380Vdc Solar Bus

[Jester]

I have been asked to build a simple voltage divider PCB for a solar array, to step down the 380Vdc solar bus voltage to 200V to be compatible with one of the 250Vdc max. analog inputs on the controller that is managing various loads connected to the solar system. I don't yet have detailed specifications on the 380V bus, however I was told that the voltage divider should be designed to withstand 600V. The procurement guy is non technical and is just passing on the somewhat vague specifications.

Does anyone have any idea why 600V max would be specified?

I do have details about the available analog input, Floating differential input, 2M input impedance +/- 250Vdc maximum.

This will likely be nothing more than three high voltage SMT resistors, perhaps 200k, 2W. Considering the differential input, the output could come from any one of the resistors, I'm not sure if there is an advantage to feeding from the center resistor so that there is some current limiting resistance to either side of the source?

I recall reading that 600V on FR4 can be problematic over the long haul because of electrochemical migration.  Assuming I design to the 600V specification, I'm inclined to design to pollution degree 3, Reinforced (20mm spacing) to be really conservative. I'm not sure if conformal coating is necessary?

This PCB will be mounted in a IP67 rated enclosure outdoors

Fusing must be at the source, It seems reasonable that I should include a warning to that effect near the input terminals, perhaps the output terminals on the PCB should be fused to prevent over heating if the output wiring or controller input  terminals fail shorted?

If anyone has advice or comments I'm listening.

[Nauris]

Simple solution would be to just place 2 M resistor in series with the input. Either solder it inline with wires and heatshrink over or screw it in an terminal strip somewhere.

[Ice-Tea]

Did a quick check on three solar converters: one has a max bus voltage of 400, another 500 and yet another 600. So I'm guessing that's where the 600V comes from. Also means that that number doesn't have any margin on it yet?

[T3sl4co1l]

Maybe transient overhead as well?  The cables could get long, I would guess, and therefore be susceptible.

Easy enough to use a kV rated resistor, or a few lower voltage resistors stacked.

FR-4 handles high voltages just fine, all you have to do is use fatter, rounder traces to reduce the electric field gradient.  This isn't very important at ~1kV, more getting into the 5kV+ range where corona is a problem.  In any case, use ample clearance, and rout slots if you can't get the required creepage (say if a component has fixed pin pitch that's too small to deal with otherwise).

If this is one-offs, I'd just as well suggest air wiring some resistors on tag or turret board.  The distances between connections on these boards is more than adequate for low kV between adjacent positions, and you'll be doing even better than that with a few resistors stacked.

Or a stack of resistors tied together alone, without terminals, then wrapped in heat shrink to make one effective high voltage resistor element, and place that wherever.  With enough strain relief, that can be placed inline as wiring.

The PCB will be cheapest in quantity, for parts + assembly cost, of course.

Tim