A bad cell and a reduce range

[Pier]

A bad cell and a reduce range

My 2012 I-MIEV, 2 years old, only 16,000 km, seemed to have no reduce range until few weeks ago. Perfect battery voltage on any cell at full charge of 4.105 volt.

Then, after full recharge, only 15 bars and 92% SOC, and cell #85 lower than all others. Every recharge, the soc went down untill it reach 75%, and 3 bars lost. Moreover, the recharge took 3 hours more to finalize. Before, the recharge ended at around 460 Watt at the wall and now it last for ever, between 200 and 400 W. ( something around or below 100W in the car).

I e-mailed Mitsu Canada those details and also to the nearest dealer. After a week, now my car is at the garage (dealership) for test and I hope repair. It was clear for me that cell #85 was at fault, but of course they do not need my advice !! They tried to find if I had any wrong doing. Then Mitsu asked the garage to extract all the data from the computer, like the way I was driving, flat or hilly, fast acceleration, etc... they told me I was driving conservatively which is true. I took care of my battery as if it was a newborn baby, never allowing it to go over 30°C at any time and driving mostly between 20% and 80% basttery capacity.

So what went wrong ?

Anywone experienced a faulty cell ?

Can you tell if a voltage of 4.105 is a sign of no deterioration of battery cell ?

 

[PV1]

Although they have to unequivocally prove it to be at fault, I'd be cautious of mentioning CaniOn to the dealer, as they may try to deny a warranty claim, and the battery warranty isn't one to gamble with.

Where are you located? What is the lowest temperature your car has experienced? How often have you allowed a full charge?

Based on earlier failures (there have been 3 or 4 on this forum, same symptoms as yours), I'm wondering if that cell got too cold during a charge and has some plating issues which developed a short. It seems that your car extended balancing time to try correcting the mis-balance, but the cell continues to self-discharge.

Not that it matters to us as far as replacing stuff, but I wonder if a balancer develops a permanent short and discharges one cell continuously. This issue seems to happen suddenly. A failure in a cell (I would think) would gradually show or instantly become a dead short and totally kill the cell. Given that the car can't correct, the cell is discharging at about the same rate as the other cells balancing. Since Mitsubishi replaces the entire pack as a single unit, we don't know what the actual cause is. Either way, cell failure rates are well below 1%.

Was the cell temperature way higher than the rest (either 63 or 64 in CaniOn)?

 

[Pier]

Hi PV1,

Giving all the details from CanIon reading written or verbally didn't cause any problem. They were just suspicious as why I know so much about electricity and ask what was my education ( job). The guy even told me that Mitsu is severe with misuse of the car concerning the warranty. I told him I was just reading from the OBD port, which is good for everyone in knowing what is happening to our battery, and everything was fine. He didn't even wanted to see my canion graph because they have their own test to perform anyway.

I am located in Quebec. It is cold but less than the guy in Sudbury and colder than Niagara falls. I take great care of my battery in very cold weather, I mean not driving or if necessary driving slowly. Being retired I don't need to go out every day. Last winter ( very cold) I experiment twice with recharging below -25°C. Noting happened, refusing to recharge until the afternoon when warmer. The second time I went driving and on return I could recharge, the battery being warmed up a little. But I don't usually recharge on very cold or warm days.

When the cell #85 went crazy it was in temperate weather. Full charge no more than once a week. Never keep battery full charge for hours, but recharge just before leaving. I usually keep battery around 20% - 50% charge.

It seems that the faulty cell is discharging at a faster rate than others. The difference in voltage at full charge ( I mean 80% with bad cell) is 70 mV while it is 180 mV at 12% SOC.

Cell t° was normal min 12°C and max 16°C that morning. I didn't know 63 or 64 was for cell #85.

So I am the 4th or 5th one with this problem on this forum. But how many members are we on this forum to calculate the % of cell failure ? In the future, out of the warranty period, it would be fine to replace ourself the faulty cell , because putting away an otherwise good battery pack just because one cell out of 88 is faulty would be expansive. Autonomy is dependant of only one bad cell.

I am waiting for more info. Mitsu ca decided to replace the pack. They are supposed to have ordered a new battery pack this morning, not knowing yet if it will be shipped from Toronto or Japan.

Pier

 

[nsps]

When my battery failed, there was a survey in which they asked me about my driving habits, but I got the impression it was more for the sake of tracking the cause than trying to refuse my warranty. However, maybe more batteries have been failing and they're trying to cut their losses.

 

[siai47]

Most likely the much longer recharge time you experienced when the cell was failing was the BMU trying to balance the voltage on the bad cell. As to replacing a cell in the pack after the warranty runs out, it remains to be seen if Mitsubishi will allow that. I think that an exchange refurbished pack, installed by the dealer, will be the only option. It seems a stupid way to go but manufacturers are so concerned about liability when you get into the high voltage system that I believe they will dictate this solution under the cover of "safety". On the Tesla Model S, there have been a number of failures of one of the main DC contactors which is located within the pack. Even though it would be something that could easily changed, it requires a complete pack swap. Not a problem for customers under warranty, but after it's over, a real hit. There are many other components in the I-MiEV pack that could fail other then a cell. Components like contactors, ground fault detector, current sensors, cooling fans and the like that would be cheap to replace but will be unavailable to consumers to purchase. If you look at the parts book for the I-MiEV, nothing in the pack (or related to it) is available. Unfortunately, in the case of the I-MiEV, an internal pack problem might render the car a total loss as the repair cost would exceed it's value. This is problem that owners of all EV's will face in the future until manufacturers come up with a low cost method of dealing with high voltage system related issues. Someone in the private sector will resolve this problem by retrieving parts from wrecked vehicles and repairing failed packs. High value (Tesla) and high volume (LEAF) will be the first ones the private sector will start working with. However, for a vehicle that is so limited in production numbers like the I-MiEV, there might not be enough incentive for the private sector to get involved.

 

[DonDakin]

Hi pier,

Thanks for posting this information. I think you are the first person to experience a cell failure and was able to look at the cells with canion.

I would love to understand how the car was dealing with the failing cell. I mean how was the car computing SOC and what was the cell voltage doing during charge and discharge.

Would it be possible for you to post your canion database. Its a file in the bt_can directory of your phone or tablet. I would like to look at it to see if i could understand better how the the car deals with a failing cell. I think others on the forum may be interested in the also.

We have discussed cell failure in the past but we don't really have any data only ideas about what happens when a cell fails.

Thanks again for sharing this information.

Don......

 

[priusfan]

Hi
just a detail: the name of this file is btcan.db ....

Xavier

 

[Pier]

Sorry Don, I don't know how to access bt_can directory, transfer it to my computer and post it on here. I am an old scientist but losing interest in fast changing ways of processing data. So much more to learn abut life. But we're not here to philosophize. With my day to day analysis of the data collected and a few notes, I will recall what happenned.

1- Since 2 years (one with Canion), the data were about the same. At full charge, all cells were at 4,105 volt, with sometimes one or few cells with a variation of 5 mV. Nothing special with temperature of cells, as it was also when the #85 cell went wrong. SOC was always 100% at full charge.

2- Come end of october, on a full charge, to my surprise, one bar less and soc 92%. I found cell #85 had lower voltage. Subsequent charges, soc went down to 87.5%, 84%, 79.5%, 82.5%, with bars accordingly of course. Then my trip to dealership, on a cold morning, 100km impossible on a single charge, had to stop at 40km for intermediate charge and arrived almost emty at dealership, and all that at 60 kmh. It seemed to me SOC was even less that morning , other than colder temperature. It was confirmed by dealership with 75% max soc after they say 24 hours charging. Technician told me they could force charge a single faulty cell.

3- Cells T° all that time, as I said, were as usual. Nothing special.

4- Difference of tension between all cells that were equal, an the faulty #85 cell increased with time. At full charge, when around 80% soc , difference was 70 mV. This is 4,105 vs 4,035. When discharged, the difference increased too, from 130 mV to 180 mV, and probably more arriving at the dealership.( Canion crashed 10km before arrival).

5- Time of charge. Regular full charge lasted 4 hours as usual. But the end charging increased considerably, going from 1 hr to 3 hrs. and even more they said at dealer.

6- Long end charging : I once checked the process of end charging on Canion. I didn't see anything particular as I never noticed before the process of equalizing cells on Canion. What I saw was a much lower energy flowing thru the charger. Around or below 100 Watt and 250W at the wall while it used to be no less than 460W (wall) before, at charge ending. That measure of energy flowing was of course varying a bit maybe mainly because of the circulatory pump.

7- Also at end charging, the SOC increased somewhat by a few % very slowly, along volt at the faulty #85 cell. But upon driving, this extra energy went away faster than usual in the first km. It is a kind of "fragile" or "false" energy. I don't know if there is a word for this "phantom" energy.

8- Only one faulty cell makes for max % SOC and range. The faulty cell again makes for turtle at the end, while other 87 cells could could drive range much further, thus range is minimized both end ...kind of...

If you have any other question, I will be pleased to answer if I can. Hoping this helps you.
Excuse my difficulty of writing or faults, I write with the right hand while holding a french-english dictionary in the left hand.......

Pier

 

[PV1]

I've semi-confirmed that battery temperature has a direct impact on how far each percentage will go. On a cold morning between 30-40°F, with about the same pack temperature, I lost my first bar about 1.5-2 miles before my normal point. Same environmental conditions on a different morning, but with a pack at 60°F, I made it to just about the spot where I normally lose my first bar in the summer. Always a full charge, and pre-heated both mornings. No heat use while driving. So, our reduced winter range is mainly due to a cold pack, not so much higher rolling and aero resistance, though these are still factors. This is the main cause of your even shorter range on the cold morning.

To access this file, you may already have an app along the lines of File Explorer or My files. If not, you can download an app called "File Expert". When you open it, you should see a tab on top for Folders. Tap that, then select SD Card, scroll to the folder called BT_CAN, and there will be the btcan.db file. Long hold on the file name, tap Send Files, tap Share, and from there you can email it as an attachment, or add it to Google Drive or Dropbox if you have one. I'd recommend using Google Drive (unless you already have Dropbox) since you most likely already have a Google account. From Google Drive, you can share a link to the file, which you'll post here for us to access the file. Unfortunately, you have to use a file hosting site like Dropbox or Google Drive, since this forum doesn't support attachments.

 

[Pier]

nsps-- I agree with you. Using Canion should not in any way nullify the guaranty. I talked openly about reading data with Canion. I even wrote it down in e-mail. Concerning searching for causes ...they told me my driving habits were extracted from my car's computer. Also after the first tests, the technician told me the head office ask for 20 (?) more tests. Some 25 pages of data he told me, send to Japan... surely tracking the causes...

Siai47 --- This is bad news for many owners, meaning the life expectancy of the car might be 8 years. Hopefully for most, the car will last longer, even at the cost of somewhat reduced range. We will see in a few years, things could change then. Did you open a pack and is it easy to replace a cell ? To me it seems that the most difficult part would be to disconnect the pack and put it down.

I liked my own battery because I took great care of it. I am afraid they will replace it with a not so good refurbished pack, practice allowed by the warranty ( let's say with a loss of 10% range). New pack in dealership language doesn't mean a brand new pack.

Pier

 

[DonDakin]

Hi Pier,


Thanks for the extra info.

For the canion database. You may find that if you plug in you phone or tablet to your computer with the USB charging cable. The computer will ask you if you want to open up the files on the device. If you click yes then you get an explorer window and you can brower thru the phone as if it was a USB key.

Hopefully This would allow you to copy the file to your computer and then email it. If you can Email it I will send you a personal message with my Email address and I can take care of posting it for you.

I would offer to come buy and do it but you are about 350 KM north of me in Montreal.

Speaking of that, was your car left outside in the winters ? Did it get exposed to very low temps for extended periods of time ?

I'm wondering if this could have prompted the cell failure.

What I would like to try to figure out is how the car deals with a failing cell. Your experience with this can help us understand it. I am thinking that the bad cell looses capacity and basically climbs in voltage quicker then normally and then falls quicker. Basically the charge algorithm becomes limited by the cell. When the cell voltage goes up too quickly the charger slows the charge rate and starts shunting on that bad cell. This forces a very long time to charge and perhaps it times out. In any case the other cells never fill because the charger slows down and the ah hour counter also does not climb to what it thinks is 100 percent and think just stop short of a normal charge. Perhaps the 180 mv you saw is all it takes to slow the charger and start the shunting. This would basically bring all the cells to the charge state (not capacity) of the bad one more or less. And thus the lower range and SOC. The sudden falling SOC you saw was maybe the actual deterioration of the bad cell. It looks like some sudden event caused it to decline rapidly.

On discharge I guess your bad cell might hit low voltage limit of 2.75 or the AH counter would just wind down and set SOC to zero and stop the car.

Does this make sense to everyone ?

I suppose we have another 6 years or so to think about cell failure and replacement and watch the progression of LEV-50 cell production. So much can happen in that time-frame. I remain hopeful that the 2012 model will have a decent lifespan and support.

Also, I would be very surprised if they gave you a "re-conditioned pack" at this stage I'm sure you will get a brand new one. I think everyone on the forum that had this kind of problem got a new pack. Hopefully it won't take forever to get is installed.

Thanks again

Don.....

 

[Pier]

Don,

My car is always outside in the cold of winter. I beleive cold temp doesn't stress battery. It is even good for battery longevity. I was keeping other Li-bat in the cold for last 6 years without apparent decrease capacity ( A drill to tap maple in sugarbush every spring ... and other chores, contrary to Ni-cad that lasted rarely more than one spring). IMO what is bad for Li-bat. is discharge or quick charge in cold weather. I might be wrong though. In very cold temp, I usually charge slowly (120 volt) and never accelerate fast ( one-sixth max capacity)... but who knows...

The bad cell doesn't climb quicker in voltage according to Canion and falls faster but not by much during discharge except at the end close to turtle. While running of course, volt of bad cell is just 10 mV belows others, because it is in series with others.

On discharge, the bad cell never felt down to 2.75 volt but at RR 3km, 12% soc, was at 3.610 volt while others were most at 3.790 volt( thus the 180 mV differential)

Now I will give you the reading volt of bad cell in this order: Time , SOC , volt bad cell / volt other cells.

After 4 hours of charging it was 11h39, batt Temp 13.6°C, batt volt 361v.
11h39 Soc 69% 4.025v / 4.105v.
12h04 Soc 71% 4.030v / 4.105v Strange ! drop of volt of bad cell.
13h12 Soc 76% 4.035v / 4.105v
14h53 Soc 79% 4.040v / 4.105v. End of charge.

On these 4 reading, Amp were : -2A -2A -1A 0A . Earlier at 10h00 it was -6A.

Contrary to what you say, the bad cell voltage doesn't go up too quickly. The other cells, 87 of them, fill nicely to their higher voltage of 4.105v. as it was before. The faulty cell #85 go from 4.025v. to 4.040v. in the last 3h14min for an increase of soc of 10%. The ds tech told me that after forcing the bad cell to charge separatly, SOC was only 75% ( don't know the veracity of that).

Someone knows how the equalization operates ? I would be surprise it charges the low cells separatly with somewhat like 4.5v.

I will try retrieving the canion database later. Thanks for the how to do. However, I don't think you'l get more info from these. There are so many screen to get the whole picture. It didn't just happen all in one trip or charge.

By the way 350 km is not that far with the new fast charger every 60 km :lol:

Pier

 

[Pier]

Here is a better table while charging the deteriorate battery. I thought it was lost.

Time SOC Volt #85 / volt other cells Volt pack RR bars Energy in

0h00 12.0% 3.610v / 3.790v. 335v. 03km 1b 2,500Watt
1h40 42.5% 3.925v / 4.035v. 355v. ---km 5b 2,500W
2h40 60.5% 4.000v / 4.095v. 360v. ---km 9b 2,500W
3h40 71.0% 4.030v./ 4.105v. 361v. 89km 11b 1,000W
4h30 75.5% 4.035v. / 4.105v. 361v. ---km ---b 300W
5h00 76.5% 4.035v. / 4.105v. 361v. 95km 12b 100W
5h15 77.0% 4.035v. / 4.105v. 361v. 96km 12b 100W and less
5h45 78.0% 4.040v. / 4.105v. 361v. 98km 12b 100W-
6h30 79.0% 4.040v. / 4.105v. 361v. 99km 12b 100W-
7h15 79.5% 4.040v. / 4.105v. 361v. 100km 12b 100W-
7h30 79.0% 4.040v. / 4.105v. 361v. 99km 12b End

 

[PV1]

Do you happen to have any data with the car on, but not driving? For example, you start the car and just let it sit for a while with CaniOn running. It looks to me that either that balancer is stuck on or the cell developed a short at about the same rate as a balancer. Given there's no significant rise in temperature, I doubt it's an internal short, based on my readings of the Boeing dreamliner battery incidents.

The only way to charge a single cell without taking the pack apart is to activate the balancers on every cell except the low one and do a slow charge, which is what the car is trying to do.

 

[Don]

My car is always outside in the cold of winter. I beleive cold temp doesn't stress battery. It is even good for battery longevity. I was keeping other Li-bat in the cold for last 6 years without apparent decrease capacity ( A drill to tap maple in sugarbush every spring ... and other chores, contrary to Ni-cad that lasted rarely more than one spring). IMO what is bad for Li-bat. is discharge or quick charge in cold weather. I might be wrong though. In very cold temp, I usually charge slowly (120 volt) and never accelerate fast ( one-sixth max capacity)... but who knows...

It's pretty common knowledge that below freezing temps aren't good for lithium cells - It would be a giant leap to say operation in such temps is 'good for battery longevity'

From a Battery University discussion on the subject

"Li‑ion batteries offer reasonably good charging performance at cooler temperatures and allow fast-charging in a temperature bandwidth of 5 to 45°C (41 to 113°F). Below 5°C, the charge current should be reduced, and no charging is permitted at freezing temperatures. During charge, the internal cell resistance causes a slight temperature rise that compensates for some of the cold. With all batteries, cold temperature raises the internal resistance.

Many battery users are unaware that consumer-grade lithium-ion batteries cannot be charged below 0°C (32°F). Although the pack appears to be charging normally, plating of metallic lithium can occur on the anode during a subfreezing charge. The plating is permanent and cannot be removed with cycling. Batteries with lithium plating are known to be more vulnerable to failure if exposed to vibration or other stressful conditions. Advanced chargers, such as those made by Cadex, prevent charging Li-ion below freezing.

Manufactures continue to seek ways to charge Li-ion below freezing and low-rate charging is indeed possible with most lithium-ion cells; however, it is outside the specified (and tested) limits of most manufacturers’ products. Low-temperature charging would need to be addressed on a case-by-case basis and would be manufacturer and application dependent. According to information received from university research centers, the allowable charge rate at –30°C (–22°F) is 0.02C. At this low current, a 1,000mAh Li-ion could only charge at 20mA, and this would take more than 50 hours to reach full charge.

Some Li-ion cells developed for power tool and EV applications can be charged at temperatures down to –10°C (14°F) at a reduced rate. To charge at a higher rate, Li-ion systems for automotive propulsion systems require a heating blanket. Some hybrid cars circulate warm cabin air through the batteries to raise the battery temperature, while high-performance electric cars heat and cool the battery with a liquid agent."

If you experience temps where you live which are below freezing for days at a time and you keep your car outside in the cold I think you're really asking a lot by continuing to use your car in those conditions. Did this have anything to do with your cell failure? Who knows . . . . but I certainly wouldn't think charging in such temps were 'good for battery longevity'

Don

 

[PV1]

I think this is the big difference. Storing cells at low temperatures may be good, but using them is not. I think the cells in the i-MiEV are more tolerant of low temperatures, given that the battery warmer doesn't kick in until 5° F. Still, I try to keep my cells above freezing, especially after reading the entry on Battery University. The easiest way to do this, besides parking in a warm garage, is to use the slowest charging rate that can meet your needs. The slower rate allows for the cells to generate more heat, as well as reducing the risk of plating. If equipped with cold weather or quick charge packages (or you have a 2014), you could also remove the servo and flip the damper in the left side front footwell (driver's side on US models) to direct warm air into the battery. You'll also have to direct air to the floor.

Lithium ion seems to be the happiest around 73°F during use.

Is cell 85 in a spot where it is more vulnerable to cold ambient air than others?

Here's that servo that controls air into the battery. To note the location, see the USB port in the bottom right corner of the picture.
https://www.dropbox.com/s/t5jezlryoo1j7ze/20140917_181508.jpg?dl=0

 

[DonDakin]

Hi pier,

As pv1 mentioned, it sounds like one cell may have developed a mild internal short. Indeed its the opposite of what i described. Instead of the bad cell going to a high voltage it somehow struggles with raising its voltage. Eventually the other cells start shunting and the bad gets charged at the lower shunt rate but never develops enough potential and eventually the charger times out. I wonder if this "failure mode" is indicative of lithium plating.

I have played around with 18650 cells, i found a lot of low capacity cells that operate normally in terms of voltage excursion. I also found one that had some kind of short. Only one. It would charge normally with good capacity but if left charged the voltage would seep down to zero after a few days.

I agree with Don that exposing the battery to low temps is probably not the best idea but mitsubishi must disagree or it would not sell the cars with the battery warmer set to run at -25 deg c. I guess the battery is tolerance of sub freezing charging.

I just wonder if exposing the car to very low temps can induce a cell failure like you experienced.

It sound like you babied the battery in all other aspects of using the car.

Perhaps it was just a manufacturing defect that popped up. Who knows.

Please keep us up to date with your situation and thanks again for posting on this forum.

Don....

 

[RobertC]

Thank you Pier for your valuable information. We now have documentation of a cell failure and know what the symptoms are.

It would be interesting to open up the bad cell to see what caused the failure. The bad cell could be compared to the other good cells in your pack, and your battery pack cells could be compared and tested against new cells.

Here is an update on the Mitsubishi i EV's battery failures documented on this forum:

Member, Date, Location, Avg. Low Temperature in January
tonymil, 6/2/13, Latham, NY, -9 °C, (15°F)
nt2w, 9/3/13, Potsdam, NY, -14 °C, (6°F)
dniemeyer99, 2/14/14, Sudbury, Ontario, -19°C (-2°F)
Sandange, 4/10/14, St. Sauveur, Quebec, -18°C (-1°F)
Pier, 12/2/14, St. Aubert, Quebec, -17°C (1°F)

 

[JoeS]

Here is an update on the Mitsubishi i EV's battery failures documented on this forum:

Member, Date, Location, Avg. Low Temperature in January
tonymil, 6/2/13, Latham, NY, -9 °C, (15°F)
nt2w, 9/3/13, Potsdam, NY, -14 °C, (6°F)
dniemeyer99, 2/14/14, Sudbury, Ontario, -19°C (-2°F)
Sandange, 4/10/14, St. Sauveur, Quebec, -18°C (-1°F)
Pier, 12/2/14, St. Aubert, Quebec, -17°C (1°F)

We also have the Salt Lake City battery: http://myimiev.com/forum/viewtopic.php?f=23&t=1918

At the other end of the spectrum, we seem to have only one incidence of failure:
mradtke, after the car was left plugged in sitting outside unused for two weeks in July in 47degC (117degF) max Arizona temperatures
(although there is a suspicion that it was defective before that)
Ref: http://myimiev.com/forum/viewtopic.php?f=23&t=2301

Any more? Perhaps a separate thread listing all battery failures?

To date, Mitsubishi has done well and replaced each failed pack with a new one under warranty.

 

[PV1]

Sandange, 4/10/14, St. Sauveur, Quebec, -18°C (-1°F)

Did Sandange's battery actually fail, or was it just the issue of going through chargers?