V4B and V4, pair of power tubes, slow red plating

[Mitch]

Hey folks,

New to the forum but have had a V4B and V4 (mid-70's, not distortion model) for around fifteen years. I having the same problem with both amps and am at wits end. On the scope with a 1K tone, each amp outputs a clean 102 Watts. When pushed to their limit with a guitar as the source, one pair of power tubes slowly starts to red plate. This is consistent in both amps.

I have basically rebuilt the power sections of both amps with no change in performance.

-Recapped both amps - 2x 40uf were replaced with 2x100uf (*tried paralleling a 40uf cap with each 100uf cap as seen in some schematics with no change observed)
-screen grid supply cap was replaced with a 20uf 600v (also tried 100uf cap here with no change observed)
-Screen grids upgraded to 1k 5W
-Plate load resistors replaced with stock value (3.6 ohm)
-R55 470omh 7W replaced
-Control grid resistors R37, R38, R43, R44 replaced with stock value
-Bias supply caps replaced with stock values
-Bias network branch resistors R35, R36 (100k) replaced with stock value
-C11 and C13 replaced multiple times (orientation swapped, no change observed)

-All diodes in amp checked out (*fly back diodes removed with no change observed)

-PI and power tube sockets replaced in one amp
-Re wired power section of one amp
-Checked all traces in power section for breaks (Went to the length of recreating the traces in the power section of one amp in a point to point fashion with bus wire. I know this adds resistance to the circuit, it was done as a mere experiment to rule out broken traces)

At one point, I drove the output tubes with a Marshall 2203 OT and observed the same red plating issue

Plate voltage is 535
Screen grid voltage is 535

One amp had old Ruby 7027's, the other amp had newer JJ 7027's - changing tube order does not alter the problem

**I did mod one amp to run 6550's and this did not alleviate the issue

Interesting observations:
-the pair of tubes that red plate exhibit a blue haloing effect
-I attached a Fluke 187 to each pair of power tubes control grids to actively monitor the bias supply. The pair that does not red plate, the bias voltage remains consistent. The pair that red plates, the bias voltage fluctuates.


Any info or suggestions anyone could send my way would be so greatly appreciated!

-Mitch

[jjmt]

Is the plate voltage the same on all the power tubes?

I originally thought HF oscillation. But then I saw that it was only on one side. And you've basically replaced everything. Except the output transformer.

Given that pointing a finger at the OT is not making me feel super helpful, the other possibility is one or more of the sockets on that side. Clean/re-tension? Maybe re-flow the solder connections on the socket? Remember that the problem might not be on the side that is red-plating. Those tubes could be biased too hot because the other side is running cold for some reason.

[Mitch]

Thanks for the reply jjmt,

The plate voltage is the same on all the power tubes. I did mention above that I drove the output section with a Marshall 2203 transformer only to recreate the same red plating scenario. Also, in one of the amps, I replaced the PI socket as well as the output tube sockets with brand new ceramics. Further, I replaced all wiring from the main PCB to the sockets to rule out any intermittent break in a wire.

The output tubes were very closely matched before adjusting the overall bias, and I also matched the triodes of the phase inverter on a curve tracer.

Thank you for your comments and keep them coming!

[AMPREPAIRNJ]

Hello Mitch,

Assuming one side is indeed fine, there are several ways to approach this:

-One issue that will offset the bias voltage is a bad ".33" cap C11/C13. A shorted or leaky cap will cause a connection between the higher voltage positive DC on the plate side of the Phase Inverter with the negative voltage on the grid supply of the power tubes. Lifting the leg on the power tube side of the ".33" and measuring the DC on both sides of the cap with the amp on will tell you if it's passing voltage or blocking it as it should. Use caution and turn the amp off and and unplug it while unsoldering/resoldering the legs. You don't need the power tubes in the amp for this test.

-With the amp off and unplugged, ohm out all bias supply resistors and connections (wires and traces) post diode. Measuring resistors in circuit may yield inaccurate readings, but having a supposedly "good" side should help to compare one side to the other. The proper way of course would be to unsolder one leg of each resistor and lift that leg up and measure the resistor directly. I've seen plenty of V-4 PCB's that will have the solder pad or trace blown due to minor soldering. These can go from obvious to hairline breaks, so be very observant.

-Observe the voltages with no power tubes in the sockets. They should all be relatively the same. If they are not, start tracing the voltage back on the sockets that are wrong until you find where the discrepancy starts. Make sure you measure the pins from the tube side of the socket as well (although I do not think you have a pin issue, it's always nice to know for sure).

-Double check your original work! The one thing both of these amps have in common, frankly, is you. Make sure you didn't cut any corners or try to make or unmake connections you found not to your liking. This could be an easy case of accidentally introducing the problem. Was this an issue before you rebuilt the outputs?



Attempt a few of these troubleshooting tests/observations and let us know how you make out. Always be careful and use extreme caution!

[Mitch]

Hey AMPREPAIRJN,
Thanks for your suggestions!

This issue was the reason I started rebuilding the output section. Those .33uf caps were the first thing I replaced. I did double check them for a DC voltage and they were clean. Also, all resistors in the bias supply checked out, including R35 and R36 (2x100k) which I had already replaced.

I did notice something interesting:
-the AC signal through the .33uf corresponding to the pair of tubes that slowly red plate (when driven very hard) is always 10v higher than that of the other .33uf. It's as if the phase inverter in unbalanced.

As part of this rebuild, I had already replaced the PI plate resistors and I double checked them tonight. C12 was also replaced at some point. The plate voltage of the PI 210vdc on each and I double checked the 12AU7 on a curve tracer to make sure the triodes were matched. I also checked R27, R29, R28, R32 and R33 - they all checked out. Cathode voltage of the PI is 9vdc.

All voltages on the power tube sockets check out. I did go back though and meticulously checked the work I had done. It seems to be good.

I will leave you with this:

I again attached a meter to the control grid on of one power tube on the "bad" half and another meter to the control grid of one power tube on the "good" half. At idle, each grid measures -60vdc. As I increased the amplitude of a 1k tone, the "bad" tubes bias voltage reached -75vdc where as the "good" tubes bias voltage reached -85vdc.

Here's a chart:

VAC VDC
R37/R33(pin3) 70 -75
R38/R44(pin3) 60 -85

After all the checking and double checking I have done, I'm dumbfound as to why these voltages are asymmetrical.

[AMPREPAIRNJ]

Did you happen to replace C16 and the bias diode yet?

What’s happening to your plate and screen voltages when the bias goes crazy?

Do you have a scope?

I’m gonna be away until Wednesday. If you haven’t solved it by then, feel free to call my shop (just google amprepairnj) and we’ll walk through your measurements with real-time observations. There’s a few quirks that are easier to explain over the phone as opposed to playing forum tag. That negative voltage swing to -85v is interesting. That should render the tube so cold it effectively barely conducts, the opposite of Red Plating. So I would say even the “good” side is not good either.

[AMPREPAIRNJ]

Ok, so I happen to have a 1973 V-2 in the shop. A typical basement find someone brought me with all the issues you'd expect, but now it's up and running 100% (V-2's actually run with higher B+ voltages than V-4's, about 50v higher). The idle Bias is -68, and when I push my signal right to the clipping line, the crossover looks great, the Bias drifts to -67 due to sag in the supply from the draw down. In order for me to get the bias to -85, I must crank the signal well beyond clipping. That will bring the bias to -85(+/-) due to the screen dropping 100 volts and the plate dropping about 65 volts. This draw down on the plates and screens offsets the bias voltage in those conditions. That's why that Bias measurement seems very odd, that extreme clipping is not a condition used to troubleshoot these amps. The observation about having separate bias voltages at that extreme clipping point may play a roll, but it's not critical enough to solve the issue since we already know one side is Red-Plating and the other is not.

Any luck with the Bias supply Cap and Diode?

Tip to anyone reading this: Anyone working on a V-4/V-series amp should always check their Hum Balance pots, they are super common to blow, sometime you don't even realize cuz of where they short in the winding. Especially if it's original.

[Mitch]

My apologies, my plan to work on these amps keeps getting thwarted. I will be back on the bench with these on Friday around 5pm EST.

C16 was replaced but the diode was not. The diode does check out, but I will replace it first thing Friday.

I really appreciate you taking the time to run that test on the V2. When I bring both of these V4's up to clip on my scope, the crossover and bias voltages also look great and it appears as if the amps are running properly. I just wanted to make it clear for everyone that the red plating does not occur until the amp is pushed into an extreme clipping state. This can be created with the volume pot turned to the 2 or 3 o'clock mark with a guitar as the source.

Just a thought I may explore Friday as well: I've have had problems with amps in the past when the manufacturer did not directly fix the control grid resistors to the output tube sockets. When the amp was pushed hard it seemed as if some sort of oscillation was created in the output tubes, causing all kinds of trouble. When I moved the grid resistors directly to the tube sockets the problems went away.


I will check back in Friday after I replace that diode.

Thank you again!

[AMPREPAIRNJ]

Ok, so as I stated before, the one thing both of these amps have in common... is you. Those settings are quite extreme. The V-4, while able to break up, was meant to be a clean tone machine, capable of one of the widest tone shaping preamp abilities in any amp both before it’s time and amazingly well after it’s time. I wouldn’t bother with the bias diode now, knowing the specifics. While the preamp and phase invterter can keep up with the amp cranked to 2-3 O’Clock, the power tubes, at those voltages, can not. These amps were designed to run the original RCA 7027’s, nothing else holds a candle to that tube and the beating it could take while still delivering amazing tones. The Ruby and JJ 7027’s are pale shadows of the original. I would strongly advise you to reinstall the fly back diodes, if you haven’t already, in order to protect your O.T. from that beating it’s getting. I’ve worked on hundreds and hundreds of V-4’s at this point, and I’ve only ever had to change one or two O.T.’s. Now that we have all the info needed, let’s look at the fact of the matter: at least two (non RCA 7027) power tubes red plate when the signal is driven well beyond clipping. The question is why. The fact is the tubes can not handle the task, but that’s an easy answer. How about what’s causing the tube to conduct so critically poorly: the signal it’s being fed. Now, while the “.33” caps may have been changed, that doesn’t mean they’re not the problem, or solution in your case. Those caps are overkill but thought necessary to pass those “Ampeg lows”, but they realistically pass frequencies that we don’t even hear, but the power tube is still working to deliver those frequencies non the less. You should start by dropping those .33 caps to .1’s and see what that gets you. Narrowing the frequency/cutting the excessive lows to the power tubes may relieve the power tubes of enough stress that your extreme clipping won’t drive it into red plating. You should experiment with the tubes you have, if a .1 doesn’t work, try a .01. You should be able to observe a difference in the red plating immediately. Whether it goes away completely will depend on your specific tubes of course. There are other paths to go down, but let’s start there!

If the amp is not red plating in normal playing conditions or at idle, without extreme clipping, then it sounds to me like you’ve got a good working amp.

[Mitch]

You are definitely correct in stating that the one thing these amps have in common is me! I know those settings are extreme and pushed to that extent, the amp doesn't even sound good. The only reason I'm testing it to that point of clip is to make sure that if I were to loan these to someone in the future, they don't melt them.

I had a spare 20 minutes today and swapped the .33's for .1's. I played the amp quickly into a dummy load at full clip and it didn't seem to display any red plating. Tomorrow I will spend more time and document voltages. You may have nailed it!