HM-1500 suspicious malfunction "PV-1: No input"

[oooomh]

Hi, the following observation is not about an error in OpenDTU, but rather of a suspicious malfunction the HM-1500 inverter itself. Nevertheless I would like to share my observation with this OpenDTU/Hoymiles community, as OpenDTU helped me to trace it down:

My system consists of a Hoymiles HM-1500 and 4 identical PV panels, each with 395 Wp. PV panel "A" faces southward at 45 degree, while PV panels "B", "C", "D" face eastward at 60 degree. The panels A-D are connected to the inputs 1-4 of the HM-1500 as A-1, B-2, C-3, D-4. This system is monitored with OpenDTU and has been running stable and nicely for about 15 months.

After a summer holiday I noticed that the daily AC peak power had been reduced during the previous days, to about 3/4 compared to its usual size. "String 1" reported only about 0,9 W during direct sunshine and the Event Log showed the message "PV-1: No input".

I interchanged the wiring of A and B from A-1,B-2 to A-2,B-1 and found then about 0,8 W on String 2, and "PV-1: No input". I interchanged the wiring of A and C from A-1,C-3 to A-3,C-1 and observed about 0,7 W on String 3 and Event Log "PV-2: No input". Going back to wiring A-1,B-2,C-3,D-4 yielded the original observation: low/no DC power on String 1 and "PV-1: No input". Grid disconnection and/or inverter restart via OpenDTU did not change the behaviour of the HM-1500.
I concluded that PV panel A would be defective and the HM-1500 was OK.

Then I measured the open circuit voltage "Voc" and the short circuit current "Isc" of all four panels using a multimeter.
Results: Voc around 40-45 Volt for all four panels, Isc at 9.59, 2.5, 2.5, 2.5 Ampere for panels A, B, C, D.
Conclusion: all four panels behaved faultlessly, with Isc values in approximate accordance with the angle of the sun at afternoon time.

During this Isc measurement, the HM-1500 was disconnected from the AC grid and also disconnected from all 4 DC inputs simultaneously. Previously, the HM-1500 had always been connected to at least one of the DC-inputs.

After reconnecting all panels as A-1, B-2, C-3, D-4 and reconnecting the HM-1500 to the grid, I made the following stunning observation. At first: Event Log "PV-1: No input" and no/low DC power on String 1, normal power on String 2-4, i.e. same problem as before. But then after very few minutes (approx. < 5-6 minutes): Suddenly all four DC Strings are back at reporting normal/expected DC power(!).

My conclusion: the HM-1500 was temporarily in a complete dead-voltage state, and this helped it recover its normal function?!
My question: Is it maybe problematic for the HM-1500 that panels A and B at input strings 1 and 2 can have greatly varying DC power (370 W vs. 70 W, or vice versa) during the course of the day because of the 90 degree difference in horizontal inclination towards the sun (east vs. south)?

Any thoughts about this, or similar obvervations?