Astral Pool Robot SMPS exploded - some repair guidance sought

[Lorenzo_1]

The control unit for my Astral pool robot expired recently with a loud bang immediately after being connected to power.  Pictures show visible damage on SMPS and microcontroller board.  I’ve annotated the reverse side with the main primary side components and added blue lines where black mask obscures the traces or there’s ambiguity. Some primary side components are hidden by heatsinks.

I’m no expert on SMPS but it looks like this unit uses some of the design elements found in this link - https://datasheet.octopart.com/EVAL6599-400W-T-STMicroelectronics-datasheet-148140.pdf.  It has a sub-board with an L6599AD controller specific to series-resonant half bridge configuration, an L6563 PFC controller and the blown IC turns out to be a Viper12A.  The blown cap was 50V 47uF.  The board outputs 28V to drive the robot and has a 5V three-pin regulator (IC2) supplying the power switch, LED etc (some models have fancier display panel options). Haven’t found a schematic etc for this board.

I have a few questions before I try and repair it:

What is the likely cause? Perhaps indeterminable but might help pin down other components to check.

Do I need to strip components out and clean the board really thoroughly, or will a good was down with circuit board cleaner, IPA etc suffice?

Blown diode D2 is marked A6 and seems likely to be a BAS16. It has an open triangle with T inside as Mfr mark, but I couldn’t tie this down.  Would a BAS16 suit here? R20 looks damaged but reads OK at 1kOhm, consistent with some evaluation circuits I’ve found.

Blown Zener Z2 is unidentifiable but the Viper12A datasheet indicates that Vdd can be between 9 and 38V and switch on/off voltages are typically 14.5V/8V.  Can anyone suggest a suitable replacement value?

Most of the other components appear to have survived. I’ve not tested MosFETs Q2 & Q3, which are 12N50NZs.  Do I need to desolder these to test them out of circuit?  And to what extent would it make sense to desolder other components to clean the board more thoroughly or test them out of circuit? It’s still showing some superficial carbon marks round the blown IC, which won’t come off easily. It’s pretty clean now but not spotless.

Lastly, is this an SMPS that I can start up on a Variac? I have an isolation transformer and lightbulb current limiter I can use with it.  But I understand some SMPS units don’t like low-V startup.

Appreciate any advice members can offer.  Not come across this type of SMPS before and its design and operation appears quite complex.

[amyk]

What is the likely cause? Perhaps indeterminable but might help pin down other components to check.

Power surge. Those single-chip mains switchers are known to blow relatively often because they're quite sensitive to overvoltage transients.

[Lorenzo_1]

Thks for the feedback.  I've ordered a 14V zener that I think should do the trick, BAS16 diode, Viper IC and cap.  Will fit them and see whether they're enough to bring it back to life.  Will be a while before I get more time to spend on it.

[Lorenzo_1]

Made some progress with this. Replaced blown BAS16, 14V zener and Viper 12A.  Cleaned up the board and tested all I could easily do.  All looked good but no 5V standby supply on startup - hence the microcontroller can't power up and nothing else happens.  Turns out one of the windings on the standby voltage transformer is open circuit.  Decided to strip it down and found a burnt section in the deepest layer - 256 turns of 0.007" (0.18mm) wire. It burnt out just a few mm short of the pin (damage was minimal - just a bit of melted insulation).  The other coils were OK - 24 turns of 0.021" wire (with yellow insulation) in layer 2 and 48 turns of 0.007" in layer 1.  The 1st and 2nd layer wires look OK for re-use. 

However, I also need a new transformer bobbin as this one's cracked near the pins. Suprisingly, a bit of research located an exact match (though with 5+5 rather than 5+4 pins) in their MCT-EE1663 on "www.mycoiltech.com" (https://www.mycoiltech.com/horizontal-bobbin-ee16-transformer-bobbin-dip-bobbin_p484.html).  I would have to try and find a retail supplier to get just a couple of these.

Is rewinding this transformer manually a viable proposition? Not too worried about the time as it would be useful experience and, if it works, a spare SMPS in case my primary one blows up.  Don't have a coil winder but a speed-controlled drill might suffice. Replacement power supply is around $400 - and board likely to be ca half that price based on past experience.  Rewiring looks very doable and there doesn't look to be any alternative if I want to get this SMPS going again.

[vk6zgo]

A pool robot SMPS???

Please, please, tell me the SMPS runs off the normal transformer isolated low voltage DC supply commonly available for pool salt chlorinators, & not direct from the Mains!

What the hell does a "pool robot" do that requires any "smarts" at all? The old "Kreepy Kraulies" just used water flow to do their job, which they did quite well.
If there is any active circuitry in the head unit, that is a failure waiting to happen, & if the SMPS breaks down, possibly putting Mains onto that circuitry, an electrocution.

[station240]

That Viper12A is something I've run into a couple of times.

Had a rummage thru my box of SMPS boards, nothing with an EE16 core.
However I do have a dead PC PSU with a Viper12A that uses an EEL19, the pin arrangement is slightly different though.

The EEL19 is really common in the standby 5V supply of ATX power supplies. So it's an option to avoid having to rewind the transformer by just repurposing one.
However, most of them have two secondaries, an extra 12V one supplies internal circuitry, so you need to figure out which is which.

Let me know if you want the EEL19 I have here, it's only got the 5V secondary and I need an excuse to reverse engineer it anyway.

[Lorenzo_1]

Thanks for the replies.  Appreciate station240 going to trouble of searching for a suitable core.  I think for the moment I'll continue search for an exact/close match rather than try to adapt something else.  Seems best path to try and bring it back to life, given it may have other things awry.  I'm looking into whether I can't temporarily fire up the secondary board with an external 5V supply given everything else seems to check out just fine. Need to think about that carefully before trying it out.

As to why smps in a pool robot, I had to go this more complex route as suction cleaners can't handle the high volume of leaves that fall in daily because we're surrounded by tropic rainforest.  The smps sits remote from the actual robot which gets a 28VDC supply through a long cable.  Robot itself just has DC motors to drive suction propellor and wheels/tracks.  I guess there is some risk of shorting in the SMPS putting AC across the output - they say it has to be run from a earth-leakage protected power source and never to get in pool while it's in there.