TickTock said:
Yeah, I would have expected to see a seasonality in the reverse of what we see on the gids (lower capacity in winter). The gids behavior at first seems like it may be an attempt to compensate for temperature variability, however the fact that the 100% pack volts doesn't change with season (394 V) means no compensation is occurring - pack is being charged to a fixed voltage year round. So the only explanation I can think of for my lack of seasonality is it just doesn't get cold enough in Phoenix for me to see this cold weather effect on capacity.
What sort of battery temperature-while-charging reductions do you think your LEAF has experienced since last summer?
What I don't understand is this:
You show a relatively large variation in the calculated charge, "0 to 100 wall" of 18.52 kWh to 19.83 kWh (~7%) over just the last ten days you have reported.
https://docs.google.com/spreadsheet/ccc?key=0An7gtcYL2Oy0dHNwVmRkNkFnaEVOQTVENW5mOTZlb0E&pli=1#gid=1" onclick="window.open(this.href);return false;
If the voltage levels are relatively consistent (392.5 to 394 V) at your "100%" charge level, why is there so much variation (and it looks like no strong correspondence to the voltage levels) in the kWh "from the wall" numbers? I thought I understood (from the expertise of others) that voltage variations that small at that voltage indicated fairly small variations in the "100%" level of total battery capacity. So if the voltage reports are accurate, and the kWh calculations are also accurate, they should be fairly close, shouldn't they?
If you are positive the battery voltage numbers are accurate indicators of capacity, what are the other factors you think could be causing the variable results? Do you use gid counts or voltage readings to indicate the battery level before you charge, in calculating your "0 to 100 wall" kWh?
I have been hopeful that recharge calculations could provide more consistent and accurate indications of battery capacity than range tests have, the biggest advantage being the ability to easily collect much more data.
However, some who have run repeated range tests (myself included)
may have seen somewhat more consistent (albeit far fewer) capacity results.