The Anatomy of a Gid

[Stoaty]

Last Friday the new Leaf at 640 miles gave me a 278 at 100% even after running the CC for 30 minutes while plugged in.
This morning the old Leaf at over 23000 miles gave me 281 at 100%. I know from the turtle to 100% tests done a week ago that the old Leaf needed about 6-7% less kWh from the wall, comparing the the new one.

Sounds like all "Gids are created equal, but some are more equal than others." I don't know why some Leafs that have clearly lost capacity still read 281 on a full charge, but it certainly shows that Gid count is only a rough measure of remaining capacity.

 

[edatoakrun]

I now have the possibility of measuring the GIDs.

Last Friday the new Leaf at 640 miles gave me a 278 at 100% even after running the CC for 30 minutes while plugged in.
This morning the old Leaf at over 23000 miles gave me 281 at 100%. I know from the turtle to 100% tests done a week ago that the old Leaf needed about 6-7% less kWh from the wall, comparing the the new one. It was a bit warmer outside this morning, but the cars still showed around 60F in the garage.

I will have to do some reading on what the GIDs are, but I just tought I put this data here.

Thanks for posting this information.

The Phoenix range test showed gids were an extremely unreliable indication of battery capacity in hot climates, overstating capacity loss, so I don't think it should be any great surprise that gids may understate capacity loss in LEAFs with less exposure to higher battery temperatures.

Could you please comment on the accuracy of the m/kWh reports from your two LEAFs?

Several months back I first noticed the following results from my LEAF, which has significant exposure to high ambient temperatures:

Ever since TickTock fist suggested the topic of gid variability, the implications of his observations have been setting in.

TickTock

1 gid *mostly* equals 80Wh


http://www.mynissanleaf.com/viewtopic.php?f=44&t=9689" onclick="window.open(this.href);return false;

If “gids” do reflect a variable amount of Wh, and they are the values used by the LEAF to calculate kWh use, then the capacity bar displays, dash and nav screen displays of m/kWh, as well as the Carwings calculations based on these same “gid” values, might be expected to be incorrect as well.

I now believe that this quite possibly could be be the case.

I have been noticing an unexplained increase in my dash, nav Screen, and Carwings m/kWh results for a few months now, not only on this test trip, but on other trips, and my long term m/kWh averages...

I think that my range tests may indicate that whatever method my LEAF uses to calculate kWh, is variable, and has been significantly understating the recent amounts of kWh use, and has probably increasingly inflated all my m/kWh reports, from the dash, nav screen, and CW.

And of course, this could reflect with TickTock's observations of variable “gid” Wh values. Gids with higher Wh content could lower the calculated kWh numbers, and raise all the m/kWh results.

Maybe this is what I am seeing, from yesterdays range test. I tried to replicate as accurately as possible, my earliest range test,of almost a year, and almost 10,000 miles ago, to test this hypothesis.

I chose a day with very close to the original temperature condition, and drove the exact same route over the first 87 miles of the trip, using the same mode (eco) and used my original trip logs to closely replicate the same elapsed times for each of the three (same distance) legs of the trip.

The results from 8/30/12 were:

97.3 miles to VLB, 98.9 miles in total, by the odometer.

CW: 96.5 (~2.5% under-report) total miles, at 5.7 m/kWh, 16.8 kWh used from 100% to about the same capacity level, slightly past VLBW.

Compare this test with my first test on 9/7/11:

91.5 miles to VLB, 93.4 in total, by the odometer

CW: 91.1 (~2.5% under-report) total miles, at 4.9 m/kWh, 18.7 kWh used from 100% to about the same capacity level, slightly past VLBW.

I do not believe that the slight increase in range over the last year reflects any increase in battery capacity. On the contrary, I expect that my total capacity (though maybe not the amount of kWh that the BMS is allowing me to access) has declined by an undetermined amount, but it cannot be detected due to the “noise” of uncontrolled variables in a range test.

But I think the decrease of over 10% of reported kWh use, is simply too great to be consistent...

http://www.mynissanleaf.com/viewtopic.php?f=31&t=9064&start=20" onclick="window.open(this.href);return false;

The Arizona range test showed m/kWh report variability similar to what I had reported earlier, but the data from that test, IMO, is not accurate or reliable enough to offer much help in determining the relationship between gid report error and LEAF kWh use report error.

 

[camasleaf]

Could you please comment on the accuracy of the m/kWh reports from your two LEAFs?

When I turtle the two Leafs a week ago they reached VLBW in totaly different conditions. The new one because I was not fully charged before I had to run an unexpected errand, the old one from 80% by doing a 45miles RT at maximum speed (without getting a fine) and with the heat at full blast/window open on the return trip. I usually do that trip and come home with at least 3 bars. From VLBW to turtle the route was similar, and I recall that there was some unexpected difference on the m/kWh between the LBW and turtle. (I did reset miles and m/kWh at LBW since I know I can count on about 9 miles).

I will have to do another test in about two weeks and I will try to record more data. The charging kWh were recorded from the Blink charger.

 

[DaveinOlyWA]

I now have the possibility of measuring the GIDs.

Last Friday the new Leaf at 640 miles gave me a 278 at 100% even after running the CC for 30 minutes while plugged in.
This morning the old Leaf at over 23000 miles gave me 281 at 100%. I know from the turtle to 100% tests done a week ago that the old Leaf needed about 6-7% less kWh from the wall, comparing the the new one. It was a bit warmer outside this morning, but the cars still showed around 60F in the garage.

I will have to do some reading on what the GIDs are, but I just tought I put this data here.

so you readded the same # of GIDs to both cars and the older one took less from the wall?

well, that probably means you have range degradation that is undetectable? "drift adjustment" not done, or could it be possible your older battery pack is colder?

couple things to mention about my heat observations. the pack "seems" to react slowly to ambient temperature changes.

for example; have seen TBs that i know should not have been a day after the fact. Like 6 TBs at this time of year is possible with quick charging and what not but to see 6 TBs the next day from a quick charge session means that heat seems to dissipate very slowly.

i "think" in your situation, you might be seeing a colder pack in the older LEAF. is it possible that one sat in the garage for a significant amount of time longer than the other one did?

i think a 6-7% difference especially if recharging a lot like 60-80% SOC could be accounted for with a thermal explanation


**edit** what i would do is record your recharge amounts and pair it with how far you drove and the miles/kw reading. reset both after each charge. as far as charge level. 100% or 80%? i dont think it matters. the skinny is that 80% is a more consistent charge but i think the difference shouldnt be a problem

 

[edatoakrun]

Could you please comment on the accuracy of the m/kWh reports from your two LEAFs?

When I turtle the two Leafs a week ago they reached VLBW in totaly different conditions. The new one because I was not fully charged before I had to run an unexpected errand, the old one from 80% by doing a 45miles RT at maximum speed (without getting a fine) and with the heat at full blast/window open on the return trip. I usually do that trip and come home with at least 3 bars. From VLBW to turtle the route was similar, and I recall that there was some unexpected difference on the m/kWh between the LBW and turtle. (I did reset miles and m/kWh at LBW since I know I can count on about 9 miles).

I will have to do another test in about two weeks and I will try to record more data. The charging kWh were recorded from the Blink charger.

IMO, it probably doesn't matter that the driving conditions differ, as long as the SOC at the beginning and end, and the battery temperature during charging, are held constant.

The question is how closely your two LEAFs' internal kWh use reports match the recharge results you get from your Blink.

By far the easiest and (probably) the most accurate way to get the kWh use report from your LEAF is from (updated) Carwings.

If you can't access CW, calculations using the nav screen m/kWh (or dash m/kWh multiplied by ~1.025) and odometer miles should give you the same results.

 

[surfingslovak]

so you readded the same # of GIDs to both cars and the older one took less from the wall?

well, that probably means you have range degradation that is undetectable? "drift adjustment" not done, or could it be possible your older battery pack is colder?

I made the exact same observation like camasleaf this past April. The Gid meter measured 281 Gids, yet the car seemed to suffer from 5% range loss. The temperatures was mild then, and the car is parked in a garage at work all day, where is can balance properly on level 2. To be honest with you, I'm surprised to see all the rationalizations people come up with just to be able to say that they don't see any degradation. It's natural that there will be some, and it should be expected. The only question is how much and how soon.

had a pretty large temperature drop last night so ran LEAF to "near" LBW (27 GID)

and plugged in. put in almost 20 kwh in BUT it charged to 281 (which dropped to 280 after car was on about 2-3 minutes) gaining 254 GID which equates to 78.7 watts per GID.

soooo, any explanation would be good

Sorry, I missed this. A new battery at 70 F should take about 22 kWh from the wall to go from VLB to a full charge. You cannot disregard charger and battery (heat) losses when calculating the amount of available energy stored in the car.

 

[DaveinOlyWA]

the thread is new but these observations are not. if you like i can post several links to posts, blogs, etc. i have written discussing the same thing.

and rationalizations is really the only tool i have. i "thought" i had other tools but they seem to measure a unit i dont understand. we have all seen several posts by reputable people using accepted methods and equipment to provide us data that gives us values of a GID. but the weather is getting colder and the value of my GID is dropping causing me to have less range than the car tells me i should have?

but the range is more than 84 miles but it should be i guess because i am getting 4.8 miles /Kwh which even if i am only storing a reduced GID and only have 18.5 Kwh to burn that should give me 88 miles?

so i have full confidence a GID is 80 wh but only in a narrow range of conditions...like 70 º or whatever batteries are measured at

and i am guessing you meant "LBW" and not "VLB"

 

[surfingslovak]

Sorry Dave, I meant exactly what I said. Based on the data you posted on this board it's very likely that you are seeing about 5% degradation. I'm no Phil, and contrary to the appearances, don't really have a whole day to argue on this board. Without giving you my full bio, let me just mention that I worked in a lab at a major automaker which was responsible for environmental simulation and signed off parts and subsystems for production release. I consider myself reputable, and I usually can put one and two together. I would also like to remind you that I was one of the few voices saying that Gids will recover in cold temps based on TickTock's data when it was not well accepted yet. Although this has come to pass, I don't think that it means very much. I'm sorry. I have looked at your blog on occasion and corrected a math mistake once, I believe. I respect what you do and the committment that it takes over the years.

That said, we seem to come from different angles. I'm a technologist, and want things to work well, or at least as advertised. I live and die by the spec. I don't get EV advocates that will go to great lengths to defend something that might not work as expected, because it's inherently good for us. What I would expect instead is that an issue is investigated, and a remedy is identified. That can be either a technical fix or something creative on the business side of things.

 

[LEAFer]

Could you please comment on the accuracy of the m/kWh reports from your two LEAFs?

When I turtle the two Leafs a week ago they reached VLBW in totaly different conditions. The new one because I was not fully charged before I had to run an unexpected errand, the old one from 80% by doing a 45miles RT at maximum speed (without getting a fine) and with the heat at full blast/window open on the return trip. I usually do that trip and come home with at least 3 bars. From VLBW to turtle the route was similar, and I recall that there was some unexpected difference on the m/kWh between the LBW and turtle. (I did reset miles and m/kWh at LBW since I know I can count on about 9 miles).

I will have to do another test in about two weeks and I will try to record more data. The charging kWh were recorded from the Blink charger.

IMO, it probably doesn't matter that the driving conditions differ, as long as the SOC at the beginning and end, and the battery temperature during charging, are held constant.

The question is how closely your two LEAFs' internal kWh use reports match the recharge results you get from your Blink.

By far the easiest and (probably) the most accurate way to get the kWh use report from your LEAF is from (updated) Carwings.

If you can't access CW, calculations using the nav screen m/kWh (or dash m/kWh multiplied by ~1.025) and odometer miles should give you the same results.

I would like to provide some data, because I'm not so sure that CarWings can be trusted.

Now, granted, the LEAF hasn't been in since Jan 2012, and does not have the door-chime update that came out shortly afterwards. But I do believe it has all the CarWings stuff, BUT I COULD BE WRONG.

See pictures below. They are the CW data from the two Turtle/Dead-to-100% measurements I took this month. Does not look like anything accurate to me. (Yes, 10/13 must have "declined" one or two trips. Is that the cause for the 9 m/kWh ??)

On 10/13 I drove a 100% LEAF to the Folsom meetup. Shuttled some other drivers, stopped for some shopping and went home to turtle it. Actual Odometer miles: 72.2. kWh to 100%: 20.722. See HERE.

In the morning of 10/24/2012 our LEAF was fully charged (100%). It was NOT charged again the next night. For 10/24 & 10/25 Carwings reports 4 + 2 trips, reaching "dead" as reported HERE. Actual Odometer miles: unknown, but 75.7 seems correct. kWh to 100%: 21.049.

 

[DaveinOlyWA]

Sorry Dave, I meant exactly what I said. Based on the data you posted on this board it's very likely that you are seeing about 5% degradation. I'm no Phil, and contrary to the appearances otherwise, don't really have a whole day to argue on this board. Without giving you my full bio, let me just mention that I worked in a lab at a major automaker which was responsible for environmental simulation and signed off parts and subsystems for production. I consider myself reputable, and I usually can put one and one together. I would also like to remind you that I was one of the few voices saying that Gids will recover in cold temps based on TickTock's data when it was not well accepted yet. Although this has come to pass, I don't think that it means very much. I'm sorry. I have looked at your blog on occasion and corrected a math mistake once, I believe. I respect what you do and the committment that it takes over the years.

That said, we seem to come from different angles. I'm a technologist, and want things to work well, or at least as advertised. I live and die by the spec. I don't get EV advocates that will go to great lengths to defend something that might not work as expected, because it's inherently good for us. What I would expect instead is that an issues is investigated, and a remedy is identified. That can be either a technical fix or something creative on the business side of things.

i agree with your estimates. i only wish i had ventured further into the bottom of the pack when it was new so i could have a better picture of what is going on. what i have found is that the more experience i get with my LEAF, the more subtle adjustments i am making to my driving style to make it work

 

[edatoakrun]

="LEAFer"

="edatoakrun"

IMO, it probably doesn't matter that the driving conditions differ, as long as the SOC at the beginning and end, and the battery temperature during charging, are held constant.

The question is how closely your two LEAFs' internal kWh use reports match the recharge results you get from your Blink.

By far the easiest and (probably) the most accurate way to get the kWh use report from your LEAF is from (updated) Carwings.

If you can't access CW, calculations using the nav screen m/kWh (or dash m/kWh multiplied by ~1.025) and odometer miles should give you the same results.

I would like to provide some data, because I'm not so sure that CarWings can be trusted...

Then just calculate your LEAF's kWh use.

I'd suggest again (as I have been for the last 14 months) that everyone get CW updated.

Since most have not been monitoring their LEAFs m/kWh for error, we are now stuck on much the same questions I asked in August 2011...

Post subject: My carwings energy numbers - CORRECT post NTB11-041 update

Posted: Sun Aug 28, 2011 7:41 pm

As I posted earlier, Since I had the Carwings update done on 8/3, my Dash and Carwings miles/kWh numbers seem to match.

Even more interestingly, the daily Driving Records/electricity consumption now seem to accurately reflect the kWh delivered from my Modified L2, as best as I can calculate by recharge time.

Has anyone tried a 100% charge to Turtle discharge drive since having the update?

What total electricity consumption (kWh) did carwings report-and do you believe it to accurately reflect total battery capacity?

If you have a meter at the wall, what L2 charging efficiency percentage did it show for your recharge, using the Carwings energy consumption report?

http://www.mynissanleaf.com/viewtopic.php?f=31&t=5423" onclick="window.open(this.href);return false;

 

[LEAFer]

@Ed: You posted your CW update was done 8/3/11. Since I was at the dealer for our first annual battery check on 1/10/2012, shouldn't the dealer have installed any and all firmware updates, whether specifically requested or not ? (The service records for 1/10/12 do not show a F/W update. The service records for 7/1/11 show "Recall P1213 LEAF VCM reprogram NTB11-038", plus a new (less noisy) charger was installed. The GOM behaviour changed after the 7/1/11 service.) Or is there a procedure to verify which CW (accurate?) our LEAF has installed ?

I could wait 2.7 months when our next Battery Check is due to ensure our CW is updated, rather than speculating on that. Or I could do an extra trip to the dealer, now that I've gathered enough data (although I'd still like to try a borrowed Gid Meter).

But don't you think it's ridiculous CW telling me the car traveled 75.7 miles (highly plausible) and used (net) 13.7 kWh from 100%-to-Dead (5.5 mi/kWh = plausible), versus re-charging to 100% taking 21.049 kWh ? ( That would be an L2 charging efficiency factor from the wall/battery use of 65% ! ) Of course, if my capacity is degraded to 85% that efficiency factor would be 76.5%, which is still VERY poor (and unlikely).

So, again, is there a way to compare CW version to your/other LEAFs ?

 

[grommet]

The CARWINGS related telematics update is not a mandatory fix; the kWh undereporting "bug" isn't even documented directly. The only reason a fix even exists is that it just happened to be included in last year's telematics connection problem TSB, which not everyone asked for.

If your Boot Ware & Application are at 147 or up, you don't have the kWh reporting bug:
http://www.mynissanleaf.com/viewtopic.php?f=27&t=5176" onclick="window.open(this.href);return false;

--Ron

 

[LEAFer]

The CARWINGS related telematics update is not a mandatory fix; the kWh undereporting "bug" isn't even documented directly. The only reason a fix even exists is that it just happened to be included in last year's telematics connection problem TSB, which not everyone asked for.

If your Boot Ware & Application are at 147 or up, you don't have the kWh reporting bug:
http://www.mynissanleaf.com/viewtopic.php?f=27&t=5176" onclick="window.open(this.href);return false;

--Ron

Thanks, Ron. I'll check car this evening.

 

[DaveinOlyWA]

The CARWINGS related telematics update is not a mandatory fix; the kWh undereporting "bug" isn't even documented directly. The only reason a fix even exists is that it just happened to be included in last year's telematics connection problem TSB, which not everyone asked for.

If your Boot Ware & Application are at 147 or up, you don't have the kWh reporting bug:
http://www.mynissanleaf.com/viewtopic.php?f=27&t=5176" onclick="window.open(this.href);return false;

--Ron

i went in requesting the carwings fix when i got the "beep nearly all the time" fix and was told my VIN was not eligible...

 

[grommet]

Sadly, then you need to "convince" them to apply it. I know a few people that had to song and dance their dealer to do it anyway. The problem is that the dealer generally doesn't get paid for optional work that isn't officially covered by a TSB.

Feel free to escalate the problem.

 

[camasleaf]

i "think" in your situation, you might be seeing a colder pack in the older LEAF. is it possible that one sat in the garage for a significant amount of time longer than the other one did?

I do not believe the temperature difference was too high. Both cars were in the garage Friday afternoon after light city driving. The old one was charged at 80% by Friday evening, the new one at 100% by 3:30 AM on Saturday. I did the measurement at 3:30 am and at 7:30 am after 30 minute of preheating. The old one had no driving on Saturday and just 10 miles on Sunday. It was charged at 100% on Monday morning when I measured the GID. All this time the temperature in the garage was around 60F and it was colder outside.

Any way, I need a good test plan for next weekend. Any suggestions are welcome. I am thinking of charging both at 80%, but one will have to be on L1. Drive them both exactly the same route (about 40 miles) and record miles, miles/kWh, and come home and charge them one at the time (one Blink only) and see what the wall consumption will be to get them back to 80%. I checked the tires last week but I will check them again on Saturday morning. All this if my wife approves it.

As for Carwings, I have not looked at it for months. Looks like there is an updated Carwings, I will look into that. The old car was at the dealer in June and it had firmware upgrades.

 

[LEAFer]

i went in requesting the carwings fix when i got the "beep nearly all the time" fix and was told my VIN was not eligible...

I'll try

Here's my Version Info:

Boot Ware (NK1):064
Application (NK2):064
Audio Unit Software:067
CAN uCOM Software:018
Front Display Software:030
BOLERO Software:00.05
Bluetooth Firmware:C230
Voice Recognition Engine:925-123-000
Voice Synthesis Engine:5.0.0.3-PS-P2
Voice Recognition Grammar Version:00001080
Voice Synthesis Data Version:00003000
Record Data Version:00001026
Part Number:3NA0A
XM Satellite Radio Software:012
Switch Software:030000
TCU Software:3NA0000622
Map Version:10/08/12/01

Looks like there's some OLD STUFF in there :shock: That would explain my CW MIS-TRUST !

 

[DaveinOlyWA]

i "think" in your situation, you might be seeing a colder pack in the older LEAF. is it possible that one sat in the garage for a significant amount of time longer than the other one did?

I do not believe the temperature difference was too high. Both cars were in the garage Friday afternoon after light city driving. The old one was charged at 80% by Friday evening, the new one at 100% by 3:30 AM on Saturday. I did the measurement at 3:30 am and at 7:30 am after 30 minute of preheating. The old one had no driving on Saturday and just 10 miles on Sunday. It was charged at 100% on Monday morning when I measured the GID. All this time the temperature in the garage was around 60F and it was colder outside.

Any way, I need a good test plan for next weekend. Any suggestions are welcome. I am thinking of charging both at 80%, but one will have to be on L1. Drive them both exactly the same route (about 40 miles) and record miles, miles/kWh, and come home and charge them one at the time (one Blink only) and see what the wall consumption will be to get them back to 80%. I checked the tires last week but I will check them again on Saturday morning. All this if my wife approves it.

As for Carwings, I have not looked at it for months. Looks like there is an updated Carwings, I will look into that. The old car was at the dealer in June and it had firmware upgrades.

i would just collect data as much as you can. i reset my trip A and miles/kwh on the dash every day. this allows me to track how far i am getting and what it costs me to do it.

as far as range tests? i am not in the camp that believes any one run will do it. just too much variables and its usually not convenient to do a half dozen runs unless you happen to have a commute or regular trip you make that is in the 60-80 mile range but if you can do at least 2-3 fairly repeatable runs to at least VLB (2nd warning) that should be enough to tell you where you are at. just make sure you are within 5 miles of home

 

[brettcgb]

The relatively flat dV/dT curve of Lithium-Ion is why coulomb counting is often used rather than voltage mapping . This is how the Leaf's battery management system (LBC) tracks SoC. The problem is that the low-cost sensor used is subject to drift and thus requires a periodic correction. Since it's not essential that the SoC be perfectly accurate in the middle area of SoC, there is not as much need to issue corrections. This works out fine, as it's difficult to do with such a flat curve. However, near either end, the voltage mapping can be much more effective, so you are more likely so see the corrections.

There is no equalization used, as the active balancing system works to keep the cells in-balance. (At least up to a point) It's designed to be able to keep cells balanced within a wide tolerance of cell condition, but the pack is only as good as it's weakest cell.

-Phil

I think you may be onto something with respect to the current sensor. I've not worked with sensors for large currents, but any sensor that covers a range of +125A..0..-225A with reasonable resolution has got to be a bit finicky. (What is reasonable? A 14-bit ADC with healthy filtering might provide 20mA resolution... with luck. (I'm a PIC18 developer with some CAN bus training, but very little CAN experience.))

I can't help but think that what you've been calling "corrections" are not adjustments to calculated vales, but may instead be outright resets based on hitting certain pack voltages or slopes, adjusted for battery load (using battery resistance and measured current). "Resets" in this case would be "forget the current SoC, here is the new SoC". Points of interest would be the low SoC knee on discharge, and the high SoC knee on charge.

When monitoring SoC during discharge (a long drive), do you see abrupt changes in SoC (skipping over a few GIDs), particularly when you are in the low SoC knee? Any correlation between these changes and when LBW/VLB/Turtle events occur? Can you depend on consistent GIDs at LBW/VLB/Turtle events? (Similar questions exist for charging near 80% and 100% points.)

Soon after the Leaf came out, you investigated the temperature display operation, and found that the display did not immediately display the temperatures measured by the sensor (I think that was you). Do you see something like that for the SoC bargraph?

At Leaf start-up, I suspect there is a time when there is heavy communications with the battery before it provides any power to the vehicle. That would be a good time to capture an "open circuit" or zero-current reading of the current sensor bias. Does that bias change after a short charge/discharge?