My SSL 4000 G just blew up!

Well Sorta,

One of the three redundant power supply has stopped working correctly, the +-20v's to be precise are no longer lighting up.
(history) I came in this morning and the SSL was wigging out... Lights out, hum in the speakers etc. I go into the machine room and find the +-20v supply off on all three of the power supplies. I switch the supplies off and then back on 2 of the 3 +-20v come back...

1. What might be going on? I know it's a broad question, but let it fly if you have any thoughts or experience...

2. I have been wanting to rebuild these supplies for a while now... like a preemptive strike. Thoughts? What typically fails? Weak areas? etc...

How many modules in the console?
Do you have changeover units?
What model Power Supplies?
I have rebuilt ALL of our SSL supplies.
Best,
Bruno2000

There are 80 channels

Yes we have 1 change over unit

the model is CF661E

On the supply that is down, are the +13V and +48V leds on?
Will the -/+ 20V leds come on if there is no load on the PS?
Bruno2000

Yes, the other Voltages on.

No, the +-20v LED's will not come on with "No load"

I have swapped around the "regulator" cards and it did not help or change anything

Do you have voltage across the large filter caps?

If I recall, the + and - rails are linked so that if one goes down the other shuts off as well.  Typically one of the power transistors shorts. If that's the case, replace ALL of them on that rail.  Another common failure is the bridge rectifier.  Sometimes I've seen both - transistors goes down and stresses the bridge and it fails, (or dies a week or so later, after you've  got the supply up and running again  ).  Also, check the crowbar SCR. Does your supply have fuses on the 20v rails? If so, you'll need to check them. If not, you should add them.  There was an SSL service bulletin outlining the procedure.

If you're lucky, you may just have a burnt crimp connection.  This is also very common - check the wires to the rectifiers. There's a lot of current flowing through those wires, and the crimped lug is the weak link.  If you see any that look baked, replace them even if they're still working.

I once spent days trying to fix my SSL psu and in the end it turned out it was a loose screw which connected a supply lead via a heatsink. The whole thing was kinda hidden between the heatsinks. If you're at it I'd do some serious cleaning / overhaul of everything that looks suspicious. Replace the fans. Replace all elyts, on the regulator boards as well as the motherboard. Everything is suspicious in that box - also connectors and leads, it all burns out from the heat and the current. I'd look for silicon (is that spelled right?) isolated leads for replacements, they can take heat much better. Probably replace the resistors at the power transistors as well, you'll see if they're falling apart (the isolation cracks / falls off) - but you need to find the same type of resistors if I remember right (for a reason I don't remember, was it oscillation?). I even had to replace several power switches, even they burned away. Carefully clean the contacts of the regulator boards. I don't know where to buy the main elyts, I remember having to order them from SSL for pretty much money. They're a special high current type, normal ones won't last long in there.
I also like to add air filters in front of the ventilators - but only if you don't forget to clean them from time to time!

And watch out - the power in that box is tremendous, if you accidentally short something you might hurt yourself seriously. My teacher used to say those psus should be good for welding...

Michael

+1 on the crimp lead. last time our ssl PSU went, it had to do with the crip falling off as the current burned/melted the wire jacket that the crimp was secured too. I ended up  soldering directly to the lug and heat shrinking over the solder point. hasn't failed since.

another common failure point it the wiring block where the AC mains cord ties in.
I have probably repaired these monsters over 100 times, I can't see replacing all the transistors every time one fails, as I have replaced only the bad one & had the supply work for years afterward, and it is such a PIA to work on this beast.
I have often fantasized about torturing the "bloke" who designed this beast!
If you meter each transistor in the chain, you can usually find the bad one by it having the lowest resistance reading across its legs. They did upgrade the pass transistors to (?) 3055s, so if yours is the old version (probably not if it really a G), do it to it.

+1 for not replacing ALL of the transistors!  I actually cut the trace to the base of each transistor, and replace with insulated wire.  That way, you can disconnect each transistor without pulling them out to test for shorts.

Also +1

I have often fantasized about torturing the "bloke" who designed this beast!

Click to expand...

Bruno2000

To each his own.  I didn't used to replace the whole bank of transistors, but got burned by it a couple of times, as have other tech's I know.  I believe it was SSL LA's policy to replace the whole bank as well.(?)  I'm not too crazy about the idea of cutting traces.  If you desolder the base and emitter pins, there's usually  enough clearance around the pins to isolate the transistor enough for testing.

It's better to do all the transistors otherwise it will bite you in the ass in the long run...

Thanks for all the input! I'm measuring a dead short across all the transistors in the +20 bank, I doubt they are all shorted, so i'm starting the process of desoldering to check which one is bad...

BTW, the other two supplies keep failing their +-20v supplies...? you can turn them off and then they will come back on.... could there be something in the SSL causing this whole thing?

jwhmca said:

Thanks for all the input! I'm measuring a dead short across all the transistors in the +20 bank, I doubt they are all shorted, so i'm starting the process of desoldering to check which one is bad...

BTW, the other two supplies keep failing their +-20v supplies...? you can turn them off and then they will come back on.... could there be something in the SSL causing this whole thing?

Click to expand...

Possible. On our ssl 4k if we power cycle the psu's and do not cut the buckets off with the maintenance switches before we turn it back on, our psu's will not power up properly. We often end up loosing our 20V rail. I suspect this is due to old age and our supplies could use a recap? I know that it has to do with current draw of our 4000.  Our quick solution has been to shut off each bucket and center section via maintenance switches under our internal patchbay. Then power up the supplies and changeover unit. Once those are stable we then power on each bucket and center section and done deal.

The +- 20v's are just shutting off intermittently...

So one 2n5886 failed and took out the Fuse... What is the best replacement for the 25a MD25a RS Fuse...? I can find them, but over in the UK only. I would like to go with something that is available from Mouser of Digikey...

I've switched the fuse and fuse holders on most of ours over to types more readily available in the US.  Something along the lines of the Cooper/Bussman type H250 fuse holders.

jwhmca said:

BTW, the other two supplies keep failing their +-20v supplies...? you can turn them off and then they will come back on.... could there be something in the SSL causing this whole thing?

Click to expand...

Have you balanced the remaining supplies? 80 inputs is a pretty good load for the two remaining supplies.  If they are unbalanced, you may be running one supply too close to it's current limiting threshold, which can cause it to shut down.  That then leaves one supply for the whole console, which will shut down for sure immediately afterward.

pucho812 said:

jwhmca said:

Thanks for all the input! I'm measuring a dead short across all the transistors in the +20 bank, I doubt they are all shorted, so i'm starting the process of desoldering to check which one is bad...

BTW, the other two supplies keep failing their +-20v supplies...? you can turn them off and then they will come back on.... could there be something in the SSL causing this whole thing?

Click to expand...

Possible. On our ssl 4k if we power cycle the psu's and do not cut the buckets off with the maintenance switches before we turn it back on, our psu's will not power up properly. We often end up loosing our 20V rail. I suspect this is due to old age and our supplies could use a recap? I know that it has to do with current draw of our 4000.  Our quick solution has been to shut off each bucket and center section via maintenance switches under our internal patchbay. Then power up the supplies and changeover unit. Once those are stable we then power on each bucket and center section and done deal.

Click to expand...

This is the way we always bring up our SSLs.  In addition, there are large relays that disengage the power to the supplies and must be manually reset if the line voltage "winks" (or takes a short nap).  The buckets are then shut down with the maint switches, and brought back up one at a time after the system is reset, and the power suplies are stable.
Bruno2000

IIRC SSL recomended no more then max of 30 channels + center section per PSU but I might be off here....

jwhmca said:

The +- 20v's are just shutting off intermittently...

Click to expand...

Had the same problem over here. Turned out to be the thermistors. try poking around in that area