I wrote in another thread about an off-grid quad Sol-Ark system that was shutting down due to parallel stop when one of the four inverters experienced a DC PV fault, and how that shutdown is far from ideal. The same system is down once again, this time due to an AC fault code.
The homeowner started getting repeated F18 and F34 AC overcurrent faults on one of the slave inverters. This, in turn, shut the entire system down due to parallel stop faults (F41). None of the other units had AC overcurrent faults, and the load is nowhere near requiring all four inverters for even the most demanding circumstances. It was designed this way for redundancy, which I am quickly finding out is not Sol-Ark's strong suit. To diagnose the issue remotely, I had the owner turn off all four load breakers, all DC PV input, and the AC microinverter input on the GEN terminals. I had them restart everything (several times). Every time, the same inverter would have repeated AC overcurrent faults, and the others would have parallel system faults. Since there were no loads connected by virtue of the load breakers being open, I suspected this had to be an internal fault. I went to the site, and Sol-Ark Tier 1 tech support had me shut off all inverters and take the suspect inverter out of parallel operation mode. As a standalone master it was able to power up and support the entire house load without issue. Then we reprogrammed it for parallel operation again and turned everything back on. We were unable to stay on the phone long enough to determine if this was successful, but ultimately, the fault returned. I was told to call and ask for Tier 2 next time if it happened again, which I intend to do on Monday. At this point, the issue can only be internal to the unit, and I intend to demand warranty replacement of suspect components or the whole unit. I had to get the system running, so I wanted to take the bad inverter out of the parallel system. I was hoping that simply shutting it down would work. This is the third of 4 inverters in the Modbus chain. When turning it off completely (all AC and DC switches disconnected), the 4th inverter would fault, presumably because the Modbus signal was not being relayed, but inverters #1 and #2 worked fine. However, I wanted #4 to also continue working while taking #3 out of service. So then I turned on the battery disconnect for #3 but left it in the off mode by not pressing the on/off button, thinking that it would allow relay of the Modbus signal from #2 to #4. That allowed the system to work momentarily, but then everything faulted out due to parallel system stop. In other words, I was going to have to physically take #3 out of the Modbus daisy chain to make this work. Of course, I didn't have a long enough Cat5 cable with me, nor a Cat5 splice connector. So I had to rig something, which I did successfully to jumper from #2 to #4. But when I turned everything back on, #4 still would not work. I eventually realized that you have to change the Modbus address from 04 to 03 in the settings. Apparently, the addresses need to be sequential for it to work. Once I did this, I was able to get the system up and running again as a triple-inverter parallel setup. No faults were observed. So the theory was proven that #3 has an issue internally. Anyway, bottom line, I am disappointed at how one inverter fault takes down the whole paralleled system, and also how taking a faulted inverter out of the system requires physical and programming changes. Turning it off should be sufficient. This is a very poor way to implement a parallel system that should provide the peace of mind that redundancy implies. Now I have a customer who thought they were getting a system with failsafe redundancy that actually requires a service call every time one of the paralleled units decides it does not want to play nicely with others. Jason Szumlanski Florida Solar Design Group
_______________________________________________ List sponsored by Redwood Alliance Pay optional member dues here: http://re-wrenches.org List Address: RE-wrenches@lists.re-wrenches.org Change listserver email address & settings: http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org There are two list archives for searching. When one doesn't work, try the other: https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/ http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org List rules & etiquette: http://www.re-wrenches.org/etiquette.htm Check out or update participant bios: http://www.members.re-wrenches.org
