Very old Gibson GA20 problem - FIREWORKS!

[EFK]

OK not really fireworks, more like little sparks. Bear with me here, this is hard to describe.

This is a very old GA20 probably 1952-53. All octal preamp. This amp has a pretty pronounced hum which I have not been able to eliminate. I have suspected recently that the hum might be transformer induced and is being picked up through the octal pre tubes, which all seem to be a bit microphonic. The OT is on the speaker, but the PT is on the chassis. So, this morning, I unbolted the PT from the chassis and was able to get it cleared off by about 1/2" and suspend it there so I could try the amp w/o the PT actually touching the chassis; my goal was to see if the hum was indeed PT mechanical hum vibrating the chassis and being picked up (and amplified) through the pre tubes. I had to keep one PT screw installed (not attached to the PT) as it is a mount point for a ground tag. Anyway, the amp fired up and seemed to definitely have less hum. As I'm moving things around a little, one of the PT mounting tabs came in contact with that screw and there was a very loud burst of static through the speakers. Also, I could actually see some little tiny arcing sparks like you'd see on maybe a 9 volt battery. I hooked my meter up to the PT tab and kept bumping it into the screw - I was getting voltage readings spiking up to 2-3 volts DC.

Is this something normal w/ a simple explanation that I have never seen before? Or is it telling me something about the condition of either the amp grounds or the PT?

[EFK]

Holy crap - I just had a thought and went back and checked AC too. With the transformer isolated from the chassis, if I clip the meter on the transformer sheet metal (mounting foot/tab) and clip the ground on the chassis, I get 83VAC !!!!!!!

Is something old and leaky going on here? W/ the PT isolated, the hum is practically gone, so I'm guessing that AC is leaking out of the PT somehow and contaminating my ground when it's mounted to the chassis?

[pdf64]

The power transformer is bad - the insulation is failing. It must be replaced, as this failure mode may have fatal consequences.
The only way to avoid replacement may be to run the line supply to it via a 1:1 isolation transformer, this will avoid the risk of directly exposing the user to the line supply. But as you've found, there may well still be functional issues (ie hum), which may get worse as the insulation breaks down more with time/use etc.
But Buffalo amps, Mercury Magnetics etc could supply an equivilant transformer, or get the original rewound.

[EFK]

I kind of had a feeling that was it after reading AC on the transformer housing! I'm not going to screw around with it - I don't feel like croaking just yet. I know Mercury sells them for the price of gold but frankly, the tube line up is pretty much the same as some of the pre-5E3 deluxes and I'll bet any lower voltage deluxe PT would work. It's no collector piece so I don't care if I have to drill a few new mounting holes.

Thanks for checking in! I appreciate it.

[pdf64]

Yes, the 'lower than 5E3' B+ voltage is important, especially with modern high line supply voltage levels - the overdrive tone seems softer with lower B+, in my experience.
A possible problem may be that 5E3 PTs are the 'drop-through' style, whereas my GA20 has 2 end bell covers and sits outside that chassis, stand up style. So you might need to cut a big hole.

[EFK]

I wouldn't cut a big hole like that, it's easier just to make stand up brackets. I'm waiting for some pricing on rewind, although it's probably cheaper to simply find a replacement. BTW, anyone know what the secondary voltages on these are supposed to be? I'll have to do some digging as I'm not turning it back on again. I know the voltages through the amp were lower than a deluxe iirc. I'm going to guess these are maybe something like 250/0/250?

[pdf64]

Sorry I meant to check this yesterday on my 57 GA20, will try this evening.

[EFK]

Mercury quoted me $325 for a rewind (!!!) but they have replacements in the $150 range. However, I got the specs from them (they have about 3 or 4 different versions) and they are all from late 50s or early 60s versions, some with trem, which had higher voltages. The real early ones - the brown leather-looking tolex w./ the G in the middle of the square centered in front of the speaker cone - had a much lower voltage PT, something like maybe 250/250, or maybe 275/275 - something like that.

[pdf64]

I checked my notebook and mine runs at 320-0-320V with the spec 200ohm bias resistor, though I fitted a 270 ohm cathode resistor and it raises the winding voltage to 330-0-330V, giving a B+ of about 360Vdc at the 5Y3 output.
The chassis has cut outs for octal pre-amp tubes, though there are octal to B9A adaptor plates, to allow 9 pin tubes to be used. I'm fairly sure these are original fittings.

[EFK]

I've read about other people finding those adapter plates as well so I'm sure they are original. There must have been a period when they were switching over to 9 pin sockets and they still had a supply of octal punched chassis. I got specs from mercury, they make about 3 or 4 different PT for these, the lowest is 300/300 and one is up at over 350/350. This amp I have has a 1952 dated P12Q, smooth cone, and the voltages at the tubes as I recall were very low. I might be tempted to think that it was due to the failing insulation, but I've seen posts from others w/ these real early ones that had comparable low voltages. I think it was the Buffalo amps guy that used to suggest that anyone who had the real low voltage PT could get a better sound by swapping out the big 10K power resistor for something around 2K-3K. I tried that and it did sound louder as well as better, although the loud hum from the PT got worse as well!