shorted OT primary

[Alex R]

A general question: could shorted turns or another non-obvious fault in an OT cause a power valve to redplate intermittently? I won't bore you with the details of the amp involved... just yet... I've always assumed the answer was 'no', but have an cathode-biased PP 2 x EL84 amp before me with no other explanation immediately obvious.

[Old Tele man]

A general question: could shorted turns or another non-obvious fault in an OT cause a power valve to redplate intermittently? I won't bore you with the details of the amp involved... just yet... I've always assumed the answer was 'no', but have an cathode-biased PP 2 x EL84 amp before me with no other explanation immediately obvious.

...typically, no, it's more often the problem of "loss of bias" that results in red-plating, which is the result of too much, uncontrolled (due to lack of bias) plate current.

...a loss of plate voltage results in sudden "burn-up/shorting" of the screen grid, which will not affect the plate at all because without plate voltage there's no plate dissipation.

[Marc]

shorted OT

A general question: could shorted turns or another non-obvious fault in an OT cause a power valve to redplate intermittently? I won't bore you with the details of the amp involved... just yet... I've always assumed the answer was 'no', but have an cathode-biased PP 2 x EL84 amp before me with no other explanation immediately obvious.

Hi Alex,

I had an amp this summer that had an OT on the way out. I initially missed the signs of impending death-one of the el84's was redplateing but when I reversed them it seemed to settle down. I had the amp a year ago in my shop for a rebuild to an 18W Marshall style circuit and it sounded very different than when it was brought back in. I suspect there was a tube failure that the owner did not report to me and/or he has pushed the amp enough during the year to make the OT lose one of the legs on the primary. He stated that when he really pushed the amp it would bloom and sound great but then it would suddenly drop out and sound thin. But did he stop the first time it had a symptom? Heavy sigh. Well the thin sound was it trying to run an unbalanced SE signal....

Once I recoginsed the OT was crippled and replaced it the performance was back almost to where it was (the replacement was not as nice as the old one was).

So yes, if the short is not completely killing the leg or is simply making it hobble (to keep the leg analogy going) the sound... until it kills the winding and or the tube.

I always thought transformers were go or no go. But a friend brought a pro reverb over last month that had a similar problem-one wonding was off on the DCR by 20 ohms from stock. Stock being about 80 ohms DCR on the OT primary. His amp had one leg around 87 and the other about 62 ohms. Sounded like ca-ca. We tacked it a replacement OT and ba-da-bing it sounded great. the owner said this amp just dropped in volume during a gig with no tube failure or other warning. I guess it was just it's time.....

Marc

[Amp Kat]

I've never seen it. Check the Cathode. If there is a Cathode 1 ohm current sense resistor and it's blown or a high resistance the tube will start conducting current and redplate. The bias resistor should be checked and the coupling cap from phase inverter should not be passing HV DC.

[Enzo]

When it red plates, check the voltages around the socket, especially the grid. That will tell you if the bias is leaving. Since it is cathode biased, we would assume hte grids to be at zero volts DC. If tehy wander positive, look for a leaky coupling cap from the PI.

[Alex R]

Thanks guys. Yeah chickens and eggs isn't it with redplating and burned OTs. I say the hot valve came first, but I did want to check.

This is a Park Vintage 20 combo, 70s PCB amp, 2 x EL84. Not fashionable but they sound good. Bloke brought it in with a report from another tech that the voltage was too high for modern tubes, asking me if I could drop it and reporting the occasional redplating. It was about 380v, ok, a bit high I guess. I did a cap job and put in a zener to drop the voltage to 350. I put 140ohms on the cathodes (stock value is 100) and two new JJs. It pulls 40mA a side; still way high (but it must have been sky-high before), but when he listened to it he said it wasn't as loud and cutting as before... so I said ok, I won't soften it any more by biasing it properly (!), give it a go and see if the tubes will take it. They still redplated after a few hours' use so here it is again. I can't make it redplate but I'll try again and check the volts if it does.

The OT has had a lot of abuse, sure, but it measures and draws current ok.

Might as well put new coupling caps on it anyhow, plug the last hole as it were.

Maybe I'll convert it to fixed bias, tap it off the secondary, that would give him the power and cut he wants and control things better. Marshall did a 20watter the same in every way except for fixed bias, I can see how they might have used the same board. Or maybe I should just bias the thing properly and tell him he was buying his power at the price of burned tubes.

I'm going to a lot of trouble I know, but he's a nice guy and he paid me well for the first job... which had zero effect from his point of view.

[Amp Kat]

Those Parks are damn nice sounding amps and I guess it's just so much difference in NOS and current production tubes. The NOS tubes are just better and last much longer. Even though with those voltages and that small of a Cathode resistor value it's naturally going to produce more power. Don't know about your voltages if they jumped higher from the 70's but ours did and it causes many amps that ran at say 340 volts now run at 390 and the heaters went from 6.3 to 7.5 so that is a consideration also. You may try bringing that plate voltage back and just rasing the Cathode resistor some to drop the current thru the tube or suggest to the guy he may need some NOS valves to keep it honest. I have some Sylvania 6BQ5's that have outlasted two pairs of Sovtek EL-84's and they are still going strong.

[Enzo]

Whenever I hear about some amp that has a problem, and the proposed diagnosis is that the B+ is too high for the tubes, I have to think about all those Fender amps where the B+ is "too high" for the 6V6s. But the amplifiers, blissfully unaware of this fact, continue to work for the last 40 years.

Guitar amps routinely use B+ higher than spec. The specs are for tube designs for consumer products - TVs and table radios - and the specs were meant to insure years of long life and reliable service.

In my experience, excess voltage will lead to tubes arcing over and failing, but not usually red plating. When I hear a dignosis of "the design is too high a voltage," I tend to assume that means the diagnoser just doesn't know why it is happening and is trying to rationalize the situation. Did the amp do this from day one? DO all the others of the same model do this? If not, then I would think the design was not the issue.

Personally I'd avoid just building a new amp in the old box as an alternative to actually fixing the problem. I know it is tough without the symptom, but static tests don't tell the story. And measurements around a part when it is NOT messing up don't tell us what is going on when it DOES. And it might be something environmental. Does he customer use external speakers with it? ANy floor toys or effects units? Posssibility something about his setup that might make the amp crank out RF for all it is worth?

Did you change the cathode bypass cap? COnsider what would happen if that cap intermittently went very leaky or shorted even - no bias.

And both tubes red plating instead of just one is a clue. Not sure what it tells us, but it is a clue. For example one leaky couling cap wouldn't make both tubes melt. it would seem to be something they have in common.

[Alex R]

Yeah changed all the electros inc the bypass.

Understood about over-volted guitar amps but it was doing 380v 60mA per tube, seemed reasonable to infer that might make EL84s overheat. And it still seems to draw more current than you might expect given the B+ and the cathode resistor value, but that's just instinct.

But actually though the guy says 'the tubes glowed' only one of the JJs I put in there has browned the printing. Coupling caps... mm. I just wish it would do it once while I was looking then I'd know. I mean I could put new coupling caps in, and it would be ok, but is it fixed?

There's also this business of him wanting it louder, more punchy, like it used to be without the v drop. I'm thinking I'll take the zener out, change the coupling caps and see what it makes of that. Is that logic? or desperation...

I put the v drop zener in because of what the other tech said & what the guy asked for on the basis of that ("should we put in NOS valves, or drop the volts?". Teach me to follow my instincts.

[Alex R]

The coupling caps are those little beige rounded-off rectangular hard cast plastic things. Any experience of those throwing up problems?

[Alex R]

Know what? I'm going to turn it on and blow some hot air at those coupling caps, see if I can get the suspect one to pass DC and make me happy.

[Enzo]

It is not the voltage that makes tubes red plate, it is the power dissipation. If they don't break down from voltage by arcing or shorting, then we need to keep the current within power limits.

If I suspect a cap, I just replace it. PArts are cheaper than labor. If it fixes it, great, if not, at least you know that part is not involved. Why should I spend 15 minutes of my time to find out if a 79 cent part is intermittent? That's now a $15 cap in my estimation if I do. No, I am not just throwing parts at it and hoping. I am looking at a part that I suspect, and I am removing that doubt.

Any experience of those throwing up problems?

Doesn't matter. Even if they don't do it hardly ever, that doesn't mean it is not doing it here. And the reverse is true. Just because a part is failure prone does not mean it has failed in a particular case. One of our shop rules is never think up reasons NOT to check something. "Oh it couldn't possibly be that, they never go" is the kiss of death to troubleshooting.

And we have another rule about other techs and their diagnoses. I don't care who it is, unless they fixed it, I don't need to know their diagnosis. That takes nothing away from the tech in question. I expect any tech who follow after me to ignore my diagnoses too. The only thing I want to hear from a previous tech is what he did. If parts are removed, I don't want to find out the hard way. If he changed out some parts or altered anything, I want to know. After all, I know I am capable of making a bad solder joint or misreading a part value. Why should I assume the other guy can't do that. If he did any work, I want to know it so I can check if it is right.

[Alex R]

Oh the last tech definitely didn't do any work! Wiser than me, maybe, but less helpful, he turned the job away by blaming the voltage - his suggested cure was NOS tubes. There is, I still think, a chance that he was right, but let's not think about that now, it's painful, especially since here I am looking at the amp for the second time after doing a lot of work and charging a fair amount too, and wondering what to do this time and whether to charge the poor guy again, or whether my 'no fix no fee' rule applies.

Yep, got it about voltage not causing redplating on its own, but it brings up the wattage, and with a high current draw going on (for some reason) too, and a bigger cathode resistor still not limiting the current by much... I think 22 watts is a lot for an EL84 to dissipate, and that looked to me like the reason they overheated... at first anyway. Well I still think that was reasonable, and the cap job made sense given its age. Even after my 'fix' I was asking the EL84s to idle at 14 watts, but I would agree that many an amp runs EL84s happily at 120% of the plate dissipation or whatever, not to mention high voltage too. AC30s, Laneys, little Boogies.

I'd just like to see the fault happen once, is all, because I already fixed it once only to find it still broke. I think a little chat with the owner is in order, but first I plan 5 minutes quiet relaxation with my wife's old hairdryer on those caps! Just to know for sure what's wrong with the damn thing.

Many thanks for the advice, Enzo, it is much appreciated. I bet it's a coupling cap. Shoulda changed them the first time.

[Alex R]

Ah well, no joy with the hairdryer, so took out the zener and put in two new coupling caps. It is a nice amp, and I can see no reason why the valves should redplate now. So we'll see.

Just a couple of things I'd like to understand better:

With a big signal starting to clip on the cathode-biased power valves, you'd expect the grids to go dc negative, by about half the AC signal voltage, so the tops of the waves are about zero volts on the scope, right? (A phenomenon I have observed before.) Driven negative by... er... what process?

This amp has .01 ceramic caps running from the centre tap to the plate connections, on top of the OT. For what purpose? Controlling oscillation? Can't quite figure how - caps from plates to ground, yes. Or some other purpose?

[Steve Conner]

Yes, those cast plastic caps are probably well past their trash-by date!

Now, regarding your first question. Any attempt to drive the grid of a tube more positive than the cathode will make the grid pass heavy current to the cathode. Think of it like a diode, where the grid is the "anode".

Therefore, overdriving power tube grids will require a net DC current, which the PI can't supply, as it's connected through a coupling capacitor, and capacitors don't pass DC. Therefore, the grid drive waveform is clipped to the cathode voltage, and the coupling capacitors charge up to whatever voltage is needed to avoid the power tubes passing grid current.

The extra charge is what you see as the grids being "driven" negative, though in actual fact they're just refusing to be driven positive. It functions like an additional bias voltage, that causes crossover distortion under overdrive. Once the overdrive is gone, the caps discharge to the usual value through the grid leak resistors. This time constant can cause blocking distortion if it's too long. Read up on self-bias and blocking distortion for more information.

Regarding the second one, any small ceramic cap anywhere is probably a band-aid to stop parasitic oscillation. Designers don't usually understand the causes of parasitics too well, and just try adding caps by trial and error. Anyway, from the point of view of RF, the OT centre tap is ground, since it's bypassed to ground by the electrolytic smoothing caps.