DaveinOlyWA
Posts: 14101
Joined: Sat Apr 24, 2010 7:43 pm
Delivery Date: 16 Feb 2018
Leaf Number: 314199
Location: Olympia, WA
Contact: Website

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 6:33 am

LeftieBiker wrote:
SageBrush wrote:
LeftieBiker wrote:I was afraid of this. It seems more unpredictable than with the Canary packs, though. Some 30kwh packs lose no bars for years, while others - even in milder climate - lose a bar a year. Nissan is playing games again.
I'll guess not games, but pack variability. I think AESC QC is poor. It seems fairly clear that Nissan wanted to exit AESC and go with LG cells but they are stuck for now with a 25% ownership stake in AESC and an obligation to use AESC cell ... and even less control over QC than they had before. Not that QC was ever that good to begin with.
The "games" to which I refer are claiming that this is all just a BMS programming error, and that their BMS update will fix the problem - which never really existed according to them. I was suspicious of this from the first, because a factory programming error should have affected ALL of the BMS units manufactured in a certain time frame - not just some of them.
I had the rare 2016 S with 30 kwh pack and likely did not need the programming update. I was at 100% SOH the day my car was wrecked.
2011 SL; 44,598 miles. 2013 S; 44,840 miles.2016 S30 deceased. 29,413 miles. 2018 S40; 15,000 miles, 478 GIDs, 37.0 kwh 109.81 Ahr , SOH 94.61, Hx 120.15
My Blog; http://daveinolywa.blogspot.com" onclick="window.open(this.href);return false;

WetEV
Posts: 3041
Joined: Fri May 04, 2012 8:25 am
Delivery Date: 16 Feb 2014
Location: Near Seattle, WA

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 6:45 am

LeftieBiker wrote:
Do you, or have you done computer programming of embedded systems for a living, Leftie?
You are engaging in a logical fallacy that is very similar to the "appeal to authority" that argues that only those who have been officially titled as experts in a field can understand that field, and are thus the only people to be believed in any discussion of that field. Like Ford engineers, in a discussion of why Pintos were exploding in low speed accidents. Or Boeing engineers, in a discussion of why their planes only needed one airspeed sensor, and why a system relying entirely on that sensor should be able to override pilot input...
Or a doctor in a discussion as to that funny looking spot on your face being cancer or not.

Yes, experts are wrong some of the time. Airplanes would crash less if experts were always correct. Yet you can't design a 737, now can you? (And it wasn't an "airspeed sensor", it was an angle of attack sensor, and there are two of them, not one. And that might not have been the fatal problem. )

Yes, experts are wrong some of the time.

So are you.

This isn't an "appeal to authority", it is an appeal for humility.

The "games" to which I refer are claiming that this is all just a BMS programming error, and that their BMS update will fix the problem - which never really existed according to them. I was suspicious of this from the first, because a factory programming error should have affected ALL of the BMS units manufactured in a certain time frame - not just some of them.
Problems can be complex.

Perhaps there is a BMS programming error that is only triggered if the car is charging at midnight when daylight savings time starts. Or something odd like that.

There also may be packs with physical problems. As well as BMS programming error(s).

And there certainly is climate and usage variability. As well as physical problems in some packs and BMS programming error(s).

Failure analysis isn't easy, even with the absolute best in tools and full documentation of the system.

Quoting a boss of mine, "The last bug in a system is likely to be found on the day the very last copy of the system is powered down and hauled to the junkyard."
WetEV
#49
Most everything around here is wet during the rainy season. And the rainy season is long.
2012 Leaf SL Red (Totaled)
2014 Leaf SL Red
2019 eTron Blue

SageBrush
Posts: 4737
Joined: Sun Mar 06, 2011 2:28 am
Delivery Date: 13 Feb 2017
Location: NM

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 11:22 am

WetEV wrote: Yes, experts are wrong some of the time.
So are you.
Excellent post, so I'll nitpick:

Should be "Yes, experts are wrong some of the time,"
but MUCH less often than non-experts in their field.

HOWEVER, this dust-up is not experts Vs Lefty, it is Nissan Marketing Vs Lefty. Lefty has the advantage here.
2013 LEAF 'S' Model with QC & rear-view camera
Bought off-lease Jan 2017 from N. California
Two years in Colorado, now in NM
03/2018: 58 Ahr, 28k miles
11/2018: 56.16 Ahr, 30k miles
-----
2018 Tesla Model 3 LR, Delivered 6/2018

LeftieBiker
Moderator
Posts: 12977
Joined: Wed May 22, 2013 3:17 am
Delivery Date: 30 Apr 2018
Location: Upstate New York, US

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 1:49 pm

It would certainly be interesting to know what the actual Nissan engineers were saying about this, both when the issue was discovered and when the "fix" was developed. Were they fighting it tooth and nail and warning it wouldn't work most of the time, or were they on board with it...?
Scarlet Ember 2018 Leaf SL W/ Pro Pilot
2009 Vectrix VX-1 W/18 Leaf modules, & 3 EZIP E-bicycles.
BAFX OBDII Dongle
PLEASE don't PM me with Leaf questions. Just post in the topic that seems most appropriate.

goldbrick
Posts: 571
Joined: Tue Aug 01, 2017 3:33 pm
Delivery Date: 01 Aug 2017
Leaf Number: 311806
Location: Boulder, CO

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 1:55 pm

I do make my living programming embedded systems and I've been programming in one form or another for longer than I care to admit. And in that capacity, I'm not surprised that a programming error could exhibit the symptoms reported here.

One possible scenario - and this is totally hypothetical ! - out of the millions or billions of possibilities would be something like the code had a bug that caused the system that read the battery voltage to not recalibrate itself as intended. So if the ADC that converts the battery voltage to a number for the BMS changed over time (and virtually all silicon devices change characteristics over time, temperature, use, etc) then the voltage reported to the BMS would be incorrect. Now imagine that there are multiple vendors and batches of these ADC parts. Then different cars would exhibit different behaviors over time. Fixing the software to recalibrate the part as needed would eliminate the problem and the BMS would get a more accurate reading of the actual battery voltage, as it would now be unaffected by the errors in the ADC.

In this field, much or sometimes most of the actual firmware is used to calibrate, normalize and compensate for the varying characteristics of the mechanical parts and handling errors in the various sensors. These mechanical parts include the chips that read voltages, measure current, monitor tire rotations, etc. It's not a perfect world and there are no perfect parts. Once things are converted to numbers then usually, the final results are extremely consistent but even then there can be race conditions (where 2 operations happen in parallel and the answer depends on which finishes first), rounding errors, memory corruption, etc. that can cause inconsistencies.

So while I'm a bit skeptical of any marketing department, including Nissan's, I don't think the fact that different cars exhibited different symptoms means they are lying as I would not be at all surprised to see this behavior from a programming error.

LeftieBiker
Moderator
Posts: 12977
Joined: Wed May 22, 2013 3:17 am
Delivery Date: 30 Apr 2018
Location: Upstate New York, US

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 2:14 pm

Now imagine that there are multiple vendors and batches of these ADC parts.
That seems reasonable as a hypothesis, but I see a couple of problems with it. First, I don't think that Nissan uses multiple vendors for the BMS, based on their use of a single manufacturing plant for many years to provide them with their North American packs. If they had done so, though, why would the problem only appear in 2016, and if they first used multiple vendors then, why aren't they recalling just the packs using the marginal BMS units for the software fix? They should know which ones they are based on the VIN. The best theory here, IMO, is that there are both defective BMS units AND defective packs, and that Nissan has chosen to address only the former in order to save money. They probably figured that the bad packs would develop bad cells and be replaced for that reason, but it seems that some of them may instead degrade quickly. There is also the very real possibility that the 30kwh pack is susceptible to heat in much the same way as the original 'Canary' pack, and the BMS problem has contaminated the data enough for it to not yet be obvious.

All that we really know so far is that the BMS update doesn't fix all of the packs losing capacity, and that some of the packs don't lose capacity quickly even with no BMS update. It's certainly an interesting situation. I personally don't trust Nissan here: either they are fumbling around like clowns, or they are try to deal with a large problem with small, relatively inexpensive band-aids.
Scarlet Ember 2018 Leaf SL W/ Pro Pilot
2009 Vectrix VX-1 W/18 Leaf modules, & 3 EZIP E-bicycles.
BAFX OBDII Dongle
PLEASE don't PM me with Leaf questions. Just post in the topic that seems most appropriate.

SageBrush
Posts: 4737
Joined: Sun Mar 06, 2011 2:28 am
Delivery Date: 13 Feb 2017
Location: NM

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 3:20 pm

LeftieBiker wrote:The best theory here, IMO, is that there are both defective BMS units AND defective packs, and that Nissan has chosen to address only the former in order to save money.
1. Programming error
2. Poor QC
3. No TMS

Welcome all, to the LEAF life.
2013 LEAF 'S' Model with QC & rear-view camera
Bought off-lease Jan 2017 from N. California
Two years in Colorado, now in NM
03/2018: 58 Ahr, 28k miles
11/2018: 56.16 Ahr, 30k miles
-----
2018 Tesla Model 3 LR, Delivered 6/2018

WetEV
Posts: 3041
Joined: Fri May 04, 2012 8:25 am
Delivery Date: 16 Feb 2014
Location: Near Seattle, WA

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Mon May 13, 2019 3:59 pm

SageBrush wrote:
LeftieBiker wrote:The best theory here, IMO, is that there are both defective BMS units AND defective packs, and that Nissan has chosen to address only the former in order to save money.
1. Programming error
2. Poor QC
3. No TMS

Welcome all, to the LEAF life.
TMS is not desirable to everyone.

Unlike good QC and good programming.

How's your Tesla stock doing today?

-5.22%? So sad. :cry:
WetEV
#49
Most everything around here is wet during the rainy season. And the rainy season is long.
2012 Leaf SL Red (Totaled)
2014 Leaf SL Red
2019 eTron Blue

cwerdna
Posts: 9666
Joined: Fri Jun 03, 2011 4:31 pm
Delivery Date: 28 Jul 2013
Location: SF Bay Area, CA

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Tue May 14, 2019 12:37 am

SageBrush wrote:
LeftieBiker wrote:The best theory here, IMO, is that there are both defective BMS units AND defective packs, and that Nissan has chosen to address only the former in order to save money.
1. Programming error
2. Poor QC
3. No TMS

Welcome all, to the LEAF life.
1) As opposed to bugs on Model 3's that cause 12 volts to run dead and be stuck in someone's garage like https://teslamotorsclub.com/tmc/threads ... le.108412/? He originally used the term "bricked" which bent some people out of shape and spurred all sorts of discussion/debate.

And putting out updates that cause stuff like this:
https://teslamotorsclub.com/tmc/threads ... -3.134137/
https://teslamotorsclub.com/tmc/threads ... -2.152200/

2) As opposed delivering or attempting to deliver cars to customers with obvious external build quality problems (misaligned panels, uneven panel gaps and some bad paint jobs):
https://teslamotorsclub.com/tmc/threads ... es.124897/
https://teslamotorsclub.com/tmc/threads ... st-2928970
https://teslamotorsclub.com/tmc/threads ... st-3153143
https://teslamotorsclub.com/tmc/threads ... ar.128682/
along w/the plethora of people finding nibs in the paint, rejected deliveries, etc. I wish I could find the pics of the guy w/the black 3 where it looked like the paint guns had run out and there was large areas w/insufficient paint.

Here over a dozen folks w/failed Model 3 drive units within a certain window of time at https://teslamotorsclub.com/tmc/threads ... st-3516347. How many Leaf drive motors have been reported here replaced due to failure since Dec 2010?

Many Model 3 folks had defective turn signal modules which required replacement to correct: https://teslamotorsclub.com/tmc/threads ... le.129414/. My guess is that eventually, Tesla was more interested in shipping out the cars w/defective modules than holding them up for properly working ones.

'19 Bolt Premier
'13 Leaf SV w/premium package (owned)
'13 Leaf SV w/QC + LED & premium packages (lease over, car returned)

Please don't PM me with Leaf questions. Just post in the topic that seems most appropriate.

WetEV
Posts: 3041
Joined: Fri May 04, 2012 8:25 am
Delivery Date: 16 Feb 2014
Location: Near Seattle, WA

Re: 2016-2017 model year 30 kWh bar losers and capacity losses

Tue May 14, 2019 6:38 am

Then we better consider the fact that Musk brags about not using automotive grade electronic components. Such as the 17 inch display in the Model S.

https://www.thedrive.com/tech/27989/tes ... de-matters
WetEV
#49
Most everything around here is wet during the rainy season. And the rainy season is long.
2012 Leaf SL Red (Totaled)
2014 Leaf SL Red
2019 eTron Blue

Return to “Problems / Troubleshooting”