Marshall TSL:100 bias drift problem

[blindboybenton]

Good Morning all

Ive had a TSL100 for a service, Blown pair of Tubes, Replaced and re biased all seems ok, Then i noticed some bias drift.
Decided to to the mod where you cut around pin 5 and wire resistors in directly to the tube base etc
See below

The Marshall TSL122 TSL100 thermal bias drift repair page did all this ,as ive done below with working results
TSL100___TSL122_Valve_PCB__60_.pdf

After about 20 mins again started to drift, stuck two DMM on and watched to see what is drifting. ( without tubes )
V5-6 drifting V7-8 are stable
So i changed all components in the bias circuit trim pots and all resistors and caps
Still the same. Used the heat gun and freezer spray found by spraying around the PI tube area voltage would drop
Changed the 82k 100k and 2 x 680K both 22nf, Removed the PI tube voltage stable, Swapped the tube still drifting

Is there any other known issues with the board conductivity

any help would gratfull, I know about the replacement board, but at this present time i need to repair if possible

BBB

[MarkusBass]

I think that if you search in google for "marshall tsl 100 bias drift" phrase, you will get a lot of results. It seems to me that the topic was mentined here lately.

Mark

[Leo_Gnardo]

I think that if you search in google for "marshall tsl 100 bias drift" phrase, you will get a lot of results. It seems to me that the topic was mentined here lately.

Mark

Indeed it was, here's your link bbb:

The Marshall TSL122 TSL100 thermal bias drift repair page

I've done this at least half a dozen times. The neat but time consuming way - remove output tube sockets, drill away board at pin 5, then replace sockets.

Or if you have the patience & customer has the dosh, you can replace the whole board.

[blindboybenton]

Thanks But ive done that mod that is stated in the link you posted. I have not removed the tube bases and cut the board away and point to point wired them

BBB

[drewl]

I've repaired many of these in the past by isolating the socket pins but recently had one that was breaking down around the phase inverter and all the way back to V2!

I imagine older units will have the board fail in many places.

The only way to fix that last one was to replace the board.

[blindboybenton]

Im the same Drewl. done loads of them with good results, But this one is i think is a new board. Its a local guitar shops customer and Money is an issue
BBB

[Mick Bailey]

Removed the PI tube voltage stable....

Am I reading this correctly - removing the PI 'fixes' the problem?

[blindboybenton]

yes by removing the pi tube the bias is stable. Changed 82k and 100k anode and 2 x 680k + 22nf decoups

[nickb]

The failure process is a consequence of conductive anodic filaments, I believe. These are copper filaments that grow primarily inside the board from one a positive plated through hole to another less positive one. Factors affecting the time to failure are the FR4 material, spacing, voltage and the temperatures involved in processing on the board with the higher temperature of lead-free soldering being a particular issue.

What little I know about this, I got from this the report below.

An interesting graph of page 34 suggests a time to failure of about 5 hours for a spacing of 1mm and 500V (btw the x axis is um not mm, I think). They state

time to failure = k x spacing ^2/ voltage

This implies a spacing of 10mm can fail in 500 hours but this does depend on the other factors mentioned above. If you Google CAF resistant FR4 you will see that there seem to be quite a few options available to mitigate this. I guess when the Marshall TSL was first built around 1998(?), this wasn't a very well known phenomenon. Must have been quite shock for many.

[mhuss]

I have a trio of newer DSL 100 watt heads in (joy.) "They're fine, they just keep blowing the HT fuse." The power tube ink has changed color showing clear signs of overheating. Does anyone know if Marshall ever fixed the FR4 problem in the DSL line? These are mid-2003 vintage (PCB Z03-14) and have the correct 5k6 grid stoppers on the output tube grids. Cold the resistance from pin 4 to 5 seems reasonable (220k+). I don't really want to disassemble and drill out three amps.

Edit: Arrgh, after warm up bias voltage dropping like a stone.

[Leo_Gnardo]

Does anyone know if Marshall ever fixed the FR4 problem in the DSL line?

Presumably amps made 2006 on have improved boards where the dreaded bias drift won't/shouldn't occur. mhuss, looks like you have a very full day's work ahead. Like they say at the dentist: drill, fill and bill.

I did the first couple of these without removing the output tube sockets. The dremel bit doesn't like to collide with the metal electrode pin on the socket. Makes for sloppy hole cut and potential busted grinding bit and/or smashed up electrode so I went for the socket removal method instead. Those who feel like it can try with the sockets in place, but for safety's sake I recommend not doing it that way.

bbb, bizarre that removing the PI tube sets the bias straight. It's beyond me, I can't come up with an explication... just expletives. Looking forward to somebody coming up with a fix. Because inevitably one such will land on my repair bench. It's a new board you say - drat - thought Marshall had got these sorted by now.

nickb, thanks for the link and your post. I'll have to read up on the link you gave. I thought heat on the board was a factor. If it's not heat dependent then leakage could occur anywhere, and if the new boards are failing too then all DSL/TSL owners are in deep doodoo.

Someone made the comment a couple years back, if these amps were cars, there would have to be a manufacturer's recall to get 'em right. Looks more like a case of planned or unplanned obsolescence, a pity because the DSL/TSL amps can sound pretty dam' good.

[drewl]

The newer improved boards are a darker green, if it is light green it's the old bad board material.

I had removed the filament traces electrically, ran wires, and anywhere filament traces would pass over any high voltage traces.
It got the bias and output section stable.
I did the same at the phase inverter and it was stable.
By the time it was breaking down around V2 it was too cluttered to try and isolate any traces that ran over each other inside the board.
That's when I gave up, it was a slow week at work which is why I dicked around with it to that extent!

I imagine eventually those board will fail all over.

[SteveJ]

FYI: Best deal I've seen on replacement boards are here. TSL100 Main Board Hot Rox UK. The bias drift issue has been corrected. However, it has affected the tone. The distortion channel sounds better (less fizz), but the clean channel does not sound as good.

[mikeydee77]

I haven't attempted the repair as I am not convinced it will be reliable. But from the ones I have seen suggest problem was centred around 2002.

Here is picture of the good and bad board colours. Bad board top, good board bottom. Flash has got in the way but the good boards are a rich leaf green and the drifting ones a paler maybe sage green.

http://amploft.co.uk/blog/wp-content...8-19.40.27.jpg

[mhuss]

Here is picture of the good and bad board colours. Bad board top, good board bottom. Flash has got in the way but the good boards are a rich leaf green and the drifting ones a paler maybe sage green.

Do they have different revision/production code/... numbers? (e.g. Z03-14) Can't really tell from the pictures.

FWIW, at least one of the failing boards looks like the darker color.

--m