VisibleTesla is cool. Among the other useful things it does is capture data for graphing. It also saves it in a parseable log. (Thanks, Joe.)
Even though I don't have VT running all the time, it has been running enough since the beginning of the year to produce useful plots of long term battery degradation.
So last month when I started noticing some change in range at the end of my daily commute, I wrote a few lines of Perl and R and got the following graphs. The graphs simply take the reported State of Charge of the battery and the reported Estimated Range, and divide one by the other to get a normalized range estimate. Then I plotted these values over time. Some events on the timeline.
October 27, 2012 - Got my car, an S85
February 25, 2014 - Received a replacement battery
June 30 - July 16, 2014 - Vacation with the car plugged in
October 15, 2014 - Received second replacement battery
Unfortunately, I didn't record the exact date I received firmware updates. But some of them are obvious on the chart.
I've driven a pretty constant 115 mile round trip daily commute plus a few trips and casual driving. My long term average energy usage is 323 wHr/mi, even though I drive "a little too fast" at times. Pretty happy with the output efficiency, a little less happy with the charging efficiency, but it's not too bad.
The February 2014 battery replacement was due to a sudden drop in peak output power. The failed component was a contactor (switch) internal to the battery. I happened to have also experienced a battery coolant leak that was the reason I took the car in, but when in the SC explained I saw this other behavior, and they found the failure in testing.
When I contacted Tesla this week to tell them my battery was failing as indicated by a steep range drop, the reply was: "We checked, and everything is fine." So I sent them these charts and asked the engineers to "please look again". The next day I got an email that they found a problem with my car, "independent of your range concern", and would replace my battery. That very same evening, my car abruptly failed to charge. Although there was no error showing on the console, the remote support folks could see an error when I plugged in the mobile charger while they were logged in - an error they had not encountered before, about an inability to balance the battery modules. The tow truck picked up the car at 7AM the next morning and had it back in my driveway at 7PM. Excellent service. Too bad I needed it. Glad I wasn't stranded (unless you count having to drive an ICE in the non-HOV lane being "stranded"). Worried it might happen again.
Yes, I got an A battery for each replacement. Yes, I asked for a newer battery this last time, but was told that was not possible and "really not necessary, as this re-manufactured battery had all new components". (I understand a new A battery is not the same as a new D battery, but there was no progress made in that argument.)
I have no idea why the data is so noisy until April 2014, but I suspect that after that date the firmware improved the reporting of range and SOC.
I thought the charts might be moderately interesting to others.
P
Even though I don't have VT running all the time, it has been running enough since the beginning of the year to produce useful plots of long term battery degradation.
So last month when I started noticing some change in range at the end of my daily commute, I wrote a few lines of Perl and R and got the following graphs. The graphs simply take the reported State of Charge of the battery and the reported Estimated Range, and divide one by the other to get a normalized range estimate. Then I plotted these values over time. Some events on the timeline.
October 27, 2012 - Got my car, an S85
February 25, 2014 - Received a replacement battery
June 30 - July 16, 2014 - Vacation with the car plugged in
October 15, 2014 - Received second replacement battery
Unfortunately, I didn't record the exact date I received firmware updates. But some of them are obvious on the chart.
I've driven a pretty constant 115 mile round trip daily commute plus a few trips and casual driving. My long term average energy usage is 323 wHr/mi, even though I drive "a little too fast" at times. Pretty happy with the output efficiency, a little less happy with the charging efficiency, but it's not too bad.
The February 2014 battery replacement was due to a sudden drop in peak output power. The failed component was a contactor (switch) internal to the battery. I happened to have also experienced a battery coolant leak that was the reason I took the car in, but when in the SC explained I saw this other behavior, and they found the failure in testing.
When I contacted Tesla this week to tell them my battery was failing as indicated by a steep range drop, the reply was: "We checked, and everything is fine." So I sent them these charts and asked the engineers to "please look again". The next day I got an email that they found a problem with my car, "independent of your range concern", and would replace my battery. That very same evening, my car abruptly failed to charge. Although there was no error showing on the console, the remote support folks could see an error when I plugged in the mobile charger while they were logged in - an error they had not encountered before, about an inability to balance the battery modules. The tow truck picked up the car at 7AM the next morning and had it back in my driveway at 7PM. Excellent service. Too bad I needed it. Glad I wasn't stranded (unless you count having to drive an ICE in the non-HOV lane being "stranded"). Worried it might happen again.
Yes, I got an A battery for each replacement. Yes, I asked for a newer battery this last time, but was told that was not possible and "really not necessary, as this re-manufactured battery had all new components". (I understand a new A battery is not the same as a new D battery, but there was no progress made in that argument.)
I have no idea why the data is so noisy until April 2014, but I suspect that after that date the firmware improved the reporting of range and SOC.
I thought the charts might be moderately interesting to others.
P