I just finished it with a full recap, alignment, and multiple test drives. Played absolutely great--played for an hour or so each day for several days. A nearby gent was going to pick it up tomorrow (Saturday) so I thought I would give it another test drive. It had played for over half an hour and when I went out in the garage to go to the gym it was smoking badly. The power transformer has melted tar all over it. I am getting ready to do an autopsy to find the root cause. I did notice during the restore that the B+ was about 40 volts higher than spec, but assumed that was okay. I did have two 5U4GB's installed rather than 5U4s.

1. Has anyone ever heard of 5U4GBs failing catastrophically like the ole 6X5's?
2. Anyone have a spare transformer out of a K-130?

Thanks for any experiences you can share on this.
Joe Miller

Did you exceed capacitor voltage ratings?

5U4 can't fail like the 6X5 since the 5U4 is a direct-emitting tube (it has no cathode sleeve as in the 6X5).

5U4GB has the same filament voltage/current rating as the 5U4G so I wouldn't expect the GB
to be the source of the problem (unlike, say, using a 5U4 in place of a 5Y4: the filament current of the 5U4 is 3 amps v.s. 2 amps for the 5Y4).

Filter caps were 450 volt. Others were 630 volt. I haven't pulled the main chassis yet but resistance measurements from the connector plug suggest that filters and voltage divider did not fail. Internal fault in transformer is a possibility but not sure how to prove that.

Sometimes power transformers just fail from age. There is acid in the paper used in the winding process and various factors cause shorted turns and inter-winding shorts. Years ago I moved to a seacoast town that had a lot of humidity indoors and after a year or so I started losing power transformers at the rate of about one a year. Installing a dehumidifier seems to have cured the issue.

Provided that you didn't have something short downstream, it could just be natural causes, or it has been stored somewhere humid for a long time.

If you take apart the transformer, you'll find damage from overheating, but finding the cause of the overheating might not be possible. IOW, the failure can destroy the evidence of what caused the failure.

I wonder if there is any evidence of the power output tubes overheating? If there were a failure in the bias circuit, the output tubes would draw too much current.

Definitely check everything before installing another transformer....

Looking at the circuit I can speculate as to a possible cause. I recently had a radiogram in for repair where that had happened, only it had a metal rectifier that went short through overload, that had a shorted OP Pentode.

Did you keep the Electrolytic caps you removed. C43 is a trap. I note 365V DC on the Plates of 6F6 fine they will handle 375V. What concerns me is 450V Ecaps. Where are they? Modern caps do not have a surge rating and with the voltages quoted I have serious doubt that those caps had a hope in hell of coping with that set. To that end I would be finding a way of seeing if they have punched through & shorted.

A filter rectifier feeding heater tubes is going to generate a surge voltage close to twice the running voltage prior to the heaters causing their tubes to start conducting & I would speculate that this set may even frighten the life out of 600V ones, that is why I would like to know what was in there? That thing seems to have some sort of voltage divider (candohm) which would hopefully drag down the starting voltage peak but not enough for 450V ones. I see no reference to the voltage on, or after, the rectifiers but have not traced the circuit. It is not unusual to feed OP plates off of the first filter cap (Choke input filter).

C43 is positive to chassis, not negative to chassis.

Heater & filament wires can short & the rectifier filaments should not ground.

Bit of homework here.

The last radio (GE) l had on the bench had a partial short in the candohm resistor, took a while to troubleshoot. You might find the same problem...or some other. If that transformer is toast, my usual recommendation is to re-wind it. However, l have that same mini-chassis stashed away if you need it. Hard to find transformers with that extra 25-volt winding, that's why it was set aside. Craig

Todays higher line voltages probably played a role in the destruction of the power transformer, too.

Clearly that is a transformer handling a fair bit of wattage and clearly if there was a rewind the primary has to be changed. Fortunately many of ours in the early days catered for the wide range of voltages then, by being tapped to handle between 200 to 260V.

For the post mortem, I would visually look for obvious wire shorts in all of the windings. One notes that debris & solder blobs are a risk. I would then remove tubes & lamps and disconnect the transformer. Perhaps even remove it. If there was a major wiring error, one would not expect it all to work

Then you can check for a short on the heaters & lamps. Lamp sockets can short,some designs can have terminals that come loose and move. I have had 5V filament wires short as soon as DC was developed. Attempting to test lines with electrolytic caps on them with an ohm meter will end badly. I have actually used a simple battery & lamp initially. The lamp will glow & act as a barretter if there is a dead short.

If I suspect a punched through electrolytic on "B", I will disconnect all of them. I have a reformer on the end of the Bench Power supply which allows me to step voltages. With it I can gradually increase the voltage on a cap up to 450V. The duds will normally have gone short long before that. & the regulator chip will "lock up". The reformer can also be used to power "B" (without tubes & electrolytic caps) to see what current is drawn. The candohm in yours will if its across plus & minus. But that can be calculated & measured. Where there is no candohm there should be no current draw. If there is current draw Ohms law would enable you to gauge where it is by the value of the resistance unless its a dead short before any. A clue in itself.

Marc

I pulled the chassis this morning and found R28 (133 ohm) resistor burnt open (see picture). When I restored the set, R28 which was part of the candohm was open so I installed a 2 watt resistor as a replacement (that is the one that opened up). Now I see that R28 sees 24 volts in normal operation which means it consumes a bit over 4 watts. I should have replaced it with an 8 or 10 watt resistor. That explains why it overheated but not why the transformer melted down. No other components in the tuner chassis appear to have been impacted.

Will do more troubleshooting tomorrow.

Here is the schematic for it:
http://www.nostalgiaair.org/pagesbymode ... 015505.pdf

You may want to take a closer look at the rest of the candohm as well.

I suspect that that just might be the back bias resistor and likely carries all of the cathode current so "Current squared by resistance applies". This is a resistor that I never recommend "beefing up": I would rather see it sail with the wind & burn rather than the transformer. The fall down of course is having an electrolytic across it often around 40V. And that's the one that should be positive to chassis.

There is a possible scenario that there is an overload from a shorted tube/ wire /solder blob / miss wire and that fried the resistor? That would cause a cap CT to B+ to cop a massive "no load" surge voltage and punch through then present as a short, frying the secondary.

Marc

I'm only using the candohm ground terminal.
Under normal operation that resistor would consume over 4 watts. I had incorrectly installed a 2 watter. My guess is that it overheated after playing for an hour or so.
But that doesn't explain the fried power transformer.
Still thinking about it.

I would not over think it. My speculation is based on the first filter letting go. As once the "back bias" lets go the first filter is likely to be the only one with a live negative.

If all cathode current goes through those resistors it's 170mA @ 24V 140 ohms. so indeed a bit over 4Watt Do you have a way of testing / electrolytic caps at working, or surge voltage?

Marc

No, don't have a way to test caps at voltage. I think I will replace all filter caps and candohm resistors with proper wattage resistors.
I still don't see how the open resistor would cause the power transformer too melt down though.

One can test an electrolytic for leakage by putting a series resistance in series with the cap and measure the voltage drop across it. At 250V stolen from another set through a 50K resistor, this would limit the current to 5mA and not unduly upset the test radio.

A good electrolytic should not draw that much current once charged. As the cap charges the voltage across the resistor falls and you use "Ohms law" to relate the voltage to the resistor V/R to give you current; a dead short would be 250V. Beware of a charged cap and any wiring to it (perhaps tack them in).

The resistor was (If I am right) responsible for causing an underrated cap to fail shorted: And I am backing the first filter cap. Do not use 450V ones. I would still like to know the ratings on the old ones. Eg 350V WKG 600V SV or 600V PV. As before most of the new ones do not quote a surge voltage, albeit I think some of Mouser's do?

Marc

IMHO failure will found in the power output, gassy 6F6's that drew too much plate current. The condition could have been triggered by the use of incorrect coupling and output bypass caps.

YMMV

Chas

All of the filter caps were rated 450 VDC. They were can capacitors. I am not interested in trying to test them. None of them show any visible signs of stress. I will replace them all just to be sure.

The resistor that failed open was the one that develops the -24V bias for the 6F6 output tubes. When it opened, it removed the return path for B+ so I don't see how its failure could have caused any overheating problem since there was no load on the power transformer from B+. So, I am thinking that the following are possibilities:

1. Lamp holder may have shorted?
2. Motorized tuning shouldn't have been the problem since auto tuning was not on. (this is the 25 volt winding on the power transformer)
3. 5U4 Plate to filament short. (Haven't tested these two tubes yet) Does this ever happen?

Joe

I can't tell for certain if connections are correct in how the new (failed) R28 replaced the open section. I would suggest this part of the radio may worth another look. I agree that taking the entire candohm out of the circuit and replace with new resistors on a new terminal strip would be a good approach.