Junction Box overheating

[nidalls]

aaaa

[littleharbor]

Are the junction boxes really oriented like the photo shows? If so water intrusion could be an issue.

[SunEagle]

Depending on the number of panels in a string you may have exceeded the max voltage rating of that junction box.

[bcroe]

I think you have a failing panel operating in bypass, which will tend to reduce the voltage of the string.
The rest of the string will slide up its voltage curve to match good strings, but will lose current capability
at that operating point. Perhaps heat from the bypass diodes is evident here. Bruce Roe

[bcroe]

the voltage for the defected string is the same compared to other strings

That is what I said. Now go back and read the rest of what I said. Bruce Roe

[bcroe]

Understood.
but can this lead to a great difference in the defected string current? 45% lower compared to other strings?

In my strings of 12 panels, a bypassing panel will make the other 11 increase their voltage
by a factor of 12/11. If you look at the voltage/current curve of cells, you will see this results
in considerably less output current. Not enough info here to compute your situation, but I
have used such measurements to find only 1/3 of a panel in bypass. If this is your problem
(as suggested by hot junction box/conducting bypass diodes), a measurement of each panel
will show very little voltage at the failure. Perhaps a thermal meter will pick out the problem.

At this location I find bad panels by going out midnight and inserting some voltage taps in
the weak string. When the sun shines, some sections will have more voltage per panel
than others. I move the taps into the suspect section (in the night, switches open) and
narrow it down. Eventually I find the exact panel. Bruce Roe