Inverters which go pop!

[Andy Warner]

Hi All,
I have recently tried parallel connecting 2 wind inverters fed from single turbine - an Aurora inverter with an early SMA Windy boy. The aurora is of course transformerless and the windy boy has a transformer and by design is galvanically isolated.

So I paralleled up the inputs post 3 phase rectifier from the turbine, and as the windy boy is a transformer type assumed that would take care of the need for an isolation transformer.

So at the point of grid connection BOOM. Quick examination shows grid breaker tripped, and one varistor gone on the Aurora, but probably much more inside... Nothing apparent in the windy boy.

Any suggestions on cause? My understanding was that the AC would be superimposed onto the rectified turbine DC by the Aurora, giving the windy boy a floating DC input, which shouldn't be a problem as it's galvanically isolated - surely it can float about as much as it likes as long as there is no path to ground on the DC side?

Any ideas or wisdom would be much appreciated.

Andy

[Rob Beckers]

Andy, the SMA may be galvanically isolated between in- and output, but the input could very well be referencing ground! If so, then that AC coming in from the Aurora (which also references ground, as a good size Voltage) will make sparks fly.

I don't know the design or schematic of the SMA inverters. This is just a theory...

-RoB-

[Andy Warner]

That's a good thought Rob. I've not found mention of it in the documentation though. The SMA has ground fault detection, so that would suggest its inputs are not grounded I think.

[Rob Beckers]

The many inverters that have a grounded input still have GFCI, so that in itself doesn't mean much. There's a number of ways to detect ground faults, even with a grounded DC input.

Of course, I don't know about SMA inverters and the cause of what you saw may be entirely different!

-RoB-

[Andy Warner]

I suppose so Rob, it could just do a sum of current on each line to detect any leakage, regardless of whether one is grounded. I have another sunny boy here, so will have a look to see if either input line is grounded.

[Rob Beckers]

Try to keep that Magic Smoke in Andy! :nuts:

[Andy Warner]

Yes! It's quite hard to put back...!

I've done a bit of measuring today, on a spare Windy Boy, and the DC side isn't grounded at all - resistances are around 10Meg ohm between DC lines and ground.

When looking at the board layout, the transformer mains side is pretty much connected straight to the grid on the SMAs. All that is between the transformer and grid is a couple of inductors and the grid relays. I still reckon this should behave much the same as a dedicated isolation transformer!