Speedball, checks pass, but smoke + L channel pop after 5 mins

[cabcabaret]

Hi,

I just finished installing the Speedball upgrade and all the voltage checks were fine.

I plugged in some headphones and listened for 10 minutes or so and heard a crackly pop sound proceeded by little/no sound in the left channel. I immediately turned everything off, and tested the headphones in another amp and they're fine.

I then switched it on again to confirm the tubes lit up, they do.

Finally I did the voltage checks. THe first thing I noticed with it flipped over was that only one LED was on initially, the one on the inside of the B+ side of the large v1.1 board. Gradually this LED dims and the others light up, albeit dimmer than before the pop.

I only managed to read 0A & 1A (both 40V) before I notice the cap on T14 & 15 is smoking! I turned it off immediately and haven't touched it since.

I think I'll try replacing the cap obviously, is it possible that it was just a bad cap? I've had the crack sans speedball for 3 years. Is there anything else I should do to avoid endlessly blowing caps or the rest of the amp?

Thanks

[Paul Birkeland]

Can you post a photo of your amp?  Are you sure it's the cap that's smoking and not the 270 Ohm resistor?


I strongly suspect that your TIP50 transistors are not properly mounted to the heatsinks, as this would cause the low voltage that you're experiencing, and that low voltage will overheat the 270 Ohm resistor around T14/T15.

I would not recommend turning the amplifier back on.

-PB

[cabcabaret]

Thanks very much for the reply, attached are some pics, let me know if you need more.

After reading your post I had a closer look and there is a small hole burnt into the 270kOhm resistor, you're right.

I screwed the heatsinks quite tightly to the transistors, and was careful to follow the installation order given the warning in the manual.  Your explanation makes sense, but I'd be surprised if they weren't tight enough.

Do I just need to replace the 270kOhm resistor and recheck everything?

Thanks

[Paul Birkeland]

It's the 270 Ohm resistor that smokes, not the 270K.  If you tighten the TIP50 transistors down too hard, you may damage the insulating washers and short out the collector of the TIP50 to the heatsink.  You can take your multimeter and set it to beep for continuity, then measure the resistance between the metal tab and heatsink of each TIP50.

[cabcabaret]

Ah yes that was a typo. i checked the heatsink-transistor connection using your multimetre method and found no short. I didn't tighten them a lot, just light wrist tightness to make then firm and secure.

If the transistors are overheating would they show the correct voltages at first (as they were for me)? In which case, are the new low voltages due to the blown resistor or the transistors?

If I replace the resistor and do nothing else, wont it just do it again?

Thanks again

[Paul Birkeland]

The resistor doesn't technically need to be replaced.

Try turning the amp on with just the 12AU7 installed and no 6080, then measure the voltages on the board up front.

-PB

[cabcabaret]

Here are the readings on the front board without the 6080 attached.

0A = 70 V
1A = 225 V
B-A/B = 0 V
1B = 225 V
0B = 70 V


Attached is a pic of the LEDs which are on if that's relevant.

How does it look?

[Paul Birkeland]

That LED on the big board shouldn't be lit up with no 6080 in the socket. Can you post a photo of the soldering on the bottom side of the big board?

[cabcabaret]

Sure, although the B+ cable broke when I lifted it of the standoffs.

The big board is sitting right on top of the caps, as I can't bend them any lower as they are.

[Paul Birkeland]

I'd replace the 100uf caps and leave a little lead length. I suspect you are grounding out one LED to the metal body of the cap.

[cabcabaret]

I tested that (I think) by repeating the sans 6080 check with the big board raised so that it's not touching the caps at all and the same LEDs lit up.

So I can redo the caps, but I think that means it won't solve the issue?

[Paul Birkeland]

Swap the white wires leaving the big PC board and see if the same LED stays lit with no 6080.

[cabcabaret]

That's a clever idea.

I switched them at the 7 & 9 terminals rather than the PCB ends. Weirdly it made no difference, the same LED lights up, while the others on the big PCB are unlit. I haven't closed the curtains to check if they are dimly lit.

All of the LEDs on the small PCB and the output tube socket are lit.

[Paul Birkeland]

I'd replace the 2N2222 and Tip50 on that end of the board. One (or both) probably touched that 100uF cap and shorted out.

[cabcabaret]

OK sounds good, is there a way to check if they were shorted?

Are all 2N2222A and Tip50 transistors created equal or should get certain brands? I'm in the UK so it's probably better to source parts here.

One other thing, what should the resistance across each LED be?

I measured all of the PCB LEDs and found that they all show inf resistance across them apart from the LED adjacent to the one lighting up. Is that meaningful?