Flames in the power section - HELP!

[Tom Phillips]

Tom,
...all the primary wires at the VAC selector's end are burned...

That's unusual. There is something strange going on in your amp. It doesn't make sense that the primary lines could carry enough current to burst into flames if they were in series with a 2 A fuse. (even with 240V line voltage) Proceed with caution of course.
Lately I sense a lot of extream safety advice being dispensed. Almost like corporate lawers were writting the response. Remember I said "What I would do is..." I would also bring up the voltage slowly with a variac and be monitoring the primary current draw. but, again, that's me. You do what you feel comfortable with.
If all else fails, you could switch to acoustic guitar.
Regards,
Tom

[Steve Conner]

Reading between the lines, it sounds like the voltage selector switch caught fire, not the transformer. In my experience transformers usually smoke, bubble, and stink of burning varnish for several minutes before they catch fire, and then stink for ever after.

If these voltage selector switches are anything like what I remember, they are somewhat poor quality and wouldn't pass electrical safety regulations nowadays. I think if it was dirty or damp it could easily have shorted. Inspect the switch for damage and burning.

Disconnect all the wires from the switch, and wire the transformer permanently for 220V for test purposes. Power up again and see if the amp works. If your voltages are missing or abnormal, and/or the transformer smokes again, then maybe it is time to go shopping for a new trannie. But I wouldn't go buying one just yet.

[Alex/Tubewonder]

Well, reading the lines not between we are told that:

amp was left on idling fro about 20 min, at some point during this time fire started.
All wires between voltage selector and transformer burnt.
100 Ohm heater balancing resistors burned
Traces of something wrong found around one of power tubes.
Transformer hot

So, apply some "what happened and how" CSI electrical forensics here.
What could burn heater balancing resistors? A faulty intermittent mains voltage selector...not likely. A short between pins 2 & 3 on power tube i.e. plate and heater is a good suspect.

It could be internal in the tube or external on the socket due to flash over or bent solder tags. This effectively loads B+ plate supply with Output transformers 1 half of primary winding whose DC resistance is somewhere around 50 Ohm + 50 Ohm heater resistors (both in parallel) PT and rectifier would want to output in excess of 400VDC, this would draw want to draw 4-5A, fortunately there's the rectifier tube that will limit available current to some 300mA. This results in ca 2.5W dissipation in each 100 Ohm resistor and these are rated what? 1/5-1W so they heat up and burn into a short.
Perhaps not enough to cause fire on PT primary side but enough for serious overheating. It most probably didn't go down this way.

What about direct short between PT HV secondary and heater winding then? This would load 345Vac with 100 Ohm heater balancing resistors which the 345V will see as being parallel i.e. 50 Ohm drawing 7 amps. OOPS over 2000W power draw, resistors go up in smoke explosively. Now since it took some time for the whole thing to heat up and burn out I would speculate that a short between HV and heater winding developed over time beginning as resistive leakage between windings and progressing with rising temperature.
A 2A primary fuse at 230V ac would allow for over 450VA power draw while the transformer is dimensioned for 135VA at best, in reality it is undersized and dimensioned for 110VA continuous operation. We have one monster sob overload of 400% of rated power - no make your own calculation of temperature rise in the transformer.
Take also into account that a slow blow fuse will withstand 1.5 times rated current for up to 60 minutes (it did blow eventually acc. to Carlo) and we arrive to some interesting conclusions regarding our worst case scenario.

Guys, I understand reluctance to replacing the PT. But the methods proposed to test its health are not safe in the long run. One should conduct full test to identify shorts between turns in any winding, shorts between windings and HiPot test to check if all dielectric isolations are intact. Less than that is not enough and not safe.

I hope to have demonstrated the importance of fault analysis to identify the primary source of fault before judging suitability of a safety critical component such as PT for further use.

BTW, the mains fuse on that amp at 230Vac should be 0.8A followed by NTC resistor as inrush current limiter.

[Steve Conner]

Your explanation doesn't explain why the wires at the voltage selector switch caught on fire. Unless they were situated near the 100 ohm resistors, and were just set on fire by the resistors when they blew up.

[Alex/Tubewonder]

Your explanation doesn't explain why the wires at the voltage selector switch caught on fire. Unless they were situated near the 100 ohm resistors, and were just set on fire by the resistors when they blew up.

You're quite right. Based on the presented evidence I can't say for certain, only speculate.
Heater 100 Ohm resistors are wired from pilot lamp to ground, well away from PTs primary wiring so they had no way of setting that on fire.

There could have been several factors setting it off simultaneously or in sequence.
Carlo doesn't say if PT wires are burned only at the voltage selector or the whole length. If the whole length than we know copper to be very good heat conductor, the temperatures that might have developed inside the PT make it possible that heat transfer along the wires was enough to ignite the sleeves.
Perhaps at the same time the voltage selector was arcing internally, that would develop enough heat to set fire to the wiring at that end.
This combined with possible fault mechanisms I described above could have been what happened.
It all goes to importance of failure analysis.

Any possibility for photos, Carlo? Pics of taken apart voltage selector too.

[slidincharlie (Carlo P)]

I attach three photos here.

Alex's summary of what happened is correct.

WHAT WE DO KNOW:
- amp was left on idling for about 20 min (maybe less), at some point during this time fire started.
- all wires between voltage selector and transformer burnt only close to the VAC selector: the rest of the wires is intact.
- 100 Ohm heater balancing resistors burned
- traces of something wrong found around one of power tubes.
- transformer hot

WHAT WE DO NOT KNOW:
- where and in which order the flames started first.

SOME MORE CLUES:
- photo 1: the heater wires above the two 100 ohm resistors are burned. This suggests that a flame developed there (but I was not there to see it), and it was not only blown resistors, like when I short the heater and plate pins with a probe in a power tube.
- photo 2: all the wires connected to the VAC selector are burned only at the selector's end, but the last 1 millimeter or so soldered to the selector's lugs is NOT burned. The selector itself is externally in perfect condition, no sign of fire or smoke. It looks like those wires were burned from below (pins 2 and 3 of V8?)
- photo 3: the really charred parts are only the hetaer and plate wire ends on pin 2 and 3 of V8. The plate wire even detached from pin 3 (as a consequence of fire, I suppose).

A flame starts to burn in the very point where it develops, and tend to burn in an upward direction, right? So it seems that Alex's hypothesis of a flame developing from shorted pins 2 and 3 on V8 could be correct.

I still wonder why and how those pins shorted?
Or could it be a short in the tube to trigger everything?
The tubes are UOS RCA 6L6GC that I have used only once in this amp.

I have no way to check the tube, btw.

I am very tempted at this stage to follow Tom's advice to unhook all the secondaries and power the amp up to see if the tranny gets hot, smells bad or make any odd noise. Is this enough to tell if the PT is okay (although maybe in need of replacement due to the thermal stress), and to find the culprit somewhere else?
In other words, are there sufficient clues to tell that the trouble started from shorting wires at the tube's end and not from the tranny itself?

[Steve Conner]

OK, so it sounds like the power tube socket arced between the heater and the plate. The arc on the power tube socket burnt the voltage selector wires, and the fault current flowing from B+ to the heater circuit burnt the 100 ohm resistors.

In this case, the power tube socket may well be trashed and need replacement, but the transformer should be OK. Does that sound reasonable?

[Tom Phillips]

I agree with Steve's assessment. I have seen similar damage many times. I clean up the wires, locate and repair the original fault and re-use the PT.
It's very helpful that you posted the photos. I especially like that you know how to take good, clear photos. My oppinion is that this all started with a plate to heater short in the power tube. Then the heater wire connection at the tube socket failed, went high resistance and had a melt down. At the same time, the heater balance resistors burnt up creating heat and smoke that damaged the heater wires at the pilot light socket. The higher than necessary fuse value contributed to the severity of the problem.

Clean the smoke damage off the wires and things will look a lot better. (I use alcohol) Then you can cut back to remove any damaged insulation and re-solder the connections. There was talk of a new tube socket but after looking at the photos, I think there's a good chance that it's serviceable.

Looks like you are missing the center key pin on your rectifier tube. Are you aware of that?

Cheers,
Tom

[slidincharlie (Carlo P)]

Thank you guys.
First of all, after cleaning and checking all the connections, I'll check the PT for odd behaviour.

Looks like you are missing the center key pin on your rectifier tube. Are you aware of that?

Yes, I know. Is it an issue, as long as I insert the tube correctly?

[Alex/Tubewonder]

Above explanations are correct, arcing between plat and heater connection on the tube socket started it. heater resistors exploded while you were not there or you would have heard it.
So you left an open amp chassis unattended where children had access to it

But be careful with that PT, as it had overheated already before the fireworks incident the isolation can have weakened.

Good luck

[Tom Phillips]

Quote: Originally Posted by Tom Phillips
Looks like you are missing the center key pin on your rectifier tube. Are you aware of that?

Yes, I know. Is it an issue, as long as I insert the tube correctly?

Nope. Not an issue if the tube is inserted correctly. Just wanted to make sure that you were aware.
Regards,
Tom

[Rob Mercure]

Hey Carlo,

I've also seen identical damage in Fender amps. Where the plate (3) arced to heater pin 2 the phenolic material will create a "carbon track" - essentially a carbon resistor which can conduct enough current recreate the problem in the future. If you've got a Dremel moto tool - or other hand held rotary device - use one of the bonded "cut off" wheels to carefully cut a slot between the two socket pins, and any other suspect ones (I tend to hit them all after a arc problem) all the way down to uncarbonized phenolic. Make sure that no black or scorched materials is left to create a resistor. Afterward you can leave the slots or fill them with super glue or hot melt glue or silicone - I like to refill them just to keep any loose conductive material out.

I've also seen a similar problem on cathode biased amps where pin 8 shorts to pin 7 and the heater ground reference resistor(s) become the "cathode resistor" and burn up.

Rob

[slidincharlie (Carlo P)]

So you left an open amp chassis unattended where children had access to it

My kids know that they never have to touch or stay too close to an open chassis but yes, you're right, it's dangerous conduct. My bad.

Thank you guys for the diagnosis and all the precious advice.
I'll start the repair job is in the next few days and report especially the PT behaviour.