Replaced My Pack!

Mitsubishi i-MiEV Forum

Help Support Mitsubishi i-MiEV Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

jray3

The sensibly-sized alternative.
Supporting Member
Joined
Dec 6, 2011
Messages
1,945
Location
Tacoma area, WA
Well, I went and did it. Swapped a collision-salvaged 2012 pack with low miles for my original traction pack that was delivering 26.4 Ah after 105k miles. After setup, a good friend and I got it done within 4.5 hours in my driveway, without a lift or other special tools, and that included plenty of discussion time!
The salvage pack had been pulled in typical wrecking yard fashion, with cables cut a few inches outside of the box, and no fewer than three crushed wiring connectors (Heater, Air conditioning compressor, and temperature sensor), which required transplants from my original pack. I left the salvaged pack in outside storage under a tarp, and it's been at least two years since the donor car crashed. SOC at that time was unknown, but I got a good buy at $1500.
The procedure was based upon a UK owner who threaded 30 cm sections of allthread (12MM, 1.25 pitch) in place of four of the eight mounting bolts. I chose those bolts immediately fore and aft of center, and put wheel ramps under all four wheels as my tallest unmodified jacking option. I threaded a nut up the shaft till finger-tight, and then locked two nuts together at the bottom of each shaft to provide a 'bolt head' to hold and prevent the shaft from turning. Turns out that it would take forever to lower the pack by backing down the nuts, even with a ratcheting box end wrench due to only getting perhaps 20 degrees of rotation per turn due to the surrounding structure. I then realized that we could let a jack support one end of the pack, spin the nut down a couple of inches by hand, and then lower the pack and repeat at the other end. This was methodical enough to prevent any scraping or cable-snagging, but greatly sped up the process while preserving safety, lowering the pack gently onto a mover's dolly.
In short, the salvaged pack woke up on the first try, displaying 64% SOC on my MUTIII clone, with only .06 V difference between the highest and lowest cell. We immediately took a spirited 20 mile test drive down to a displayed 1 bar, and plugged into L2 for a full recharge to recalibrate the BMS. Now after 3 full L2 recharges and two DC fast charges, the MUTIII is still reporting the same capacity of 28.6 Ah, and I haven't seen higher than 50 miles RR, after a typical 38 to 41 miles RR in recent weeks on the old pack. I will try to add a full swap procedure in the coming days and share some photos. My data gathering will not be complete enough to satisfy all, but the car's utility is already increased. I hit the sleepy battery hard, with over 110 miles of use and the two DCFC sessions in it's first 24 hrs back on the road, because it was either drive hard, or drive a gasser! Sure, the optimal procedure would likely be gentle driving and L1 recharges only as the pack comes back to life, but full power delivery is no problem for it. Full power acceptance is another issue, however. The first DCFC was on a brand new Electrify America station that delivered 125a, 50 kW to a buddy's SOUL EV the day before, but my session started at 11 kW and quickly tapered to 9 and then a measly 6 kW. The second DCFC, 24 miles and 30 minutes later, was on a BLINK DCFC, which started at 100 amps, 35 kW, but then tapered down similarly.

I removed the lids on both packs for inspection, and found that though both contained LEV50 cells, not LEV50N, my originals were yellow and the replacements are aqua blue in color. Since the temperature sender is enclosed in the rear portion of the pack lid, I simply transplanted the lid rather than extracting the sensor. Removing the damaged HVDC cables from the busbars that lead to the main contactors was challenging, as they pass through unthreaded busbars into a square nut that is caged in brittle plastic. Using steady pressure on a hand wrench cracked the plastic and bent wiring terminals, even with great care to brace the terminals with flat blade screwdrivers, but an electric impact wrench was quick enough to break the others loose without breaking the plastic, no bracing needed.

If you need replacement cells that only have 26 Ah capacity remaining, send a PM. My current plan for the old pack is to put it on a trailer and engage the CHAdeMO contactors with an external 12V signal, enabling it to be used as a battery range extender, either in parallel with MR BEAN, or behind one of my homebuilt EVs, using a Manzanita Micro PFC 50 as a 12 kW DC-DC converter, shuttling charge from the 350+V pack down to a lower voltage pack in my conversions in a manner more civilized than jumper cables (my conversions run 120, 144, and 240 VDC). It also has DIY Powerwall potential, in parallel with my solar arrays.

Many thanks to our own JoeS and prolific online EV commenter BrianH for their invaluable assistance on this project!

Stay tuned,
Jay
 
Great story :) congrats for change. Did u needed use any programming to Vin code changed battery to your car? Or did u changed BMS brains also from old to new battery? Just wondering if do change new cells to battery, will bms Brains adapt the new AH size of battery? I know that BMS wont tell over 50.0Ah or 100.0% status in Evbatmonitor. I got garanty battery change for 1 cell damaged battery. Maybe just BMS blansing resistor stayed on in it.. cell dropped 5% every day.. but charged and were ok when drived.. so cell wasnot bad.

-bj
 
Thank you Bjorn,
There was no VIN lock, so no brain transplants required.
I got a definite sign of trouble this morning. I had noticed that the first bar didn't last very long in driving, and last night it only charged up to 15 bars, 48 miles RR. :cry: The datastream says only 88% state of charge. So I pulled the trouble codes, which I hadn't done before on this battery, as there were no warning lights. I got P1AC6- Each Cell Volt diff (Hi side) and P1A4D, IGCT line voltage. The voltage difference between hi and low is 0.05V (4.10V on cell 19 and 4.05V on cell 38). Cell 38 has been the lowest every time I've scanned. We'll see if #38 gets better or worse with some 8A L1 TLC over the next few days.
 
jray3 said:
The procedure was based upon a UK owner who threaded 30 cm sections of allthread (12MM, 1.25 pitch) in place of four of the eight mounting bolts. I chose those bolts immediately fore and aft of center, and put wheel ramps under all four wheels as my tallest unmodified jacking option. I threaded a nut up the shaft till finger-tight, and then locked two nuts together at the bottom of each shaft to provide a 'bolt head' to hold and prevent the shaft from turning. Turns out that it would take forever to lower the pack by backing down the nuts, even with a ratcheting box end wrench due to only getting perhaps 20 degrees of rotation per turn due to the surrounding structure. I then realized that we could let a jack support one end of the pack, spin the nut down a couple of inches by hand, and then lower the pack and repeat at the other end. This was methodical enough to prevent any scraping or cable-snagging, but greatly sped up the process while preserving safety, lowering the pack gently onto a mover's dolly.
In the next couple of months I'm going to do much the same thing to swap 4 cells. My plan is to use the M12 x 1.25mm threaded rod method (you say 30cm was enough length ?) and like you use a jack to support it while the nuts are spun quickly then lower the jack a bit more. Should work quite well I think.

I was originally going to use a dolly (made of 4 caster wheels and some 25mm plywood) to support it all the way down - in other words put the jack under the middle of it and jack the dolly right up in the air under the pack to support it from the middle and lower it down from there, however your description of simply supporting one end of the other alternatively with the jack sounds like it would be easier still and just require the dolly to wait at the ground.
In short, the salvaged pack woke up on the first try, displaying 64% SOC on my MUTIII clone, with only .06 V difference between the highest and lowest cell. We immediately took a spirited 20 mile test drive down to a displayed 1 bar, and plugged into L2 for a full recharge to recalibrate the BMS. Now after 3 full L2 recharges and two DC fast charges, the MUTIII is still reporting the same capacity of 28.6 Ah, and I haven't seen higher than 50 miles RR, after a typical 38 to 41 miles RR in recent weeks on the old pack.
If you see another thread I posted in not long ago you'll see that the BMU won't make large changes to the Ah capacity by itself in any reasonable amount of time. I reset mine to the factory default (which is 45.8Ah) using the "battery replacement" service operation in Diagbox and drove the car for several days and several hundred miles and it did not budge even though I ran the car right down to the point where it nearly shut down on one occasion.

If you swap the battery pack you must either do the "battery replacement" diagnostic process (not sure what it is called in MUT-III) or do the "battery calibration" diagnostic process. The battery replacement procedure simply resets the capacity to a factory new 45.8Ah and does not measure the true capacity of the battery - in the diagnostic tool it says this should only be used if the battery pack is verified to be less than 6 months old from date of manufacture - clearly not true in your case! :lol:

So instead you need to run the battery calibration process - whatever it is called in MUT-III. Set aside a whole day to do it, it takes a long time as you have to both discharge to about 20% then fully charge to 100% on Level 1/2, but it will accurately measure the performance of your replacement battery and update the usable Ah capacity figure accordingly.

After this you will see increased range, assuming the replacement battery actually does have better capacity.
Full power acceptance is another issue, however. The first DCFC was on a brand new Electrify America station that delivered 125a, 50 kW to a buddy's SOUL EV the day before, but my session started at 11 kW and quickly tapered to 9 and then a measly 6 kW. The second DCFC, 24 miles and 30 minutes later, was on a BLINK DCFC, which started at 100 amps, 35 kW, but then tapered down similarly.
Early and excessive throttling of DC rapid charging speeds suggests one or more high resistance cells. My car has 3 cells which have developed higher than normal internal resistance over the time I've had the car and the symptoms of this are rapid charging speeds starting to throttle much sooner and at low SoC and also reduced regenerative braking strength. This high resistance is one reason why I am going to replace these cells. (They also have low Ah capacity)

It's relatively easy to diagnose high internal resistance cells by rule of thumb using Canion and DC rapid charger. Monitor the cell voltage page on Canion before starting rapid charging, if soon after you start rapid charging certain cells go much higher in voltage than others (say more than about 15-20mV) and they were equal or lower voltage than others before charging they have high resistance.

When the highest voltage cell hits 4.105 volts the rapid charge rate is throttled back to keep it exceeding this voltage. If even one cells has high internal resistance it can dramatically slow down rapid charging as the BMU attempts to keep the cell going over voltage.

You can see this in the following post I made where the cells highlighted in red have gone to a higher voltage than the rest during rapid charging:

http://myimiev.com/forum/viewtopic.php?f=23&t=4014&start=50#p39013

Before jumping to conclusions I would perform the battery calibration though and make sure the BMU has a correct figure for the Ah of the battery.
I removed the lids on both packs for inspection, and found that though both contained LEV50 cells, not LEV50N, my originals were yellow and the replacements are aqua blue in color.
The cells I just bought are aqua blue, I don't know what cells are in my car though. Purple is another colour of cell for these as well.
Since the temperature sender is enclosed in the rear portion of the pack lid, I simply transplanted the lid rather than extracting the sensor.
What temperature sender is this ? There are 66 temperature sensors on the CMU PCB's to monitor cell temperatures but I am not aware of any other temperature sensors inside the battery enclosure ? Are you sure that cable wasn't for the extractor fan ?

jray3 said:
Thank you Bjorn,
There was no VIN lock, so no brain transplants required.
Correct, there is no VIN coding of the CMU boards in the pack.
I got a definite sign of trouble this morning. I had noticed that the first bar didn't last very long in driving, and last night it only charged up to 15 bars, 48 miles RR. :cry: The datastream says only 88% state of charge. So I pulled the trouble codes, which I hadn't done before on this battery, as there were no warning lights. I got P1AC6- Each Cell Volt diff (Hi side) and P1A4D, IGCT line voltage. The voltage difference between hi and low is 0.05V (4.10V on cell 19 and 4.05V on cell 38). Cell 38 has been the lowest every time I've scanned. We'll see if #38 gets better or worse with some 8A L1 TLC over the next few days.

P1A4D according to this means that the 12v supply voltage to the EV-ECU is below 8 volts or above 16 volts with the car in the READY mode.

Check the health and charge of your 12v battery! Has it been replaced in the time you've had the car ?

P1AC6 according to this is due to a difference in voltage between the cells of more than 0.05v. However a difference of this magnitude is perfectly normal, and shouldn't result in a fault.

So what I think this fault really means is the consistency check it does where it measures the entire string of 8 or 4 cells in a module and compares that to the sum of the individual measurements has failed.

In short, you might have a faulty CMU board. So don't throw away that old pack yet! You might need to salvage a CMU off it and swap it over.

If you look at the realtime voltage data for each CMU in the MUT you should see individual cell voltages for each cell in the module and one overall module voltage which is 31 volts for an 8 cell module - this overall voltage is not simply the sum of the individual measurements but a separately taken measurement.
 
Thanks Simon that is a lot to review but of course I did a calibration charge with my very first charge and twice since. The pack is still reading 28.6 Ah, but DCFC Performance is rapidly improving I am currently at an EVGO station (BTC Power 50 kW dual cable unit) and it’s just like old times, charging at 109 amps for 8 minutes, and still pulling 37 amps at 72% SOC 11 minutes in.
Voltage differential is running between 0.04 and 0.06V, with cell#38 always the low one.
This session charged from 24% to 81% inside of 18 minutes, better than the previous sessions that shut down early.

Yes, that broken wiring connector that remained with the lid must be for the fan.
 
jray3 said:
Thanks Simon that is a lot to review but of course I did a calibration charge with my very first charge and twice since.
You misunderstand. I'm not saying to do a "calibration charge", by which I presume you mean running the battery down until the bars are flashing and charging it back up again. That will not recalibrate the battery for such a large change and is a bit of a misnomer to be honest.

There is an actual battery calibration process which must be initiated with the dealer diagnostic tool where you will be asked to follow a number of steps such as turning on the heater to run the battery down to 3.775 volts per cell, then plug it in to charge etc - all whilst the diagnostic tool remains connected. (So plug your laptop into power!)

This takes nearly all day to complete, but it will immediately update the Ah figure at the end of the calibration process and you will then have an accurately calibrated battery. The fact that it's still reporting the old figure of 28.6Ah means that you have not done this calibration process.

I can't give the exact steps to perform on the MUT-III as I use Diagbox and the menus differ greatly but rest assured there will be a battery calibration procedure in there somewhere. It will be in the repair/replacement options under the BMU.
 
Thx again Simon. I'm happy to let the car take its time to recalibrate, and see if any problems persist or pop up anew as we cycle the salvaged pack in daily use, but am tempted to let my dealer take a peek. We've built a bit of a relationship and traded some favors. The reported capacity is rising about one Ah per day, so at that rate- I'll have a good as new battery in two more weeks!
Reviewing my photo logs, the reported capacity for my first drive on Tuesday night was 27.6 Ah, rising to 28.6 after 12 hrs, one L2 and one DCFC, and then 29.6 after another 24 hrs, one DCFC and one L2 full recharge. Who wants to bet it'll be 30.6 Ah when the wife gets home today after another full cycle on the battery?

Cell #38 was still the laggard, running about 0.04V lower than the highest cells at rest and as much as 0.07 low under load. Cell #38 corresponds with CMU ID05 Cell F when I scroll through the datastream on my MUTIII clone, the iCarsoft i909. This unit appears to be just a scanner with no reprogramming ability beyond resetting the OBD codes. https://www.icarsoft.com/Product/s-338-iCarsoft_Multi_system_Scanner_i909.html

The old pack is certainly not going to be thrown away, it'll be reused to the greatest extent possible.
 
hobdrive can align cells, but for a very long time, approximately 0.005v per charge. but you need to keep the program open and press the calibration if it turns off.
straightens my 8 cells for 2 months, and for a long time he will do it.
with full 4,100v in 80 cells, 8 replaced cells and today they have 4,065v
 
Just an update as the salvaged pack continues to improve, albeit at a slower rate. We’re now reporting 30.5 Ah capacity, and the laggard cell #38 is only 0.03V low in the morning now. DCFC times are back into sub-30 minutes, with a good duration at 50 kW. I’m now doubting that this pack will recover to more than 50 practical miles, which would continue to serve our needs, but piques my enthusiasm for a long-range rebuild of the original pack. Anybody know more than revealed in this Aussie thread, which was planning a 100 Ah rebuild? https://forums.aeva.asn.au/viewtopic.php?f=49&t=5515&start=25
 
Do you have a MUT diagnostic tool ? If so, I'm not sure I understand why you're so reluctant to perform the battery calibration procedure.

It's the only way that you're going to get the BMU to make an accurate measurement of the Ah capacity with a sudden large change in cell capacity such as after an entire pack replacement. Without this it will not let you use the true usable capacity of the pack which makes replacing the pack a little pointless IMHO...
 
As discussed earlier, I have only the MUT3 scanner clone by iCarSoft. It doesn’t send commands beyond resetting OBD codes, as far as I can tell. Besides, if the factory says to only reset to a default capacity if the battery was manufactured within the past 6 months, would it not be unwise to do so with a 7 year old salvaged battery? A gradual learning process by the BMU seems wiser, especially with one cell showing weakness...
 
With capacity remaining at 30.5 amp hours for over a week with no improvement and the range estimates standing at 56 over several days of sedate wifely driving, I was presented with an opportunity for some more serious testing today. The first leg of today’s 75 mile outbound trip could have either been 24 or 50 miles so let’s go for it! We made it the 50.0 miles To the first DC Fast charging stop with the last mile on turtle, arrival state of charge was 7.5% (or 13% according to the Blink DCFC) The good news is that the lowest cell was only 0.03 V down and the second lowest in recent history; number 20, shared the same voltage reading. DCFC commenced at a healthy 119 A and continued down to no less than 30 A when it disconnected at 14 bars and 86.5% SOC (I had asked Blink to charge to 90% though the car usually stops at 80%).
After that hard highway drive, the 86.5% range estimate is 43 miles. Looking pretty good.
 
Measured pack capacity continues to climb slowly with use. My MUT3 clone now recognizes 33.3 Ah of capacity on the 2012 salvaged pack. I’m doing a lot of driving, as much as 170 miles per day and 3 to 4 DCFC sessions on weekdays and plugged in on L2 the rest of the time...
Cell #38 came back into line weeks ago and the hi/lo cell numbers now seem pretty random, with very little differences.
 
So what you're confirming here is what we anticipated - it takes a LOT of driving for the BMS to gradually increase the estimated Ah capacity of the cells if you do a battery pack swap and don't tell it about it by doing the normal calibration routine...
 
Indeed, Mandrake- Point Proven as Expected, but not something that bothered me enough to purchase an expensive tool that I might need only once.
My car is currently at the dealer getting a warranty replacement of the onboard charger.
I'll see about having them do the recalibration.
 
Hey Jray3,

I have been away from the forum for a long time. Still own 2 imievs that are still working nicely.

I was just wondering if your charger failed or if the dealer is just replacing chargers ? I got a letter in the mail recently saying the charger was under warrenty for 10 years now with unlimited KM's

Interesting read on the battery replacement and re-calibration.

Don.....
 
Thx Don,
My OBC failed, so a dealer is replacing it with the updated design. They won't replace a functional OBC, and want to see the correct diagnostic code still being generated. A statement that I cleared the codes while trying to diagnose and restart the charger was sufficient to get the work approved after I delivered the car with bad charger. They wouldn't consider the replacement unless a failed charger was currently installed in the car, so I obliged. :roll:
 
The replacement battery continues to improve, or more likely, the slow BMS recalibration continues. For the first time in several years, I made an airport run without charging. That’s a 65 mile round trip, on a battery that was cold-soaked near freezing at both ends, on snow tires! I arrived home with 8 miles RR and two bars, so had to hook up the MUT3. Result was 36 Ah capacity, with those two flashing bars representing 19.5% SOC.
:eek:
 
jray3 said:
The replacement battery continues to improve, or more likely, the slow BMS recalibration continues. For the first time in several years, I made an airport run without charging. That’s a 65 mile round trip, on a battery that was cold-soaked near freezing at both ends, on snow tires! I arrived home with 8 miles RR and two bars, so had to hook up the MUT3. Result was 36 Ah capacity, with those two flashing bars representing 19.5% SOC.
:eek:
What?! My pack is over 45 Ah and going that far would require rather attentive driving or a heat wave. Koorz is a 50-mile car and it takes all it has to make the (now) 45-mile round trip to work. I had to bail today and borrow a LEAF to get home since my OBC decided to quit working.
 
Neither of your two sets of figures make sense to me, once seems extremely optimisitc one very pessimistic ?

When mine was at 39.9Ah my summer range was 63 miles and winter about 43 miles. Now it's at 32.8Ah my summer range is 54 miles and winter range 35 miles...
 
Back
Top