HELP: Crack Speedball problem

[TripleM]

Hey guys,
I have built my Crack some time ago in stock form and enjoyed it a lot with my HD800s. Back in the day I ordered it with the speedball upgrade and now decided to install it (finished yesterday).
After measuring everything as in the speedball manual (which turned out fine) I started it and after the start up period (maybe 30sec) heard a loud buzz (mainly left channel). The troubleshoot guide pointed towards a lifted ground, I measured the whole ground path, no luck.
With our electronics guy at the lab I work we hooked up 300ohm resistors to both output channels to simulate a headphone, hooked the channels up to an oscilloscope and voila, hum/buzzing frequencies visible. After removing one PCB after the other and replacing them with equivalent resistors (as in the original Crack) we traced the error to the left channel small PCB of the speedball upgrade.
He didn't have the transistors at hand, only 2 LEDs, so we changed those, soldered the PCB back in and didn't see oscillations anymore. We were not sure if it was the LEDs or maybe by resoldering we managed to repair a cold solder spot or something. So I put everything back together, started some music, sounded good. Now today I listened to it again, and after some time noticed some crackling and scratching and then suddenly a super loud pop/crack in the right channel... I was scared to hell that I shot the right driver of my 800s, but so far they sound good with my other small amp (O2 amp).
But now I don't dare to hook up the Senns to the Crack again before properly solving the problem.
That's where I need your help. Does anyone have similar experiences with the speedball?
What would you recommend as next steps? Revert it to stock? Unfortunately I clipped the leads of the resistors I took off (2x 22.1kOhm 3/4W metal film, 2x 3kOhm 10W wirewound), so I can't put them back in. And I haven't found replacements here (I live in Switzerland). So that would mean getting replacements maybe from Bottlehead directly, which takes time of course.
Strangely the right channel popped loudly, not the affected left channel from before...
Maybe blown tube? Can I see that from the outside?
I hope you can help me!
Cheers
TripleM

[Chris65]

Try another set of known good tubes if you haven't already. Unless there's obvious damage, you can't tell if a tube is bad by looking at it. At least that would eliminate the tubes as the source of the problem, but if they were ok before the Speedball then most likely an issue with the boards.

[Paul Birkeland]

It sounds like you still have a cold solder joint somewhere.

How were the voltages?

[TripleM]

How were the voltages?

Hi, quick update on the situation:
I built it back to stock (was able to salvage the 3kOhm 10W resistors, leads were long enough, and found equivalents for the 22.1kOhm ones).
Resistance check (as per Crack manual) went well. Voltage check (thanks PB) was good for all except terminals 6 and 10 (right and left channel). They were not 0 but oscillated around 0 with peaks of around 15mV (also with 2 300kOhm resistors simulating headphones connected to the TRS jack). I didn't hook it up to an oscilloscope yet, but already on the DMM I could see the oscillations. Measured again the whole ground path, checked out fine.
Tomorrow I'll also bring some cheap sacrificial headphone to see if I hear those. Also I plan to hook it up to the oscilloscope again and track the oscillations and see where else they turn up (the voltages elsewhere were of course too high to see small deviations in the mV range on the DMM). I hope that will lead me to the problem. I visually checked the solder joints and resoldered some suspicious ones, no success.
Any other ideas?
Cheers from Switzerland

[Paul Birkeland]

Peaks of 15mV are quite small, and wouldn't be abnormal across those 2.49K resistors. 

15mV will not damage your headphones.

[TripleM]

Great, thanks for the quick reply! I'll try with cheap phones tomorrow and if that sounds fine hook up the Senns.
Then the next step would be to track the problem down on the pcbs.
I will check for connectivity and proper paths on them, is there a possibility to check the transistors (I assume I'll have to desolder them to check them right?) any other checks I can run on the PCBs?
Thanks a lot for the support!
Hope to get it to run properly with the speedball to enjoy some sweet sweet tunes!
Cheers

[Paul Birkeland]

I will check for connectivity and proper paths on them, is there a possibility to check the transistors (I assume I'll have to desolder them to check them right?) any other checks I can run on the PCBs?

Generally, removing the transistors will likely destroy the PCB's, and possibly the transistors themselves.

Checking the voltages is typically sufficient to provide information on the health of the transistors.

[TripleM]

Ok thanks for the info.
I checked with cheap earbuds, sounded ok, now running with the Senns for about an hour, everything fine till now, so I assume the stock build is fine, so the problem is either in the PCBs or in the connection of the PCBs to the Crack.
I'll check the leads and resistances of each PCB and also compare the left/right boards to each other as a next step. If I don't find anything obviously wrong I'll carefully solder one after the other back in and then measure the voltages with the speedball again. Unfortunately I won't be able to do that probably till Sunday.
Cheers