubes: 144mA (plate current: 90mA).
I stopped putting the other tubes because it is clear that the problem is not related to preamp or PI tubes, right?
So where's the problem?" - Rectifier, B+ rail, tube sockets? Measure the voltage drop accross the 470ohm screen grid resistors, then divide by resistor value to ensure that screen dissipation is as expected (around 5mA per tube).

Voltage at V3 plate seems reasonable.

Troubleshoot V5 later. If it's working ...then it's working.

[Steve Conner]

I noticed you used the OPT shunt method to read plate current. Did you also use this method to bias the amp? It can sometimes underestimate the actual plate current, because some of it can still flow through the OPT. So, the amp may just be biased hotter than you think.

[MWJB]

Hi Steve,

What method would you recommend, given that the 1R cathode resistor method can read slightly high (includes screen current)? I've not encountered a shunt measurement that read higher than the 1R cathode method.

Carlo, I'd look at 40mA as your upper limit in any case, so you can afford to lose a few mA off the the new tubes. Having the tubes draw very high idle currents isn't going to help with any heat issues.

[slidincharlie (Carlo P)]

Rectifier, B+ rail, tube sockets? Measure the voltage drop accross the 470ohm screen grid resistors, then divide by resistor value to ensure that screen dissipation is as expected (around 5mA per tube).

1) With an old Sylvania rectifier there is almost 30mA draw less, which brings the current draw from the amp to about 110mA (less the preamp tubes).
2) Screen grids: there is less than 1V drop across each 470 ohm screen grid resistor.

To Steve: I was about to make the same question as MWJB I thought that the OT shunt was the most accurate method to set the bias...

[MWJB]

Well that 110mA sounds more reasonable, I take it that the Sylvania is of the same type as the original rectifier (5U4G)?

I'm not surprised that there is less than volt dropped accross the screen grid resistors, that 5mA per tube was on high side (e.g. up to 5mA), it doesn't look like you have a problem there.

[Steve Conner]

Hi guys

The issue is the meter's own internal resistance. Typically, a DMM on the 400mA range will use a current sense resistor of about 1 ohm. So, if one leg of the OPT had a resistance of 100 ohms, then only 99% of the plate current would get shunted through the meter, thus it would underread the plate current by 1%. Which is nothing worth worrying about.

But if we tried it on a large OT with a resistance of 10 ohms per half primary, then it would only read 90% of the actual plate current. More of a significant error now.

I tend to use the 1 ohm cathode resistor method and just ignore the screen current, which overestimates the bias by a few percent. Or else I permanently install some cathode resistors and hook them to a meter on the front, which I think is a cool gimmick...

[MWJB]

I see. A Vibrolux OT might measure 130-150ohms per side, but I take your point, especially with regard to bigger 100W+ style OTs.

I must admit I don't use the shunt method so much these days, too much risk of forgetting to reset the meter to volts & then attempting to read a plate voltage! Not recommended! I use bias probes mostly, or make sure I have 2 meters handy with one set for current & the other for volts/ohms.

It's worth reiterating that whatever the method, a recogniseable datum line is the important thing, so long as you're not pushing your luck with plate dissipations (and there's really no reason to be most of the time), either method works well.

[slidincharlie (Carlo P)]

So with the old Sylvania (yes, a 5U4GB of course) and with colder bias setting I might be more in the ballpark now...
The V5 is still to be troubleshoot, with its 400V on the first plate and 0V on the cathodes. It is quite far from any bf/sf Vibrolux schem. Any hint? Or may I leave it and be happy?
Have a nice weekend in the meantime...

[slidincharlie (Carlo P)]

Now the amp has the voltage selector on 240V (mains around 235VAC here), which brought the heater voltage down to specs. BUT this setup allows for lower than specs B+ (390V instead of 450V of silverface specs). With the 6L6 bias set at 40mA I am at 15.6W or 56% max plate dissipation. As long as I like the amp's tone, do the power tubes risk to be overbiased at such values? If not, I would be happy with the current setup since it sounds safer for the PT and should let the whole amp run cooler.
I'd love to hear a final comment on this current scenario.

[pdf64]

That's great, there's no danger from too much bias voltage (though your bias set up is fine).
If you're happy with the tone, go with it.
If you want to experiment, you could try a GZ34 in there to see how much more plate voltage that gives you - remember to re-bias though.

[slidincharlie (Carlo P)]

If you want to experiment, you could try a GZ34 in there to see how much more plate voltage that gives you - remember to re-bias though.

You mean like the blackface version of this amp... I have a new GZ34 here.
Isn't there any risk with a stock silverface PT, once rebiased?

[pdf64]

No, infact it will less less load on the rectifier heater winding (2A instead of 3A), saving 5 watts, though a higher B+ would draw a little more current from that winding, unless you re-bias. Peter.

[Dave Curtis, dB AudioTech]

No, infact it will less less load on the rectifier heater winding (2A instead of 3A), saving 5 watts, though a higher B+ would draw a little more current from that winding, unless you re-bias. Peter.

With the extra B+, and the static dissipation @ just over 15W with the 5U4, I'm thinking he'll have to rebias. I run a SS rectifier in my '76 Pro Reverb; no problem after a rebias.

Re: 15W dissipation: tube life will be extended. The VT22 I have came with the original Magnavox 7027s biased at around 14W; still going strong at 35 years old.

[slidincharlie (Carlo P)]

Hey...
I put a JJ GZ34 tube in place of the 5U4GB and for starter I put an eye on the total current draw... In a few seconds the amp draw as much as 190mA and was still rising, then I shut it off.
The I turned it on again to check the plate voltage on the 6L6's and I got 430V (instead of the 390V obtained with the previous stock rectifier).
What can cause such a high current draw?
I haven't even checked the power tubes bias for fear of cooking something...
If you check my posts above, I had around 120mA total draw with the new (old actually) rectifier and the mains selector on 240V...

[pdf64]

Always turn your bias up full, ie to highest negative voltage on the pot, before fitting new power or rectifier tubes.
Then adjust bias to a suitable level.
Sorry if I took that as said, guess I should have noted it in previous post.

[slidincharlie (Carlo P)]

hmmm... I believe that something strange is going on here with the GZ34.
I rebiased the 6L6's to 40mA @ 430V on the plates, for 17W dissipation. Apparently an optimal scenario.
Yet the amp is drawing much more current (160 mA) than with the the stock 5U4GB and the power tubes biased at 40mA @ 390V on the plates.

Unless you guys can give me a possible explanation of this and redirect me to a different solution, I think I'm better off with the 5U4GB.
That said, if 400V on the plate (pin 1) of the tremolo tube are not an issue, I think I'm finished with this amp. Looking forward to road test it next Wednesday at a gig!
Thanks.

[MWJB]

"That said, if 400V on the plate (pin 1) of the tremolo tube are not an issue, I think I'm finished with this amp. Looking forward to road test it next Wednesday at a gig!" The Fender schems say measure vibrato tube voltages with effect "switched off at the footswitch & speed & intensity at full CW rotation".

If you had 0.0v at the cathodes the tube would be switched off & not working.

I don't know why 17W dissipation is "optimal", you seem to be consistently setting to the high side of normal (30-35mA, +/- 2-3mA).

Out of interest do you have paralel or series filer caps at the main B+ node, you say caps were done a few years a go, I'd doublecheck that they are wired correctly, whether it be series or parallel.

[slidincharlie (Carlo P)]

The Fender schems say measure vibrato tube voltages with effect "switched off at the footswitch & speed & intensity at full CW rotation".
If you had 0.0v at the cathodes the tube would be switched off & not working.

Vdc readings on the tremolo tube are the same irrespective of the footswitch and of knob rotation. For whatever reason today I have 360+ V on the V5 plates (40V less than the other day): maybe due to low B+ (370V instead of 390 of yesterday) and low mains supply (230V instead of 235+).
I have just re-checked the cathodes of V5 and I have 0.0V on pin 3 and 0.021V on pin 8 (unless my Fluke DMM went nuts).

I don't know why 17W dissipation is "optimal", you seem to be consistently setting to the high side of normal (30-35mA, +/- 2-3mA).

I am no tech, so I rely on techs' and experts' advice that I get on books and in reputable forums like this one. Some like you tell to stay in the 30-35mA range, others say to stay within 70% plate dissipation. With 17W and 56% of plate dissipation I thought I was in a correct setup...

Out of interest do you have paralel or series filer caps at the main B+ node, you say caps were done a few years a go, I'd doublecheck that they are wired correctly, whether it be series or parallel.

I replaced the old caps with new caps one by one, paying much attention to wiring... But who knows, I'll check again.

[MWJB]

"With 17W and 56% of plate dissipation I thought I was in a correct setup... " Well, there's no "correct" set up, zero current will stop the amp from working and too much current will burn up tubes/blow fuses...maybe damage OT. Anywhere between the 2 extremes that gives you the performance you like is "correct"

It's just a bit odd that you are, on one hand, worried about excessive current draw & on the other hnd you are biasing on the high side of normal? On a Fender with 6L6s i set to 30-35mA and have a listen...by all means if you prefer 40mA & you are within a reasonable plate dissipation, then that's fine...but bias it there because of the tone, not to meet some arbitary percentage figure. A 6L6 Fender will still sound ball park at 30mA, whether it runs 400vdc or 525vdc.

[Prairie Dawg]

Divide and conquer is the watchword for repair issues.

For power supply problems, the chief thing to be divided is problems internal to the power supply from problems external to the power supply. On tube amps, this is easy - pull all the tubes. Now put a voltmeter on a secondary and let'er sit until you either get the power turn-off or until you're tired of waiting for it to turn off and therefore conclude it was the amplifier pulling too much current.

If the transformer does not get hot and go open, the problem was past the rectifier tube.

So then put in the rectifier tube, but no others. This then adds the filter caps and choke to the loading. If the trannie now overheats and goes open, the problem is with the filter caps and possibly (I've seen this only once) a short from the choke winding to its core, or a wiring short in the power distribution network.

If the trannie does not get hot and go open, the problem was past the filter caps and choke.

Now put in the power tubes. Try it again. The power tubes or wiring to them could be bad and causing them to pull too much current. This could be a tube problem, bias problem or capacitor problem if it's emitter biased or screen component problem if it uses screen resistors and capacitors.

That set of debugging will get the vast majority of all power problems, as the preamp and PI tubes use only a tiny fraction of the amp's power, even if a tube is shorted.

If it's not one of those, you have a real rarity, and therefore an interesting problem.

That's basic troubleshooting theory, RG. I used it a lot in my former life as an aircraft engine mechanic/test technician.

The first thing to do is define the universe you're working in and work to limit its geography. In this case the problem is located in a 12x24 piece of real estate. Already things are starting to look up.

As my old crew chief always used to say, "Think 'system'" that is, visualize parts working together in a systematic way and not in isolation. Then, localize the problem within the system.

then, you can proceed to test variables and see where it leads.

Amps are nice that way because the power's organized in a linear fashion.

The 'interesting problem' is, indeed interesting, and I have found it to be almost always attributable to lack of a certain speck of knowledge.

One time it was a check valve the size of a pencil lead that is buried underneath a prop governor that appears on no schematic diagram. If a bit of crud holds it open you can't schedule forward pitch.

The other time it was a mis-drilled hole that hand been plugged with epoxy and redrilled. 700 hours later it blew out and the oil pressure dropped 30 pounds while the pilot happened to be looking at the pressure gauge. Massive freakout in the cockpit.

That one was harder to solve.

[Steve Conner]

I agree. Since I work with computers a lot, I use binary divide and conquer. That means cutting the "universe" as nearly in half as you can, to see which half the problem is in. Then cut the part containing the problem in half again, and repeat until you're down to the faulty component.

This is not so good for "systems" problems: with one of these, you might cut the system in half and discover that both halves worked just fine when tested in isolation, but when you join them back together the whole system fails. Even that tells you something, though.

[slidincharlie (Carlo P)]

It's just a bit odd that you are, on one hand, worried about excessive current draw & on the other hnd you are biasing on the high side of normal?

Actually I've never thought to an amplifier in terms of current draw... In this thread excessive current draw has been identified as an issue in this amp, possibly causing excessive heat in the PT.
To make things more complicated, voltages have varied quite a bit in the last two days within the same tube setup.
I have just tested again the two different recto setups, and this is what I got:
- with the 5U4GB = 380V B+, 35mA bias on the 6L6's and 133mA total current draw;
- with the GZ34 = 430V B+, 35mA bias on the 6L6's and 147mA total current draw.
If it was for tone I'd like the GZ34 setup better becuase the amp sounds a little louder, cleaner and bolder; in a few words I find it closer to its original silverface personality. BUT 147mA current draw I've been explained it's very high. Yet I haven't found a culprit for such high current draw. As reported in one of my earlier posts, almost all the current is drawn with only the rectifier and the power tubes installed.
Is there anything more that I can do to identify what is sucking all the current?

[MWJB]

You have checked tube sockets & power supply wiring.

How much current is drawn accross the standby with just the GZ34 in place? Then try with a SS rectifier if you have one. Do a similar test & see what is "normal" in another, similar amp. Establish a datum line. I would really worry about the current draw test if it was massively out of whack.

The vast majority will be drawn by power tube plates, preamp tubes don't draw much.

If heater voltage is OK you probably don't have anything to worry about, but 50v drop between the 2 rectifiers seems a lot (high side of "ball-park", could be down to brand/performance of individual parts).

If your PT is just a little overstressed/borderline, it might last for years & years, still run hot & the amp may compress a bit more.

[slidincharlie (Carlo P)]

You have checked tube sockets & power supply wiring.

Besides looking for carbon traces that might cause maybe a current draw (?), what else should I check in a socket? With a meter, I mean.

[MWJB]

Any possible shorts accross pins/from pins to chassis...you already know plate & screen draw for power tubes & they seem normal. Visual inspection should do it.

[slidincharlie (Carlo P)]

How much current is drawn accross the standby with just the GZ34 in place? Then try with a SS rectifier if you have one. Do a similar test & see what is "normal" in another, similar amp. Establish a datum line. I would really worry about the current draw test if it was massively out of whack.

I tested everything again.The current draw with only the rectifier in place (GZ34, 5U4GB or ss) is a fraction of one mA.
Adding the power tubes brings the current draw up to 123mA with GZ34 and 110-111mA with 5U4GB and a ss recto.
Adding the preamp tubes one by one adds about 15 mA more. This means around 140mA with the GZ34, and around 125mA with the 5U4GB or ss.
I don't have other similar amps to compare unfortunately, but the PT is a stock late 70's unit and, as you said, I expect it to last for many more years if 140mA isn't too much... But isn't it too much?

[pdf64]

'Adding the power tubes brings the current draw up to 123mA with GZ34 and 110-111mA with 5U4GB and a ss recto'
That's with the power tubes adjusted for 35mA each each time?
Or the bias adjust left the same for both?
Whatever, I don't understand how even with the 5U4, power tubes can have total of 75mA (including screens), but the B+ current is measuring 110mA.
There must be a measurement error somewhere.
Suggest that you break into a power tube cathode ground return and measure the cathode current directly, to se if we can find the missing 35mA somewhere. Peter.

[slidincharlie (Carlo P)]

That's with the power tubes adjusted for 35mA each each time?
Or the bias adjust left the same for both?

Adjusted each time to 35mA each.

Whatever, I don't understand how even with the 5U4, power tubes can have total of 75mA (including screens), but the B+ current is measuring 110mA.
There must be a measurement error somewhere.
Suggest that you break into a power tube cathode ground return and measure the cathode current directly, to se if we can find the missing 35mA somewhere. Peter.

I put a 1 ohm 1% resistor between the cathode of each 6L6 and ground. The direct current reading taken with the OT shunt method are confirmed (that is, about 38mA from each tube, as expected across the cathode resistor reading).
Any other hint to chase that excess current draw?

[MWJB]

Voltage creep accross the circuit board? Measure various points on the board for voltage, especialy hi voltage eyelets, also try stuffing an extra layer of insulation between the top circuit board & the lower board. Voltage creep on the board is usually accompanied by some unwanted noise, but have a look see anyway.

PT should take 140mA happily, at higher demands the B+ will just drop, if it sounds OK to you is it necessarily a problem?

[pdf64]

In the spirit of investigation, put 1 ohm resistors between the OT centre tap lead and the B+ capacitor, and between the screen grids and their B+ cap.
Then, with just the power tubes in, compare the volt drop across the 1 ohm B+ resistors, 1 ohm cathode resistors, and your normal current measurement (current meter across the standby switch).
If the OT and screen 1 ohm resistor volt drops add up to equal the cathode resistor volt drops, but the current meter is still higher, then I would put it down to being a measurement error.
Because we already know that with no tubes in, there's almost no current.
The measurement error possibly caused by the high level of ac ripple, and the higher equivilant resistances of a 5U4 than GZ34.
I'd assume that your current meter isn't giving accurate measurements in that situation.

[Steve Conner]

While we're clutching at straws, have you checked the accuracy of the mV range on your meter? Maybe it underreads.

If 35mA were flowing through a leaky tag board, I'd expect clouds of smoke.

[slidincharlie (Carlo P)]

I've been away from home for two weeks now... Hope somebody is still willing to follow me...

Voltage creep accross the circuit board? Measure various points on the board for voltage, especialy hi voltage eyelets, also try stuffing an extra layer of insulation between the top circuit board & the lower board. Voltage creep on the board is usually accompanied by some unwanted noise, but have a look see anyway.

The only place where I can read some AC voltage is on the OT center tap eyelet (5.3Vac).

PT should take 140mA happily, at higher demands the B+ will just drop, if it sounds OK to you is it necessarily a problem?

I've been told above that this amp, with its tube complement, is expected to draw around 100mA. So there are about 30-40mA that are being drawn in excess with the GZ34 (or about 20-30mA with the stock 5U4GB). I ignore the maximum continuous current rating of this PT... You say it should handle 140mA happily... I guess I am happier with the GZ34 tone-wise... So if you guys believe that I am not stressing the PT too much, I'd be incline to leave it alone.
BTW, I play this amp only a few times per year 'cause it's too loud for most local venues.

Re: meter misreadings. I have a good old (ICE brand) needle meter. Is it good to check my DMM against this meter?

[MWJB]

You should be looking for dc, not AC on the board, between the eyelets.

Under large signals the PT's current rating is going to be exceeded anyway, when this happens the secondary voltages drop...you would just seem to have a little headstart on that front, so I'd live with it.