Taking the Leap: Changing Caps In My DSL

dfrattaroli

New member
Before I start I want to point out that I've never worked on amps before. I'm a pretty "handy" guy and have built a solid body electric from scratch that I've been playing every day for 11 years now. I also do carpentry as a hobby and household electrical work as well. But amps always scared me. I spent a good deal of time researching things like discharging caps as well as working on this amp in particular. Before today, my DSL chassis had been laying on my bench for 2 weeks fully discharged with my bleeder wire attached at all times. I would walk in and read my meter just for the hell of it. I even switched the amp on and watched the meter just to see the bleeder wire work, speeding up the bleeder circuit already in the amp. My point is, I take nothing for granted.

Now, before I touched the filter caps, I got my feet wet by doing some simple and commond mods for the DSL: pulling C12, replacing C46 with a 1kV cap and doing the speaker jack mod from under the board.

Last night I decided to pull the 330uF 315v caps and replace them. Below are pics of the place where the old caps used to live. I also learned first hand, when desoldering, add more solder. :)

Below the board...
20160620_215651.jpg


On top...
20160620_215739.jpg


Now here are the new Nichicon caps. These are larger in diameter than the stock Samwha caps that came with the amp 16 years ago. I put one in as low as I could and then got the next one in as low as I could without putting too much pressure on the leads. The original caps were snap-in style. I used a zip-tie just to hold the orientation before soldering in the 2nd cap.

20160620_221140.jpg


20160620_221259.jpg


And from below...

20160620_221218.jpg


20160620_221149.jpg


Going to continue with the rest of the caps today. My choke gets here on Thursday.

Thanks!
Dave
 
Now, before I touched the filter caps, I got my feet wet...
That sort of made me chuckle the way you worded it. Good luck, have fun and be careful!

Some of those older DSL/TSL amps had circuit board issues. The board would heat up during use and become slightly conductive. Since the pins for the screens and the grids are adjacent to each other the traces on the circuit board are next to each other. The high positive voltage on the screens could then creep into the grids and make the negative bias voltage less negative making the idle current increase. The power tubes would start to glow cherry red and just get worse until the amp shut down.

Here's a page on it (not the easiest read but whatever).
 
SpiderWars":2hdy5d4u said:
Now, before I touched the filter caps, I got my feet wet...
That sort of made me chuckle the way you worded it. Good luck, have fun and be careful!

Some of those older DSL/TSL amps had circuit board issues. The board would heat up during use and become slightly conductive. Since the pins for the screens and the grids are adjacent to each other the traces on the circuit board are next to each other. The high positive voltage on the screens could then creep into the grids and make the negative bias voltage less negative making the idle current increase. The power tubes would start to glow cherry red and just get worse until the amp shut down.

Here's a page on it (not the easiest read but whatever).


Haha. I didn't realize the wording. Thanks for the tip on the board. I'll give it a read.

Dave
 
I read with interest the mod provided in : http://www.hullerum.de/Marshall/TSL122repair.html

FR4 from the pre non-leaded RoHS had a much lower Tg, 140C or lower, Whereas non leaded compatible assemblies use high temp FR4 with TG in the 160-180C range. Not sure that the mod addresses the symptoms more than the cause. Typical resistivity for FR4 TG 135C in 0.62" thickness (used by many but a less than desirable choice where tubes are used) runs: Volume Resistivity, min. 2x10(7th)megohms/CM and Surface Resistivity, runs about two orders of magnitude less. It would takes significant material failure to exhibit a 1-4 megaohm path, even than you are looking at 0.1ma of current to be integrated into the bias circuit and exhibit shifting of the bias voltage. So burning of the board by overheating tubes which surely could damage the FR4 and result in a compromised path between the nets, but what really is the cause?

Looking at the picture of the tube socket, the spacing appears to be less than ideal for a HV net. One thing that could be tried is to use an exacto knife to create opposing "D"'s in the pads (both sides) to create a larger air gap. You could also just make a small slot between them as you would for plate to heater pins.

If the HV is simply radiating and causing the net to be compromised, the a opening of the gaps and rerouting the net (can be either net) is needed, and cutting/peeling away the offending trace

Not sure if there was evidence or arcing (tell tale is carbon deposits/blackened path on the PCB) If you use conformal coating you get a 16X-30X (depending on voltage) gain in arcing performance ( as in rating required to arc over) compared to non coated board, free air measured at sea level It drops dramatically at high elevations. Conformal coating was used a lot in old picture tube TV's in the HV sections.

Comments about heater nets is spot on as you need at least 150 mil width (1oz copper) to get to a 25% margin over the 6 amp requirement. Personally I would go wider and with heavier copper.
 
Back
Top