Fender Twin possible output transformer fault?

[MWJB]

The OT has a turns ratio, if we can determine this and ensure that each end of the primary (to CT) has the same ratio, we can tell if the OT is good or not. This test can be done with the OT in situ, just with the amp disconnected from wall AC & the speaker & power tubes pulled.

Let's say you apply 0.5VAC to the speaker jack socket (measured with a DMM), you should then measure ~10.9VAC from end to end of the OT primary.

10.9/0.5 = 21.8 (21.8 is the "turns ratio")
21.8*21.8 = 475
474*4ohms speaker load = 1900 (primary Z)

A Fender Twin should have a primary Z of perhaps 1800 to 2100ohms.

Of that 10.9VAC (+/- 0.5VAC) you should see half the AC from each end to the centre tap. The dcR test can show a slight mismatch due to the wire length difference from one end of the primary to the other, but the VAC test shows us the number of turns, which should be the same from each end of the primary to the CT.

[MWJB]

"It did give me the idea to set up a test rig to put a variable AC source on the secondary and monitor the primary through a load resistor/spark gap/HV probe and see at what point the transformer fails." It would probably cope with <800VAC at the primary quite happily...this does sound mor lik a destruction test than a diagnostic test.

[Steve Conner]

FWIW, violent arcing inside a tube (that follows the socket, not the tube) sounds like a RF parasitic oscillation problem. Are all grid stopper resistors present and correct on the power tubes?

But I'd also try swapping out the tube's parallel mate with another known good one, in case it's going bad and turning both tubes fully on (which it can do because their control grids are connected together)

[MWJB]

Or it's an arcing socket, perhaps carbon deposits were left on the socket during the initial failure. If it arcs with a known good tube, before reaching operating B+ I'd change the socket.

[Steve Conner]

Yeah, good call, if the amp was run without a speaker load, that can arc the sockets. The damage might be visible, or the arcing might be visible in a darkened room.

[Mick Bailey]

The AC test showed no problems - 5.64v + 5.63v on the primary from a 0.5v 1khz sine signal on the secondary. Grid stoppers fine, but changed anyway. New socket. Spare valves installed (previously tested in another amp). I scoped the amp at 300v B+ and can't see any oscillation.

On power up the same tube arcs internally - its parallel does to a much lesser extent. No arcing on the socket.

It just occurred to me to switch the blue/brown leads and see if the fault moves to the other pair.......

[Steve Conner]

That'll make it oscillate, unless you also pull the PI tube or do something else to break the feedback loop.

[Mick Bailey]

Just switched the leads and removed the PI - nothing else. The fault didn't fully move, in that the violent arcing didn't occur within the tubes. However, there is clearly a problem as there is a bright purple glow filling the two tubes that were previously OK, which wasn't there before the switch - then the HT fuse blows. The original faulty side looks fine. The fuse blows too quickly to get a voltage reading, but again this happens around 330v.

Next stage is to swap out the OPT. If this works then a post-mortem on the old OPT may reveal what was wrong.

[MWJB]

Why are you so sure that it's the OT...even if the OT was shorted, why would you expect that arcing to happen at a specific B+ & & a B+ less than full power. The amp does not "work properly at 330vdc"....because it is supposed to work at over 450vdc. Guys who have brought me amps with shorted OT primaries only did so because of the change in sound at normal operating voltages.

At <330vdc on the B+ what does the negative voltages at pins 5 show? Should be around 3/4 of negative dc shown on the schem. Purple glow (if severe, not normal blue/purple glow) might be indicative of problems with bias voltage?

You say you have tested the power supply caps for shorts...how old are they, perhaps it would make sense to change them (& the bias cap)? Rectifier diode good? Hell, it's only a 2p 1N4007...& takes 2 seconds to change. I'd change the B+ diodes too.

What hapens if you just power up with 1 tube & try it in each socket one, by one?

[Mick Bailey]

I didn't say it worked properly, I said "All is fine below a b+ of 300v". I've monitored the B+, screen and bias voltages on the tube up to a B+ of 330v until the arcing occurs. Also the full B+ holds up all day long running into a resistive load (at least until the oil gets too hot for comfort) and with a full supply voltage. Ripple is low, filter/bias caps ESR low.

I don't think it is a shorted OT. I know what that sounds like (in fact I have a Laney at the moment with this very fault) - to refer back to my original post; "My thought on this is that one half of the OPT may be going OC under load and generating a flyback voltage of some KV". The reason for this theory is that I'd seen something similar on a power supply for a 1940s Muirhead fax due to a transformer with a break in the PT secondary winding. Under load the open circuit was causing a rapid make-break effect and boosting the B+. It tested fine on the bench. When it got rewound there was a break in the secondary winding. I wondered if a similar thing was happening with the B+ across the OT.

I've just tried the amp with a tube on each side and it doesn't arc (either with the inside pair or outside pair). It only happens with a full compliment. Tried groove tubes, Ruby and EH.

Now, it seems to me that a set of dummy load 'tubes' could be knocked up on octal bases to test such things. Wouldn't have to worry about frying a set of good tubes. I already have a standardised tube used to calibrate my tube tester. Just a bigger version required. Anyone done this?

[Steve Conner]

These 4 month old screen resistors... Are they the right values and wired up correctly?

If it's not that, then I don't know what else it could be except the original idea of a faulty OT.

[Bruce / Mission Amps]

Try that again with reversing the primary leads but disconnect and NFB from the OT secondary too.
Too bad you don't have an extra 100 watt OT to just jumper in.... most of us who do this for a living probably have lots of extra trannys for a 5 min tests like this... sorry, it's just hard to explain waway a failure mode without simple tools and parts on hand to swap out.

[MWJB]

So it was a different Mick Bailey who said "The amp words perfectly up to a B+ of 330v with the power tubes fitted"...? ;-)

What happens to screen, & bias voltage at 330vdc? In fact what are they with just 2x6L6s fitted at full B+? You say you have monitored them ...why keep them a secret.

I'm all for subbing another OT (even if a 40W item for 2x6L6) to see if the arcing at idle is OT related...but I would generally exhast the more obvious lines of enquiry before removing a customer's OT.

If the OT is going OC with 4 output tubes installed, but not with 2, what does it with 4 at reduced idle current (in fact what is the idle curent? We don't seem to have any figures).

"Ripple is low, filter/bias caps ESR low." that's not what I asked, what age are the filter caps, how do you know it wasn't a cap failure that cused the initial failure? These amps are 20yrs old or so? Have you observed the caps/circuit in a darkened room to see if you can detect any arcs elsewhere (PO, intermittent shorts....any number of failures could cause a pop, this pop would be reflected by flashes in the tubes, which you have referred to as purple...serious arcs may be white/yellow/bright pale blue).

[Mick Bailey]

Oops...

Yes I did say that. My apologies.

330v is the tipping point so the following voltages are at 300v B+ with 4 tubes;

Bias -38.7v
Screen 299v
Idle current 230ma/240v supply

Currently have the amp running with a spare transformer. I've got quite a few OTs but most are from scrap equipment and have not been tested so can cause more problems than they solve. In the end I used a new 100VA UK mains transformer (120v-0-120v/6v-0-6v) as a temporary measure just to eliminate the OT. Its been powered on and off for a day now on a dummy load with no arcing. With this arrangement B+ is 458v, screens 456v, bias -59.1v.

Screen resistors are 3w 470R, correctly wired.

[Mick Bailey]

I rigged up some load resistors to substitute for the tubes and reconnected the original OT. At 330v B+ the faulty side jumps to over 1kv before blowing the HT fuse. Looks like it's behaving like a fluorescent tube starter circuit. Or at least it was - It's now permanently OC on that side. Good.