Blown Diodes - Parallel Panels

[bob-n]

My friend said that he is seeing multiple blown diodes on his panels. He has sets of two panels in parallel and blames the blown diodes on paralleling panels. I can't think of a reason why paralleling panels will cause diodes to fail. Can you?

I'm tempted to blame either inadequate diodes or bad connections. By the way, these panels are quite a few years old.

Here's my model of parallel panels. Does this agree wtih your understanding of panel construction? Unless the inverter is doing something really ugly, it seems this should be safe.

ParallelPanels.jpg
Yet this happens:

blowndiodes.jpg Thank you.

[Mike90250]

Parallel panels has nothing to do with those diodes..

Those are BYPASS diodes, that activate when a portion of the panel, is shaded and stops producing power. They heat up and eventually fail. Panels in full sun will never blow bypass diodes, panels with regular shade will eventually fry the diodes.

[bcroe]

Those diodes are of quite marginal capacity, certainly not likely to have
a long life under stressful conditions (shading in good sun). Nothing to do
with panels in parallel. Failing open circuit they will have little effect on
output, shorted they will kill the panel.

Bruce Roe

[bob-n]

Mike,

Thank you. Yes, all of that makes perfect sense to me, except "panels with regular shade will eventually fry the diodes." I'm not saying you're wrong, but it seems to me that high-current low-frequency diodes are really old technology and we should be very good at making reliable diodes, even if operating hot 24/7 at full current. But they will certainly get HOT, and that must be the issue!

[bcroe]

Mike,

Thank you. Yes, all of that makes perfect sense to me, except "panels with regular shade will eventually fry the diodes." I'm not saying you're wrong, but it seems to me that high-current low-frequency diodes are really old technology and we should be very good at making reliable diodes, even if operating hot 24/7 at full current. But they will certainly get HOT, and that must be the issue!

Old technology yes. Diodes for carrying these currents typically are heat sink
mounted, not floating in an extremely confined air space with a marginal sized
plastic case. Bruce (older than silicon diodes) Roe

[bob-n]

Bruce is certainly right.

I just did some quick math. I didn't try to be exact. I assumed a 200 watt 40V panel, so 5A operating current. It made the math easy.

The R-6 diode package is rated 15C/W thermal resistance. Conventional diodes would have at least 1V drop at 5A, so 5W dissipation. That could add 75C to the diode operating temperature. Add to that 25C air temperature and another 15C or more temperature rise for heat from the sun on the panel, and we can see 115C diode temperature. That doesn't sound terrible for silicon, but a typical 6A diode is derated to 2.2A at 115C (curve below). Also, I really doubt the reliability of those spring contacts at high temperature.

All told, this sounds like abysmal engineering. Perhaps they get away with it because of some assumptions about low light level and reduced heating when part of the panel is shaded.
diode.jpg

[sdold]

The picture seems to show connections that got really hot, not necessarily bad diodes.