klapauzius
Posts: 1658
Joined: Thu Feb 24, 2011 4:19 pm
Delivery Date: 25 Jan 2011
Leaf Number: 0197
Location: Seattle, Wa

Re: Discuss data obtained from the LEAF Battery app

Sat May 25, 2013 9:48 pm

drees wrote:Ah readings appear to be directly related to GIDs.
If that was the case, then the simulated GIDs and the"real" Gids should match?
If I have to set a reference value for the simulated gids that is different from the reported AH value on the CAN bus , then they are not directly related?

Stoaty
Posts: 4490
Joined: Fri Jun 18, 2010 9:50 pm
Delivery Date: 12 Jun 2011
Leaf Number: 3871
Location: West Los Angeles

Re: Discuss data obtained from the LEAF Battery app

Wed May 29, 2013 9:08 pm

JPWhite wrote:Daily degradation appears to be .02 or .03 Ah. So once battery temp gets above 86 or so, the degradation accelerates by one order of magnitude !!

This is just one day, but it does look like single incidences can impact overall battery health.
It looks like I can keep the battery temperature to a max of 86 degrees as long as the ambient temperature for the day doesn't exceed 90 degrees and I park in the shade (based on test done today where max was 90 degrees). Last summer I wouldn't drive my Leaf to work if the high temperature was predicted to be 95 degrees or greater. I may have to consider lowering the threshold a bit.

Other data:

--2 hours of L2 charging started after the Leaf battery had reached ambient temperature of 67 degrees caused a rise in battery temperature of 3.5 degrees F. My rule of thumb will be: approximately 2 degree rise in battery temp for every hour of L2 charging.

--2 hours of cooling in parking garage when Leaf battery temperature was 10 degrees above ambient produced a 2 degree F. drop in battery temp.
Rule of thumb: about 1 degree F. cooling per hour with a 10 degree temperature differential. Obviously the rate of cooling will slow as the temperature differential decreases.
2011 Leaf with 62,000 miles given to Nephew
2013 Tesla Model S85 with 251 miles rated range at full charge
Leaf Spy Manual
Battery Aging Model Spreadsheet

edatoakrun
Posts: 5222
Joined: Thu Nov 11, 2010 9:33 am
Delivery Date: 15 May 2011
Leaf Number: 2184
Location: Shasta County, North California

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 8:46 am

LEAFfan wrote:
JPWhite wrote:I noticed I had lost quite a bit of battery capacity in short order. I found it was Thursday (May 16th) and did an analysis of what was going on to cause the 0.3 reduction in Ah in a single day.

As one might predict based on what happened (and is happening) in Arizona, high battery temp is the culprit.
...

Battery Ah never recovered overnight after cooling.

Daily degradation appears to be .02 or .03 Ah. So once battery temp gets above 86 or so, the degradation accelerates by one order of magnitude !!

This is just one day, but it does look like single incidences can impact overall battery health.
Yep, a lot of research shows battery degradation begins at 87F. The App should stay yellow (up to 86F) until the BT reaches 87F.
You are, IMO, in danger of repeating the experience of those watching their gids last spring, who equated the decline in gid counts due to rising ambient temperatures with battery capacity loss.

My indicated battery capacity has increased ~1% over the last few weeks (currently at 58.48 AHr and 88.26%) since my P 1 post below:
...My capacity reports at app-reported temps from ~59 f to ~86 F are extremely consistent, varying only ~0.20 % since the first capacity report I got at the SF BayLEAFs meeting over a month ago. Most recently:

Image

That being said, I think they are consistently wrong, in showing an ~12.5% loss of capacity, which has not shown up in my LEAF by either recharge or range capacity tests...
I don't think that my battery is healing itself and increasing in capacity. I just think the previously large error (likely a ~5% to ~10% overstatement) in capacity loss has decreased slightly.

My battery is ~20 f cooler today than it was in my post above, but the reported changes in indicated capacity do not seem to be reflecting a simple relationship with the current battery temperature.

Consistent with my p1 post, similar overnight temperature changes have not resulted in significant changes in the reported AHr/capacity variations, but the current increases may have been induced by my lower average battery temperatures over a longer period of time.
no condition is permanent

GregH
Posts: 860
Joined: Fri Jul 01, 2011 4:16 pm
Delivery Date: 13 Jun 2011
Leaf Number: 26967
Location: Irvine, CA

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 9:13 am

klapauzius wrote:Well, all I would care for is the AH value, so if it is indeed 67.57 Ah, that would be terrific.
So there were 2011 Leafs with higher than 66.2556 Ah? My understanding was that there werent GID readings higher than 281?
So how do GIDs and AH (apparently both from the CAN, albeit with different applications...the GID meter would read the CAN-GIDs, right?) reflect capacity then, if they are not the same?
The question is, is the AH reading a more accurate reflection of capacity than the CAN-GIDs? Or is it some fudged number that the car makes up from inaccurate and variable readings? You would think for coulomb counting the car would have to be discharged and charged completely?
67.57Ah is awesome! My 2012 started a tad above 66 but is now down to about 63.5 despite my efforts to baby it. :(
One of the LeafDD test pilots had a 2013 SL with 67.36Ah.. I'm jealous..

There were a few Gid readings over 281 but not many. Fixating on the number 281 is somewhat counter productive but understandable from the early days when folks were hunting for a true SOC percentage in the passive CAN data. As many have pointed out, Gids are actually more useful than SOC because they are constant units. On the LeafDD I offer an option to display the Ah Capacity * SOC which should give a similar scale (and is how Jim is calculating his estimated Gids on Android).

Yes the Ah cap is a better reflection of the pack capacity than Gids.. not just because of the higher precision but because Gids on a "full" charge are not consistent.. Actually Gids on an 80% charge would be better. For example sometimes I've seen my car charge to 94.6% with say 272 Gids and sometimes 95.6% with 275 Gids. So a better capacity measure would be something akin to Gids/SOC as "full" can be anything from 94% to 97%.. 80% is always 80% though. I used a similar measure to check Leafs before buying my 2012.. I think I found Gids/(SOC-3) worked best over a range of SOCs.. but why bother when we have a precise Ah capacity right from the car now? I've also found another number in the Group 1 data that seems like a percentage health number.. ie 100% for new and 70-80% for highly degraded cars. Although that 2013 I tested the other day came up at 110 so I'm not so sure what it is. Ah/Coulomb counting doesn't work so great on the Leaf because the current sensor is only precise to 0.5A and only accurate +/- 1.5A.. makes it very difficult to count Ahs when charging at 3.3kW!
'17 blue Volt Premier w/ACC
'12 SL black Leaf
'11 SL blue Leaf
RAV4-EV 2002-2005
Gen1 & Gen2 EV1 1997-2003
PV 2.4Kw, 10kWh lithium battery SCE TOU-DA

klapauzius
Posts: 1658
Joined: Thu Feb 24, 2011 4:19 pm
Delivery Date: 25 Jan 2011
Leaf Number: 0197
Location: Seattle, Wa

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 9:35 am

GregH wrote: Ah/Coulomb counting doesn't work so great on the Leaf because the current sensor is only precise to 0.5A and only accurate +/- 1.5A.. makes it very difficult to count Ahs when charging at 3.3kW!
If you integrate a noisy random variable over time, the sum becomes much less random. If you integrate e.g. 8 Amps with an error of +/- 1.5 A every second, the result will still be very accurate, because the error grows only with sqrt(N), where N is the number of measurements.
If your unit of measure is AH and you measure current with a random error of +/- 1.5 A every second (lets assume it is Gaussian with a standard deviation of 1.5 A) , your error on the AH measure is just 0.025 AH or 1.5 *sqrt(3600) =90 As. So Coulomb counting will actually work very well.

Anyway, does anyone have an idea how the AH number is calculated?
Maybe this has been discussed elsewhere already?

GregH
Posts: 860
Joined: Fri Jul 01, 2011 4:16 pm
Delivery Date: 13 Jun 2011
Leaf Number: 26967
Location: Irvine, CA

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 10:24 am

klapauzius wrote:
GregH wrote: Ah/Coulomb counting doesn't work so great on the Leaf because the current sensor is only precise to 0.5A and only accurate +/- 1.5A.. makes it very difficult to count Ahs when charging at 3.3kW!
If you integrate a noisy random variable over time, the sum becomes much less random. If you integrate e.g. 8 Amps with an error of +/- 1.5 A every second, the result will still be very accurate, because the error grows only with sqrt(N), where N is the number of measurements.
If your unit of measure is AH and you measure current with a random error of +/- 1.5 A every second (lets assume it is Gaussian with a standard deviation of 1.5 A) , your error on the AH measure is just 0.025 AH or 1.5 *sqrt(3600) =90 As. So Coulomb counting will actually work very well.

Anyway, does anyone have an idea how the AH number is calculated?
Maybe this has been discussed elsewhere already?
When my 2011 Leaf was ON, not moving I'd typically see about 1A draw. When my new 2012 Leaf is ON, not moving I typically see about 0A draw. I haven't seen any indication of self calibration in the BMS (you'd think it wouldn't be very hard.. just measure after the contactors open or before they close). Without calibration the difference of 1A over the course of a 5 hour charge at 8-9A can be rather large. We haven't found anything akin to a running Ah counter in the data so far.
How the car calculates the Ah capacity (or Gids for that matter) is not entirely understood... but it seems to be a mixture of Ah counting and voltage and temperature analysis..
[edit] Actually I should amend that.. Doing a bunch of partial charges on the new car (say from 30% to 70% up and down) does introduce a drift in the SOC such that at lower voltages sometimes the car snaps down to the lower Gid/SOC over the course of a few minutes when it realizes it's off. I've seen this a few times and it's somewhat alarming. Now I pay more attention to voltage than Gids when I'm under LBW.
'17 blue Volt Premier w/ACC
'12 SL black Leaf
'11 SL blue Leaf
RAV4-EV 2002-2005
Gen1 & Gen2 EV1 1997-2003
PV 2.4Kw, 10kWh lithium battery SCE TOU-DA

GregH
Posts: 860
Joined: Fri Jul 01, 2011 4:16 pm
Delivery Date: 13 Jun 2011
Leaf Number: 26967
Location: Irvine, CA

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 10:42 am

klapauzius wrote: Anyway, does anyone have an idea how the AH number is calculated?
Maybe this has been discussed elsewhere already?
We only deciphered this portion of Group 1 about 6 weeks ago.. See the Active CAN sampling thread for a history..
'17 blue Volt Premier w/ACC
'12 SL black Leaf
'11 SL blue Leaf
RAV4-EV 2002-2005
Gen1 & Gen2 EV1 1997-2003
PV 2.4Kw, 10kWh lithium battery SCE TOU-DA

lukati
Posts: 288
Joined: Wed Aug 17, 2011 3:23 pm
Delivery Date: 07 Dec 2011

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 11:06 am

Now that we can read battery temperature accurately, an interesting experiment can be done. Park a black and a white car next to each other in a sunny lot and determine the battery temperature as a function of time. That would tell us whether it is a good idea to buy a light color Leaf in warm climates.

GregH
Posts: 860
Joined: Fri Jul 01, 2011 4:16 pm
Delivery Date: 13 Jun 2011
Leaf Number: 26967
Location: Irvine, CA

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 11:18 am

lukati wrote:Now that we can read battery temperature accurately, an interesting experiment can be done. Park a black and a white car next to each other in a sunny lot and determine the battery temperature as a function of time. That would tell us whether it is a good idea to buy a light color Leaf in warm climates.
Since I've just invested in a black Leaf :shock: I'm actually somewhat pleased to report that the insulation between the battery and the passenger compartment works pretty well. I've come out to a baking car after an hour of direct sunlight and found the temperatures still calm.. mostly because it WAS in the shade and thus the pavement under the car was still cool. I actually notice slight cooling during the morning commute which tells me cooling the underside of the car is more important than cooling the passenger compartment. Kinda flies in the face of thermodynamics, but that's what I've seen so far. Yes, though, that would be an interesting test.
'17 blue Volt Premier w/ACC
'12 SL black Leaf
'11 SL blue Leaf
RAV4-EV 2002-2005
Gen1 & Gen2 EV1 1997-2003
PV 2.4Kw, 10kWh lithium battery SCE TOU-DA

Stoaty
Posts: 4490
Joined: Fri Jun 18, 2010 9:50 pm
Delivery Date: 12 Jun 2011
Leaf Number: 3871
Location: West Los Angeles

Re: Discuss data obtained from the LEAF Battery app

Thu May 30, 2013 11:27 am

GregH wrote:I actually notice slight cooling during the morning commute which tells me cooling the underside of the car is more important than cooling the passenger compartment. Kinda flies in the face of thermodynamics, but that's what I've seen so far.
Makes me wonder if putting a small fan under the car at night to circulate air better would improve night time cooling during the summer months when I need it most. The air in my underground parking garage doesn't move much.
2011 Leaf with 62,000 miles given to Nephew
2013 Tesla Model S85 with 251 miles rated range at full charge
Leaf Spy Manual
Battery Aging Model Spreadsheet

Return to “LEAF CANBus”