ntermittent faults are some of the most frustrating faults you’ll ever come across. Not only are they difficult to trace, but they’re also difficult to confirm you’ve actually fixed.
Whenever I get called to an intermittent fault, one of the first things I try to do isn’t actually troubleshoot it – it’s to figure out how to recreate it. If you can consistently make the fault happen, you’re already halfway to finding the cause. It also gives you a reliable way of testing your repair afterwards.
I had one site where a Deye inverter was giving intermittent battery CAN bus communication losses (you can read about that case here). Looking through the monitoring data, I noticed the communication losses only ever seemed to occur when high currents were flowing through the grid input conductor.
To test the theory, I simply recreated those same conditions. I discharged the batteries before charging them back up, first at a low current and then at a much higher current. Sure enough, the fault only occurred during high current charging. At that point, it became fairly obvious the problem was likely interference on the communication cable.
After replacing the communication cable with a shielded Cat6 cable, I could perform exactly the same test again. This time the fault never returned, giving me confidence that the issue had actually been resolved.
I had a very similar experience on another site with a Deye 50kW inverter that was giving intermittent W7, W8 and W9 (Fan Out) warnings. The warnings had only appeared twice over the course of a week, making them extremely difficult to investigate.
Again, the monitoring platform gave the first clue. Both events occurred while the batteries were charging and the inverter temperature was increasing.
This particular site only allows a relatively small amount of self-consumption because the PV array isn’t large enough, so getting the inverter into exactly the same operating conditions wasn’t as straightforward. After allowing the batteries to discharge slightly, I forced them to charge again until the inverter reached the same operating temperature. Sure enough, the warnings appeared again.
Now we had something we could actually work with.
After cleaning the cooling fans and checking all the fan connections, I could immediately repeat the exact same charging test. Rather than waiting another week hoping the fault had disappeared, I was testing the inverter under the exact conditions that I already knew caused the problem.
I’ve found that in many cases, simply figuring out how to recreate an intermittent fault often points you towards the root cause. Even if it doesn’t immediately identify the problem, it gives you a controlled way of testing every repair you make.
Whenever I’m dealing with an intermittent fault, I don’t start by asking “How do I fix it?”
I start by asking “What conditions make this fault happen every single time?”
Once you can answer that question, the rest of the troubleshooting process usually becomes much easier.