1 gid *mostly* equals 80Wh

[TickTock]

My recent observation that my car came back with more gids on an 80% and 100% charge while still drawing the same power from the wall led me to go back and check more logs on the value of 1 gid. Using three turtle to 100% logs (so non linearity won't be considered a factor), I computed the total energy into the battery as the integral of V*I*delta_t and compared to the power from the wall for the same charge. Then I divided both by the number of gids that charge added:

11/21/2011:
input:
gid start: 6
gid stop: 256
wall start(kWh): 508.93
wall stop(kWh): 531.57
battery start(kWh): 0
battery stop(kWh): 20.02
computed:
charger efficiency: 88.4%
wh/gid into battery: 80.0
wh/gid from wall: 90.6

7/6/2012:
input:
gid start: 6
gid stop: 217
wall start(kWh): 2857.34
wall stop(kWh): 2878.11
battery start(kWh): 0
battery stop(kWh): 18.50
computed:
charger efficiency: 89.1%
wh/gid into battery: 87.7
wh/gid from wall: 98.4

8/11/2012:
input:
gid start: 5
gid stop: 232
wall start(kWh): 2979.18
wall stop(kWh): 3000.01
battery start(kWh): 0
battery stop(kWh): 18.82
computed:
charger efficiency: 90.4%
wh/gid into battery: 82.3
wh/gid from wall: 91.8

I think since the efficiency (power_into_battery/power_from_wall) is pretty consistent, the power_into_battery computation should be valid --> although 80Wh may be the nominal target for 1 gid, it can be off by as much as 10%. Just something to keep in mind if you are counting gids and cutting your range close. Maybe my car just got out of cal and they recalibrated it for me when it was [Edit]at the Nissan Maricopa testing facility[/Edit] (between the 7/7 and 8/11 measurements), but I am going to be more cautious using a gid as an accurate unit moving forward.

 

[Stanton]

You keep excellent (historical) data. This is very interesting, but I'm beginning to wonder if/when we'll be able to get an "absolute" capacity reading on any given battery pack.

 

[smkettner]

I don't think battery capacity is as easy as measuring a gallon of gasoline. More like trying to measure a rubber band.

 

[RegGuheert]

You keep excellent (historical) data.

+1 ...and you publish it! Thanks, TickTock!

Just a few thoughts on this new data:

- It seems like the reports back from Nissan based on the Casa Grande testing are on the level. Hopefully we will hear more details from them soon.
- Based upon TickTock's measurements of energy accepted by the battery, it appears his battery has lost about 6% capacity in 9 months. This is 6% of the capacity 9 months ago, not original capacity. However, these two measurements are with the battery at different temperatures, so the number may be off (since it's possible the battery has different capacity at different temperatures). It would be interesting to have these measurements again in November.
- I wonder if the calibration had somehow degraded or if they have merely recalibrated at a different temperature point. I would suspect the latter, since I would think temperature compensation would handle temperature variations fairly well. So I'm wondering what may have moved off cal? I wouldn't expect voltage measurements to move much over time. Do the Hall-effect sensors degrade over time?
- The car will have to use something as ITS absolute energy reference, and I will assume that is GIDs. That implies that not only did the car think it had less energy available but that driving efficiency would be higher. Energy consumed reports by the car should also have been low.

Hopefully Nissan can use what they have learned to improve the accuracy of the instrumentation in the LEAF.

 

[jpa2825]

@TickTock -what is your method of charging? L2 I presume with that efficiency. I have heard tale of expecting 75% efficiency from wall to battery in other threads.

 

[TonyWilliams]

(since it's possible the battery has different capacity at different temperatures). It would be interesting to have these measurements again in November..

Waiting until November won't provide accurate data either, unless you think degradation STOPS, and will wait until November to then measure the battery at a lower temp. The only way to make that measurement is now (and attempt to prove the temp compensation theory), by cooling the battery.

 

[TickTock]

@TickTock -what is your method of charging? L2 I presume with that efficiency. I have heard tale of expecting 75% efficiency from wall to battery in other threads.

Yes. L2 with an AV EVSE and an EKM-25IDS power meter.

 

[TickTock]

You keep excellent (historical) data.

+1 ...and you publish it! Thanks, TickTock!

Just a few thoughts on this new data:

- It seems like the reports back from Nissan based on the Casa Grande testing are on the level. Hopefully we will hear more details from them soon.
- Based upon TickTock's measurements of energy accepted by the battery, it appears his battery has lost about 6% capacity in 9 months. This is 6% of the capacity 9 months ago, not original capacity. However, these two measurements are with the battery at different temperatures, so the number may be off (since it's possible the battery has different capacity at different temperatures). It would be interesting to have these measurements again in November.
- I wonder if the calibration had somehow degraded or if they have merely recalibrated at a different temperature point. I would suspect the latter, since I would think temperature compensation would handle temperature variations fairly well. So I'm wondering what may have moved off cal? I wouldn't expect voltage measurements to move much over time. Do the Hall-effect sensors degrade over time?
- The car will have to use something as ITS absolute energy reference, and I will assume that is GIDs. That implies that not only did the car think it had less energy available but that driving efficiency would be higher. Energy consumed reports by the car should also have been low.

Hopefully Nissan can use what they have learned to improve the accuracy of the instrumentation in the LEAF.

Thanks for all the encouragement. I do have a lot more data on incomplete charges (n-80%) just not many from turtle to 100 since I do try to avoid both of those levels. If I can get my boys off Minecraft for an hour or so I will plot wh/gid over time to see if it tracked temperature up until the service.

 

[Stanton]

Thanks for all the encouragement. I do have a lot more data on incomplete charges (n-80%) just not many from turtle to 100 since I do try to avoid both of those levels. If I can get my boys off Minecraft for an hour or so I will plot wh/gid over time to see if it tracked temperature up until the service.

I have that SAME problem with one of my sons :lol: Oh well, guess there's worse things they could be doing :roll:
I also wanted to comment that I appreciate the fact you post excellent info on a difficult subject without all the venom I see in some other posts/folks regarding the heat+battery issue.

 

[TickTock]

OK. I did see some temperature dependance on the gid unit. I added two more sets to my plot 0 to 100 and 0 to 80 wall energy. For these, I compute the wh/gid=delta_wall_power/delta_gids then multiply that by the final gid value to approximate the wall power required to go from zero to 80% or 100% charge. When I do this, the seasonality almost disappears and the loss over time is more gradual. Full disclosure: I did delete a few anomalous datapoints that were far off the trend. I am attributing those to days I ran the climate control to pre-heat/pre-cool the car and/or topped off away from home. I am missing a huge chunk of data for this computation during the winter (I got lazy and was only recording data in the morning so don't know what the starting gid value was).

Definitely much less alarming degradation than looking at gid count alone. I can now see why Nissan didn't want us looking at it.

 

[DaveinOlyWA]

can you provide a rough chart on capacity verses temp as a guideline? i realize that without knowing time frames that are needed for any real BMS change to occur, the data will be all over the map but i think you have really got something here

 

[edatoakrun]

="TickTock"

OK. I did see some temperature dependance on the gid unit. I added two more sets to my plot 0 to 100 and 0 to 80 wall energy. For these, I compute the wh/gid=delta_wall_power/delta_gids then multiply that by the final gid value to approximate the wall power required to go from zero to 80% or 100% charge. When I do this, the seasonality almost disappears and the loss over time is more gradual. Full disclosure: I did delete a few anomalous datapoints that were far off the trend. I am attributing those to days I ran the climate control to pre-heat/pre-cool the car and/or topped off away from home. I am missing a huge chunk of data for this computation during the winter (I got lazy and was only recording data in the morning so don't know what the starting gid value was).

Definitely much less alarming degradation than looking at gid count alone. I can now see why Nissan didn't want us looking at it.[/

And I still don't understand your method of calculating final gid value , TickTock.

Please explain.


Definitely much less alarming degradation than looking at gid count alone. I can now see why Nissan didn't want us looking at it.[/quote]

What degradation?

Doesn't this mean you believe that your battery may be accepting nearly the same kWh from the wall now, at both 80% and 100%, as it was when you began tracking it, on 10/01/2011?

 

[vegastar]

Tick Tock extrapolates the total energy from the wall from 0 to 80% or 100% by calculating the Wh/gid used in each charge session.

I noticed that in my case the value varies a lot. Sometimes the Wh (from the wall) per gid is BELOW 80. In 100% charges the value is always higher than on 80% charges.

Here is my data:

https://docs.google.com/spreadsheet/ccc?key=0AvuZhK4s7poqdEJfYTBCb3lwakZDNGI1cGNJUlFIUmc

 

[RegGuheert]

Thanks for the data, vegastar!

I noticed that in my case the value varies a lot. Sometimes the Wh (from the wall) per gid is BELOW 80. In 100% charges the value is always higher than on 80% charges.

I think that is to be expected since the LEAF does cell balancing at the top end. The result is that there will be additional energy dissipated in shunts during 100% charges.

 

[TickTock]

="TickTock"

OK. I did see some temperature dependance on the gid unit. I added two more sets to my plot 0 to 100 and 0 to 80 wall energy. For these, I compute the wh/gid=delta_wall_power/delta_gids then multiply that by the final gid value to approximate the wall power required to go from zero to 80% or 100% charge. When I do this, the seasonality almost disappears and the loss over time is more gradual. Full disclosure: I did delete a few anomalous datapoints that were far off the trend. I am attributing those to days I ran the climate control to pre-heat/pre-cool the car and/or topped off away from home. I am missing a huge chunk of data for this computation during the winter (I got lazy and was only recording data in the morning so don't know what the starting gid value was).

Definitely much less alarming degradation than looking at gid count alone. I can now see why Nissan didn't want us looking at it.[/

And I still don't understand your method of calculating final gid value , TickTock.

Please explain.


Definitely much less alarming degradation than looking at gid count alone. I can now see why Nissan didn't want us looking at it.[/quote]

What degradation?

Doesn't this mean you believe that your battery may be accepting nearly the same kWh from the wall now, at both 80% and 100%, as it was when you began tracking it, on 10/01/2011?

Vegastar summed it up pretty well. It looks like I am down 6-8% from when I started tracking last October but I was already down 10-12% by then (why I bought a gid-o-meter and started logging).

 

[vegastar]

The rate of degradation does seem to be slowing down, as from October to now we have more time than when Tick Tock's car was bought to last October, and less degradation.

 

[TickTock]

And I still don't understand your method of calculating final gid value , TickTock.

Final gid count is recorded from the canbus at the end of the charge (not calculated - at least not by me).

 

[edatoakrun]

I asked:

What degradation?

Doesn't this mean you believe that your battery may be accepting nearly the same kWh from the wall now, at both 80% and 100%, as it was when you began tracking it, on 10/01/2011?

...Vegastar summed it up pretty well. It looks like I am down 6-8% from when I started tracking last October but I was already down 10-12% by then (why I bought a gid-o-meter and started logging).

When you say, “down 6-8%”, I assume you are referring to the three charge sessions that you posted on P.1 of this thread?

I have been looking at your reports, in their entity.

https://docs.google.com/spreadsheet/ccc?key=0An7gtcYL2Oy0dHNwVmRkNkFnaEVOQTVENW5mOTZlb0E&pli=1#gid=1" onclick="window.open(this.href);return false;

Maybe I am reading your entries incorrectly. Do columns P and Q. “0 to 80 wall”, and “0 to 100 wall” not represent your kWh use from the wall, for 80% and 100% charges respectively? For example, the earliest entries in these columns for October 2011 are 16.41 and 20.15, and your most recent are 16.90 and 21.20 respectively. Column Q has no entry for the 11/21/11 charge session you report on P.1.

Are P and Q reported from a meter, or calculated by other means?

As I interpreted the data in those columns, as plotted on your “LEAF Battery Capacity” chart, I thought I was seeing an account of relatively stable capacity, with no clear indication of capacity reduction over the 10/1/11 to 8/11/12 time period.

I understand charging efficiency “from the wall” will vary, but I would not think it would distort the trend line so significantly, in this sample.

What am I getting wrong?



As to prior to October 2011, do you now believe you had 10-12% greater (new LEAF “normal”) capacity on delivery, and lost that entire amount by October?

IIRC, at one point you suspected that you'd had reduced capacity in your LEAF, from the time of delivery.

 

[TomT]

No longer having a Gid Meter, I decided I wanted to get an idea of where I was... Thus, last night, in the interest of science, I ran it down to turtle and charged to 100%
It took 21.1 Kwh which means I am down about 15% from what it was when new...

 

[TickTock]

Maybe I am reading your entries incorrectly. Do columns P and Q. “0 to 80/100 wall”, and “0 to 100 wall” not represent your kWh use from the wall, for 80% and 100% charges respectively? For example, the earliest entries in these columns for October 2011 are 16.41 and 20.15, and your most recent are 16.90 and 21.20 respectively. Column Q has no entry for the 11/21/11 charge session you report on P.1.

Sorry for the brevity of my descriptions - I try to keep it short so people will actually read but that often means lack of detail. The 0 to 80 wall, are extrapolated computations. Since the starting gid value varies from charge to charge, I extrapolate the power measured from the wall to estimate the power I would have consumed had that charge session start from 0 gids. In most cases I am pretty low on the gid count so the extrapolation should be reasonably accurate.

Are P and Q reported from a meter, or calculated by other means?

Yes (see above)

As I interpreted the data in those columns, as plotted on your “LEAF Battery Capacity” chart, I thought I was seeing an account of relatively stable capacity, with no clear indication of capacity reduction over the 10/1/11 to 8/11/12 time period.

Yes, much more stable, but it does appear to be slightly lower - although not nearly as low as the gid-count base estimation.

I understand charging efficiency “from the wall” will vary, but I would not think it would distort the trend line so significantly, in this sample.

No it doesn't change the trend and once I have more data leading into the winter, presumably the long-term trend will match (IOW, the gid base value should start recovering as it did last October). I removed some extreme out-lyers (like 0.25 wh/gid, etc.) which clearly has something else going on. When I started recording the data, I really didn't know what information would be important later (at the time I thought I was alone with the degraded capacity) so I didn't always comment when I ran the heater or ac to pre-condition the cabin, etc.

What am I getting wrong?

As to prior to October 2011, do you now believe you had 10-12% greater (new LEAF “normal”) capacity on delivery, and lost that entire amount by October?

IIRC, at one point you suspected that you'd had reduced capacity in your LEAF, from the time of delivery.

On this I am not certain. I didn't keep good logs immediately and as you recall from the discussions last October, I suspected my battery was delivered to me already with reduced capacity. However as others later started seeing degradation I think a more likely explanation is I saw rapid degradtion during those first few months of ownership when I was just driving and enjoying the car