DuellingInverters

This blog is being put together in the hope that it assists anyone who encounters the same problem that it has just  taken me almost two months to resolve. What problem? Read on!

At the end of January I had a new Solis PV inverter installed to replace an 11 year old inverter that suddenly expired. I chose a Solis 1.5KW as I fitted one of these to my garden PV array 2 years ago and it's been no trouble.  The new one was for my rooftop PV array. Both arrays are quite modest at a little over 1KW. The installation was straightforward and I was exporting to the grid within minutes of switch on. Job done. Or so I thought.

At around 8:20 next morning the earth leakage trip was activated and all the house power went off. I didn't think much of it, just reset the trip and all was normal. Next morning at the same time the same thing happened and the penny dropped - the trip was when the light level was sufficient for the inverter to start exporting.  Again I could reset the trip and all was well. So I started switching out different mains loops to see where the problem lay. A tedious process as there was only one potential trip point each day. After a week or so I got down to the garden mains feed, the feed to the other inverter, some 150 feet away. I isolated the cable and checked for insulation with a low voltage test meter  and all seemed good. I eventually discovered that if I switched off the PV DC at the garden inverter there was no trip. So I used a CT to monitor the current from the garden inverter and built a  battery driven data logger based on a Raspberry Pi Pico. I discovered that at the time of the trip the garden inverter was also coming on line, certainly within the one minute resolution of the logger.  So it seemed to be that the two inverters were somehow interacting with each other and causing some sort of power surge which tripped the mains. Quantum entanglement or something more prosaic? As a simple test I covered a small section of the garden PV array with a towel (before trip time) and sure enough there was no trip. Reducing the light level effectively delayed the second inverter going into export mode.  

The next step was to go onto a few forums to see if this was a known problem. The most useful was renewableheatinghub.co.uk where, chatting to some very knowledgeable and helpful people,  I discovered that the problem was not entirely unknown, but that finding a solution was not at all easy, the answer being in fitting spike and surge suppressors.  So I bought a good quality surge suppressor and tried it in-line with each of the two inverter mains feeds. This this didn't help at all.  I decided not to bother trying spike suppressors as I couldn't think of a safe and easy way to install them.

I then had the idea of talking to the manufacturer, Solis. To my surprise they were very helpful and seemed to think an inverter software update would  resolve things. I got the feeling they knew about this issue but of course didn't say that. What I had decided would be an easy fix was to up the specified PV cut-in voltage from the standard 120vdc to, say 140vdc, effectively delaying export. But they had their own ideas. Anyway a week or so later I was in possession of a special firmware installation device that they had despatched. I loaded a firmware file that they emailed me and climbed into the loft to do the upgrade.  

My loft seems to be a permanent home for a lot of flies and there's fibreglass dust  so it's not the ideal environment to work in.  As it happens working in less than ideal conditions  is just the sort of thing I did when I had a paid job, except it wasn't lofts I was in, it was huge industrial boilerhouses. And I wasn't uploading someone else's software, it was usually someone else uploading the software fix that I'd sent them. Only when the fix didn't work would I have to go to site to sort things. So an interesting bit of deja vu with some role reversal.

In the event the firmware upload didn't go well. It took half a dozen attempts going though menu sequences on the inverter control panel and then suddenly it worked. Except that it didn't work properly, it stopped the inverter working at all. Two steps back for one forward. 

More discussions with  Solis resulted in me talking through it all on the phone with a field engineer who had sent me yet another software file. So I donned my battle gear (respirator, glasses, hat, headlamp) and tried it all again. But again it failed at each attempt. I told the Solis engineer I was giving up. However I then decided to give it one last try and lo and behold it all worked - and the new firmware was running and the inverter was exporting! Joy! I was reminded at this point of the buzz of getting things working after so much failure. And I suddenly really missed my old job. Never thought I would!

The proof of success was next morning at sunrise when both inverters powered up and the mains didn't trip. At the time of writing I've had 3 days in a row with no trip so I'm confident it's all sorted now. And I'm also reminded of how many hours I wasted during my life chasing software bugs that didn't exist (the hardest ones to find!) because the problem was in the hardware and how many hours spent changing hardware when there was an easy software fix! 

So I'm feeling pretty good about all this. Shows that analysing a situation in a logical manner usually pays off in the end. I'm 80 now so well past my engineering peak (!) but I'm still building electronics and writing firmware. My current interest is environmental monitoring - for a peek of where I'm up to (it's a never-ending development) you can visit http://www.mecol.co.uk/weather/logplotter.html. And maybe have a read of another blog from a few months back on a Raspberry Pi Air Pollution Monitor.

 Incidentally the logger I built to help track down the inverter problem is now incorporated into the graphical display as the garden PV KW measurement. Waste not want not.