Lexicon 480L Blowing Fuses

Dear Group DIY, many thanks for all I've learned reading this board over the years! A particularly tricky project has come across my desk and I'd love to hear any insights you may have.

I'm working on a Lexicon 480L that has had a power supply problem and is now blowing fuses. This box came to me with problem on it's 15V rail. I isolated the problem to the unit's mother board and I found a shorted 7815 regulator. Unfortunately when I replaced that, it started blowing fuses…

I've gone through and replaced all the regulators in the linear power supply along with the filter caps, and all the tantalum caps on the board. Disconnected from the circuit, the power supply works fine but as soon as I connect it the fuses blow. Obviously something is drawing too much current but when I measure between power and ground I'm seeing a load in the range of 2.5K, not an outright short.

It's a huge PCB but I'm not seeing any visual damage or suspicious parts. Any ideas how to track down what's blowing the fuse (hopefully leaving the ~15 custom ICs on this board intact!) ?

Thank you!

Have had this problem last year, and sadly it was one of the Lexicon output drivers - not available anymore.

Silent:arts, thanks for the information! Oh I really hope it's not one of the output amps... I'll report back with what I find.

Time and time again I see problems with these proprietary boards/chips potted in plastic (Roland Juno 106, API 2520 etc...) I get it from an intellectual property point of view but it really seems to cause problems in the long run. I'm happy to see that DIY discrete op-amps are starting to reverse the trend.

On a theoretical level I'm still puzzled by this short circuit situation. If something is pulling too much current, why is my meter (a nice Fluke 115) reading ~2.5k between +15V or -15V and ground? I've seen fried semiconductors do strange things and there are a lot of unusual parts on this board but I would still still expect a short to read as a short.

Well thanks again Silent:arts, you nailed it. One of the output amps is bad and it seems like this is a very rare part indeed (if anyone has a scrap 480L sitting around, let me know!!).

Looking at closely, it seems like this card is buffering the signal and providing two outputs, a normal one and an inverted to make an active balanced amp. From what I can see I'm guessing there are a couple op amps in there to flip the phase and three transistors (maybe to beef up the current for each output). On the back there are some big SMD caps and maybe a few big resistors, my guess would be that one of the SMD caps is tantalum and failing short but I'm not getting anywhere with my meter.

At this point I'm wondering if I can replace these output amps with something like an SSM2142 or THAT1646. My other thought is to melt the plastic coating of this board using acetone like people do to Juno 106 Chips (https://www.youtube.com/watch?v=W4lF0xSVQlg) and go in for the surface mount repair.  ⊙︿⊙

I already tried the "acetone" methode with my Juno-106, and it really worked!
In case of the Juno106 the plastic coating itself causes the problem.
This plastic coating is becoming slightly conductive after all these years >>> Oscillator doesn't work anymore.
Just removing that coating solves the problem.

I don't know what happens in case of the 480L... but that methode will at least help you to "see what's inside".

...markus

Hard to believe it's been 6 months since I posted this! I wanted to follow up and share what I learned encase anyone is looking at a similar problem in the future. Again, many thanks to silent:arts for the tip about the output amps, that was essential in sorting this out.

From calling around to all the best techs I'm aware of, reaching out to folks from Lexicon/Harman and speaking with folks who worked on the 480L design but left Lexicon years ago, there seems to be very little information floating around about this LLA5200 amplifier and I wasn't able to find any spare chips. The only place I found with any interest in working on this was Beamish Electronics in Cleveland.

It turns out that although these problematic output amps are hybrid boards from the same era as the Roland 80017a, acetone is definitely not the right solvent to open them up  – after a week's soak there was no effect at all.  That said, it is possible to get into these boards the hard way: scraping away the plastic by hand.

After scraping away the plastic I found there's actually a fair bit going on with the circuit – the front side of the board has a couple op amps (LM833 and uPC4570), a few smaller transistors and resistors printed right on to the ceramic substrate. Without vias through the ceramic board, many of the pins on the edge of the chip serve to connect one side to the other. That side has a number of larger SOT-89 transistors (to beef up the output current I assume) and tantalum caps. In this case it seems that one of the output transistors and one of the op-amps failed which brought down the whole analog power supply. Since everything was already exposed, I ended up replacing all the output transistors, tantalum caps, and both op amps.

Ultimately this was a damn tricky repair but I just wanted to say it is possible to do, but I couldn't have done it without GroupdDIY!