I can usually squeeze in a few updates on the live system without anyone noticing...assuming it all goes right. The key word in that statement is 'usually'. We had a slight hiccup this weekend and had to take the site down to sort them out. I though it was only going to be for about an hour....
There is a Murphy's Law for time estimation of projects. 'To estimate the time required to do a task, multiply the estimated time by 2 and increase to the next largest unit of time. Therefore a 1 hour job will take 2 days.'
Well that law held true again. The updates were completed Monday morning at 6:00 AM.
As for my roof not showing data for the weekend, that is an old issue with micro inverters. When you have 20 of them, and the weather goes from cloudy to sunny frequently, the RCD will trip. This is a known feature in buildings with lots of flourescent lighting (and no solar panels). In my case the solution was to replace the RCD from a B unit to a C. The reduced the frequency of this happening from weekly to yearly, but as I was working on the monitoring, I didn't notice the tripped RCD.
We also solved a very interesting issue with a customer who thought his array was failing. He had 15 inverters reporting intermittently. When I looked into the data, we also found 4 other inverters that occasionally came through his gateway. The only way this could happen was if another inverter was sending its data out via his gateway. It occured to me that if other inverters were connecting to his gateway, perhaps his were connecting to a different gateway, hence the intermittent reporting. After contacting him, he confirmed that his neighbour had 4 panels with Enecsys inverters. After disconnecting the neighbours gateway, all 19 panels begain reporting from the same gateway.
All this between getting the kids to orchestra on Saturday, the concert Sunday evening (we didn't get back till after midnight) and getting one child off to school for 6:00AM for a trip to Snowdonia.
Hopefully today will be a nice quiet Monday.