π
2010-05-12 01:01
in Computers, Solar
It's been one year since I started graphing power and solar usage with the Brand One Powermeter. To be honest, I've been very underwhelmed with that device and would not recommend it to others, but I was still able to get stats, even if they aren't quite correct.
What's wrong with the brand one and its CTs is:
it randomly corrupts itself and sends me totally bad data that I have to filter out in software (know bug, owner says he filters it out too instead of fixing the broken hardware)
the overall software is very sad, not just the 2010 bug which is not a bug but plain lazy programming (2009 turns into 200: because ':' comes after '9' in the ASCII table). You should see the limited serial programing interface and error handling or lack thereof.
the fact that the unit is so slow that it can only upload once every 2mn and it totally locks up while doing so
I can't change its IP or any of its settings without sending it back to the owner and waiting 2 weeks or more.
My last support round trip to fix a misprogramming took almost 2 months to get it back fixed with weeks at a time where I got no Email answers.
Anyway, despite its flaws and confirmed incorrect measurements that are off by more than 2% and not in a consistent fashion (I tried to derive an error per day or kwh, but it was all over the map, from 2kwh short per day to overcounting during the winter. I think part of the problem is that the CTs are too big and pick up too much noise in the case of the solar panel probe.
Anyway, the PV system had its own counter and reported 8566Kwh for that period of time while the powermeter measured 8260Kwh (a bit over 5% off).
This graph and the graphs below are selected in zoom mode, so you can use the mouse cursor to draw a rectangle on a time region and the graph will refresh on the time slice you selected. Note that you can also zoom out by clicking the graph with the right button:
House Power without AC, you can guess which day was our wedding day :)
House Power
AC doesn't look like much, but it still comes in handy 6 months a year:
Photovoltaic production looks pretty good, even during the better winter days:
Also, the daily production start and end (hour the sun starts hitting the panels and when it sets past the roof while the end hour can be a bit early because of trees on the west of our roof). Note the one hour jump on Nov 1st and March 6th because of the daily savings change:
And the interesting part is the PG&E Meter. 0.25Kwh average is 2170Kwh/year, but due to time of use our bill is actually slightly negative:
Anyway, even if the scientist in me is quite bothered by the inaccuracy of the Brand One Powermeter (the data being over 5% off is just no good), it's still good enough for nice trend graphs and seeing the results one year later. |