Repairing B&K Precision 9115-AT 80V 60A Power Supply

[magudaman]

Well crap I have really done it this time…I popped my B&K Precision 9115-AT.  I was testing my luck charging a battery at nearly full power and the Battery Management System (BMS) shutdown the pack to continue balancing.  While the BMS was down I made a minor adjustment to voltage and walked away.  About a minute or two later the BMS kicked back in connecting the pack to the power supply and a loud pop occurred, my stomach dropped.  I ran back and my whole side of the shop was dead.  Breaker wasn’t popped but the GFI tripped.  I reset the GFI and powered the supply back up hoping the pop noise was just the GFI, sadly it wasn’t. The power supply started self test and promptly dimmed the lights and popped the unit’s fuse.

So I got this thing for an amazing deal on ebay a couple years ago and I had to pull some financial strings to buy it at that time and a non-warranty repair today is $620 plus shipping to them.  All things considered I don’t think I can get a similar capability supply (80V 60A 1300W) for $620.  However I don’t think I can absorb $620 and would likely just downgrade to some thing less capable
Ok story time is over.  I need to attempt to repair this beast!

Here is a crappy overview shot of the board layout (sorry I thought it was centered)(added addition picture with center stage removed but better framing.





Left to right, looks like input stage,rectifier, filtering, and PFC (maybe, I see switching fets) then goes to the center stage.  The center looks like the primary switcher with DC output at maybe 100vdc.  It appears to only have voltage in (350vdc?), voltage output (100vdc), an additional 3 pin header than goes to the front controller (logic power and OK signal?), and lastly all the logic look to be on a daughter board.  The right side is output stage DC/DC.  I’m making educated guesses (all fringe knowledge).

Here is wide shot of left side board, input stage:


Close up of logic on left side board:


Wide shot of front control board:


Wide shot of right side output stage:


The center board is visibly damaged here is a wide shot:


You can’t see it but the two of the four primary FETs are blown along with their respective drivers and possibly the support passives.  Maybe overload or MTBF achieved?  I don’t think I back fed the unit but I could have.  B&K told me if you back feed the unit it usually blows the output stage capacitors from the inrush current input the supply.  The also confirmed they would not provide electrical diagrams .

Close up center board daughter board:


Overview of center board 4 primary fets and drivers:


Close up of center board primary fets and drivers:


Close up of center board other two fets:


Overview of center board traces underside:


Finally to my question.  Can I just get away with replacing the primary switching fets, supporting drivers, and their respective passives? Or do you think I need to dig into the logic passives on the center board too?  Is this all a lost cause to find out the DC/DC stage is cooked?  Any other advice as I get started?

[TheMG]

You could replace the drivers and passives, but leave all the FETs out. Power up the unit and use oscilloscope to see if there is a switching waveform at the gate connections of each FET. Important note: must use isolated scope probe or an isolation transformer when poking around the primary side of the power supply.

If you get a switching waveform on all the gates, this means you should now be safe to reinstall the FETs. Leave the 100VDC output disconnected from the next stage, power up and check for 100V.

The output stage looks like a DC-DC buck converter. As long as its input and output isn't shorted, it's probably safe to reconnect. With a buck converter if the FETs are shorted or being switched always-on, you get full output. If the FETs are not being switched on at all, you get no output voltage.  Only if the diodes are shorted could things get interesting, but this would be measurable as a short circuit at the output, since all that stands between the diode(s) and the output of a buck converter is an inductor.

[edpalmer42]

Before you do *anything*, build a dim-bulb tester.  Otherwise, I can almost guarantee that you'll run out of fuses before you get it fixed!

[magudaman]

Thank you very much for the help TheMG and Edpalmer42.  I got called out for work this last week so I didn't get any time to work on it further.  I am going to assemble a parts list this weekend and get an order in next week.  I will update with my progress as soon as I make some.