SSL 661E power supply issue

[Azatplayer]

Hi guys, first post with trouble written all over it! Hopefully Keith has some gold for me as ive been digging up dirt!

I tech for a couple studios here in Melbourne Australia, guitar shops as well and also do custom valve gear.
One of my clients has a 6000 seres SSL. Ive done various repairs on the old girl, repaired a few cards and had cleaned up the power supplies some time ago.
I have had zero experience with these things, tho they all have a beginning and an end and ive managed to get around them with some thought.
I have a power supply issue on an older 661E thats the latest problem shes having.

Its dropping the -20 rail as soon as a loads put on it. In situ, it ends both rails, flick power off and on and they return, but with any load at all, it fails both rails.
On my bench i ran it under load one rail at a time and the -20 is the culprit. I replaced the bridge recs 6 months ago in this unit as well as the fans at the time. It had failed on one side and i did both recs as a matter of course.
Its a pretty untouched unit, has no fuses and thats a mod on the board to do, original caps, cards were till i recapped them and replaced two of the dead trannies found.
Been running fine since, but was called out to look at them as the second unit a 661G was running terribly hot and turning off. They had figured it was the issue but i guessed that it was doing all the work and this older one was not drawing current.
2 units into a changeover unit, all appears ok. Turn the old unit off and the G keeps going, turn the G off and the old unit fails.

I can not see anything wrong anywheres visually or with a meter. Ive swapped cards around, replaced the components in the -20 PSU monitor circuit thinking maybe that played a part, still am suspicious of it.
It has no protection anywhere, no relays, nothing to do what its doing that i can see asides the psu monitor. Ive tightened all connections. It does it when cold so i cant see it being a faulty thyristor, as it doesnt trip them all as it should as all the thyristors are in series.
Measured all resistors in there and all seem ok. Checked all the outputs and theyre all reading same same. After it trips that rail and I turn the power switch off, that very first movement, the lights come back on briefly as it turns off.
Just seems like theres something very lightweight at fault, nothing big. But where..

Appreciate all and any help offered by some more experienced minds. Thanks in advance, Donald

 

[jwhmca]

Did you recap both supplies?

 

[sthenic]

Hi, you can borrow the 48V regulator card and place it in the -20 position to rule it out. Like wise, you can slot the -20 reg in the 48v hole and check it under load.

Last time I had a bipolar rail collapsing under load it was a bad transistor, and replacing the whole bank (for good measure) brought the rail back to life.

You can desolder the joints, to diode test out of circuit.

Brilliant thing is.. apart from the regulator card, it really just a handful of components to go wrong.

Also check out the dedicated SSL tech forum, it’s mainly E and G series guys:

http://forum.sslmixed.com

 

[nielsk]

These are very prone to bridge failure, this can exhibit the symptoms you describe.
For some reason (purchasing mistake?) the secondary voltage is very high for the design, and the +/- circuits are stressed as a result (part of the reason so much heat is generated and MUST be evacuated).
There are also multiple places that corrosion can happen in the mains circuit, also suspect the power switch...

 

[SSLtech]

When you say "it drops both rails" I presume that it's NOT the crowbar circuit... because the crowbar would kick in even with NO load. -It's only triggered by voltage RISE, not voltage sag... so lets move forward with voltage SAG in mind.

I recently had this exact symptom and I diagnosed a failed rectifier, because I was measureing AC voltage going in, but no DC voltage coming out. -The client replaced the rectifier, but same problem...

When I returned and re-investigated, it seemed that one of the connections on the back panel (I think it was a crimp or barrier strip, from memory) was high-resistance. -under load, this caused the available voltage to sag, triggering protection.

You've got some kind of series resistance somewhere, is my suspicion. -Possibly a rectifier, but you have to look at exactly WHERE the voltage disappears. Even under shut-down, you should still see voltage at the same sort of numbers all the way along BOTH circuits (they're identical, after all!) all the way until the collectors of the power transistors. -If that's NOT the case, work backwards from there, and find out where the readings diverge. (-DO remember that both legs use different "low" reference points!)

Start by quickly swapping regulator cards to either eliminate them from suspicion or to locate one or other as faulty. -You can even swap with the 250V or 48V cards if you like. If one of those bears the fault, then I'd be surprised, but we can go into it from there. -Apart from that, it's basically a few passive components (like trim-pots), a rail of active transistors (which seem to be working, though there may still be a high-resistance point somewhere), and the "holy trinity" comprised of power transformation, bridge rectification and reservoir capacitance. A series resistance "upstream" would cause a momentary "sag" across the appropriate reservoir cap, and for tracing something like that, an ANALOG multimeter is superb. Digital meters may not react quickly enough to intuitively alert you.

Check for that stuff first. -Make absolutely[/b] sure that you have SOLID power at the collectors of the transistors before you start looking at the complicated bits... and most of those can be eliminated with a quick card-exchange to prove/disprove suspicions.

That should be enough to nail it, I'd expect.

Keef

 

[electroslut]

Or just buy a new PSU with a warranty 8)

https://www.atomicinstrument.com/

 

[pucho812]

Or just buy a new PSU with a warranty 8)

https://www.atomicinstrument.com/

yes... their supplies are quite good.