Thanks for your take on this - I was basing the PCB thing based on this video, where somebody is applying heat and seeing the Resistance drop across a certain spot in the amp:
The internet seems to be full of the claims that the PCB material actually is faulty, and applying heat makes the substrate slightly conductive. I have NO idea however if that is true or people are just full of it. Marshall support just gave me a generic reply as in "Hi, only solution is replacing it, but boards are not prodfuced, so nothing we can do, bye!"
I have applied a fix shared in this video on a TSL100 and just at this moment watching if the bias is rising. so far, within 30 minutes of activity, the bias has risen from ~75 to 85 per side...
ANYHOW, this a Santiago Alvarez AMA thread, and I don't want to hijack your thread for this, as you had no involvement on the TSL100. IF you have any other info which might be useful, I'd be super grateful though and then leave it as is afterwards :-)
Hey, no "hijacking" any thread at all, that's why is called 'ask me anything', no 'only write here to tell me how great I am!' (you can do that too btw
)
ok, jokes apart, DSL, TSL... what I know, and this is based on my experience and not on what I read out there, there are a couple of fundamental issues with that amp. One is that the bias circuit is designed with some capacitors in series with the bias winding so you depend on those electrolytic capacitors to have your bias voltage. Electrolytics are well known for having huge tolerances, extremely poor temperature stability, they dry over time and, in general, don't like to be stressed for long time. Second issue seems to be that the clearances between pcb tracks aren't as good as they should be, which usually isn't a problem in itself but when you add heat issues it doesn't help
Unfortunately your DSL/TSL bias is made that way and over time you will suffer those variations as 'drifts' and this is a design issue. So what happens when your bias drifts to positive or fails: the amp becomes too hot because the output tubes start drawing more and more current. If you are lucky, you'll burn fuses but if not, or that takes long what happens is that you start stressing the pcb in this case. When the pcb tracks are too thin or are too close to the hot spots of the power tubes, or any combination or those factors, you will "cook" the pcb, which is made of some sort of substrate, probably fibergalss and some sort of polimer to glue all together. When you overcook it, it carbonizes and becomes conductive. Is this that the pcb material is bad?, not at all, is that you are using it under conditions it hasn't been designed for and, over time, it fails.
just ask yourself why none of any other amps and any other electronics fail: just because they are not designed like that, they have a 'normal' bias circuit that is reliable and stable, it should actually be the latest circuit to fail in the whole amp. Now don't ask why Marshall started using this in the early 90s, no idea but my 6100 failed so many times that I ended replacing the bias with a 'plug-in' circuit just to make sure it didn't happen again. The JCM900s seems to be more reliable in this sense.
You know, the main issue with this internet thing, is that people read that "bias drift" thingy and then everything becomes a "bias drift". I have had few people over the years claiming that their amp had a bias drift and it was a power tube failure that had taken a coupling capacitor with it, making the amp impossible to bias. Is this a bias drift?, no, it is a power tube failure with colaterals. Like if you blow a car tyre, crash and you blame the steering system.
long brick of test again, hope you made it to the car crash bit!
