Aftermarket iMiEV battery upgrade project

[nuggetgalore]

Well done Andreas! Yes it's always a bigger job than first imagined. Glad you got everything to fit in the end. The two 4-cell modules weren't too difficult to manage?

I had to replace both side (resp top and bottom) walls with thinner ones. The thickness of these modules is critical, the cover sits flush on the top of the retainer bracket with just about zero clearance.This also forces the cooling air to flow past the vertical sides of the modules.
The bus bars (for those interested : laser cut C11000 high conductivity copper) were quite tricky to design and bend to fit the gap between modules 1+2 and 12 resp. 4+5 and 6. Then I had to modify the original busbar from module 12 to the main contactor in such a way that in case the new cells are not not good enough I am able to refit the originals. Same for the bar connecting module 6 with the link to the 280 Amp fuse. By the way, the link from the main fuse to the service plug is the only bit I did not have to replace or modify (very much appreciated,ha ha ha).
My training as a pattern-maker did help though,lol.

[nuggetgalore]

By Sunday afternoon I had the battery back in the car and it is charging.
Thanks to coulomb's explanation the fault was easy to find.
It appears that the track on the intermediate board was disconnected from the soldering pad.
I either missed that checking the individual continuity from flywire terminal to 3mm nut or it broke later and I misread the voltage from module negative to cell 7 positive (about 27.xy V but 25.xy V when the connection to cell 7 is open circuit).
Either way,quite embarrassing error by the quality department(that's me!)
I am pleased to report that I was able to clear the two DTCs with the i909.

The GoM is as expected not knowing what is happening, battery is I think fully charged now but GoM showing 14 bars. A bit disappointing that the max cell volt at 4.11 is now 0.04 higher than the min at 4.07 after charging 3.65 kWh off the wall. I did a quick check on some balancer drive status while it was charging, they were all on, for what ever that means (did not have enough time to track the high/low volt cell(s) with the balancer drive, next time maybe).
I wonder if that high volt cell was shutting the charging down? But I did find out during pre-charging the modules (obviously without any balancing gadget) that the cells get out of balance quite soon. After I was happy with the overall module voltages I paralleled high volt cells with low volt cells of an other module until all cells where inside 0.02V, that was then confirmed by i909 and EvBatMon.

[acmotor]

Great work there !
I am among the many who are hanging on your every update !
Please keep up the posts.

[nuggetgalore]

Please keep up the posts.

Yes I will keep posting. But one important caveat right here. Even if these cells do satisfactorily perform for a reasonable time,it is not a practical proposition, considering the enormous effort Chris has put in plus the many hours I spent cajoling the cells to a reasonable state of balance and then installing the modules.
Anyway, the first preliminary results are in. The not quite full first charge (to 14 bars but next morning was 15) allowed me to drive 58 km until the last bar disappeared (zero on GoM), 2km later the tortoise lit up but with no noticeable limit of power when I kept driving the last 2 km to home.
The pack voltage as soon as I was home(at rest) was 324.9 V max cell 3.72 V min cell 3.68 V.
I captured an EvBatMon screen shot at 11% SoC drawing 41 amps:pack 317.5V; max cell 3.63 ;
min 3.48 V.
Comparing the voltages with those I posted on page 13, there should be a fair bit of headroom for the machine to learn the new capacity. It better does, as otherwise it would be no better than the old original cells. Several full charge and discharge cycles hopefully show an improvement.
I also might have to get hold of a MUT3 or pay Mitsubishi for a battery re-calibration. Sadly the i909 is unable to do this afaik.
tbc

[acmotor]

My thinking is that it is not the practicality of those particular cells but rather the breaking of the knowledge ground around the imiev brain accepting an aftermarket upsize. For this work you are to be applauded.

[nuggetgalore]

Not a lot of success so far. I made a silly mistake today.
I had three bars left with GoM showing 18 km after I got home yesterday driving about 44 km starting with the battery supposedly fully charged at 259 V.
Tortoise lit driving another about 10 km.
As the cell voltages were way above what they used to be with the original cells at tortoise, I intended to drive a bit further so I could pump a bit more juice in when charging.The car showed no slowing down symptoms at all,accelerating and regenerating normally,I had EvBatMon displaying currents and voltages. But about 2 km later than what GoM predicted it suddenly dropped power and I could just limp into a safe spot at the side of the road. So the "full" charge gave me about 64 km before it died. Had it towed (RACV gold membership, thanks !) and an hour later it was getting charged, no problem with that.
Max cell voltage at start of charge was 3.71 V min 3.68 V,pack 324.3V.
This charge took just under 10kWh off the wall(about 9kWh into the battery) until the BMS said at 4.11 V max cell volt and 4.07 min cell volt (pack 459.90 V) battery is full.
I think the cells (amongst other things) are not a matched set,the BMS is struggling to balance the cells (the original cells always topped up at 4.1 within 0.02 or better and pack volt 362). With the exceptions of just a few(at the lowest volt cells), all cell balance drives are on all the time, battling to keep the cell voltages in line.
Interesting that at the end of charge, the fuel gauge showed 12 bars and GoM 91 km.I took it for a short (1km) drive,and fuel gauge went to 15 bar and GoM to 96 km. I will be surprised if I can drive more than 60 km with that charge.
The real bad news is that when I took the car for short drive, I noticed that regeneration is not working anymore.
I checked for any fault codes but none are set, that is even more worrying.I can only hope that somebody has an explanation/solution for this problem.

[dracekvo]

You not try run battery calibration proces with diagnostics software?

[zzcoopej]

The real bad news is that when I took the car for short drive, I noticed that regeneration is not working anymore.

You've probably considered this, but anyway - I have found with our iMiEV, however much more noticeable with the PHEV, that when the battery was near new and is fully charged, it can't accept regen, or only very weak regen, until the battery has some capacity.

[nuggetgalore]

The real bad news is that when I took the car for short drive, I noticed that regeneration is not working anymore.

You've probably considered this, but anyway - I have found with our iMiEV, however much more noticeable with the PHEV, that when the battery was near new and is fully charged, it can't accept regen, or only very weak regen, until the battery has some capacity.

To be honest, I never considered that, thanks for mentioning. I cannot recall that I ever noticed this before though,I must have driven the car many times with a full battery. I will find out soon enough.

[whimpurinter]

Yes, when the imiev is fully charged, it does not regen until the charge has dropped a little.

[coulomb]

with the battery supposedly fully charged at 259 V.

I hope you mean 359 V.

Max cell voltage at start of charge was 3.71 V min 3.68 V,pack 324.3V.

So you still had probably around 55 km of range left. Turtle for these cells should be around 3.5 VPC.

...until the BMS said at 4.11 V max cell volt and 4.07 min cell volt (pack 459.90 V) battery is full.

I hope you mean 359 V.

I think the cells (amongst other things) are not a matched set,the BMS is struggling to balance the cells (the original cells always topped up at 4.1 within 0.02 or better and pack volt 362).

Graeme reported earlier that with the 94 Ah versions of these cells, he had a lot of trouble balancing them, so he had to take all the cells out of the first conversion and parallel them for a week. You might be able to get away with some manual balancing (sets of headlight bulbs with clip leads (I have a box of these) and some dual chargers. Though I'd want to break the pack into at least two pieces before using chargers on them; 360 V is a bit much to trust to the insulation of a cheap charger / power supply.

With the exceptions of just a few(at the lowest volt cells), all cell balance drives are on all the time, battling to keep the cell voltages in line.

I wish I knew more about those cell balancers, and what it means when they are "on".

Interesting that at the end of charge, the fuel gauge showed 12 bars and GoM 91 km.

12 bars at the end of a charge? That seems strange. Did you perhaps mean 16?

I took it for a short (1km) drive,and fuel gauge went to 15 bar and GoM to 96 km.

The RR meter increasing as you drive it a familiar situation.

The real bad news is that when I took the car for short drive, I noticed that regeneration is not working anymore.

I have no idea why that is; I never noticed that. It is probably due to the difference in voltage versus SOC at high SOC.
Edit: it may also be that the 94 Ah cells can take more regen than the 60 Ah cells, since regen current is smaller relative to 1C.

[nuggetgalore]

with the battery supposedly fully charged at 259 V.

I hope you mean 359 V.

Sorry yes a typo

Max cell voltage at start of charge was 3.71 V min 3.68 V,pack 324.3V.
So you still had probably around 55 km of range left. Turtle for these cells should be around 3.5 VPC.

...until the BMS said at 4.11 V max cell volt and 4.07 min cell volt (pack 459.90 V) battery is full.
I hope you mean 359 V.

I make this type of typo typical....
I think the cells (amongst other things) are not a matched set,the BMS is struggling to balance the cells (the original cells always topped up at 4.1 within 0.02 or better and pack volt 362).

Graeme reported earlier that with the 94 Ah versions of these cells, he had a lot of trouble balancing them, so he had to take all the cells out of the first conversion and parallel them for a week. You might be able to get away with some manual balancing (sets of headlight bulbs with clip leads (I have a box of these) and some dual chargers. Though I'd want to break the pack into at least two pieces before using chargers on them; 360 V is a bit much to trust to the insulation of a cheap charger / power supply.

I had all the cells near perfectly balanced when I installed them into the pack. with the exception of cells 33 and 34 (different reason), both i909 and EvBatMon confirmed this.
As bladecar and zzcoopej posted, the lack of regen was because the high volt cell was not able to take any more charge,after a few more km it did regen again. I am now fairly sure that the relatively big inconsistency in the cell capacity make them nearly unusable. At relatively high pack volt, cell #6 always becomes the high volt cell when I take the foot of the accelerator, and another cell takes that spot as soon as I ever so lightly accelerate again.

With the exceptions of just a few(at the lowest volt cells), all cell balance drives are on all the time, battling to keep the cell voltages in line.
I wish I knew more about those cell balancers, and what it means when they are "on".

I have now been able to cross reference the balance drive [OFF] status with the low(est) volt cells.
So my deduction is that if the drive is on,it bleeds a small amount from the cell,allowing the lower cells to catch up. But it appears the out of balance (resp the capacity mismatch) of these 60Ah cells is too great for the BMS to handle. The original cells even after nearly ten years in service always were able to get to within very close tolerances during charging and of course at the end of charge. On a related topic, the target voltage is always the voltage of the lowest cell, I assume (with caution!) that as long as there is head room for the high cell, the charging continues until the target reaches the top.But if the high cell runs out of space, stiff cheese for the poor straggler cells,it then follows that the pack volt does not reach full voltage resp all the possible Ahs.

Interesting that at the end of charge, the fuel gauge showed 12 bars and GoM 91 km.
12 bars at the end of a charge? That seems strange. Did you perhaps mean 16?

not a typo this time!

I took it for a short (1km) drive,and fuel gauge went to 15 bar and GoM to 96 km.
The RR meter increasing as you drive it a familiar situation.

Never happened before,quite the opposite, just 10 meters driving used to reduce RR by a km.

The real bad news is that when I took the car for short drive, I noticed that regeneration is not working anymore.
I have no idea why that is; I never noticed that. It is probably due to the difference in voltage versus SOC at high SOC.
Edit: it may also be that the 94 Ah cells can take more regen than the 60 Ah cells, since regen current is smaller relative to 1C.

I have a good idea why I never noticed the lack or reduction of regen with the original cells.They are much fatter and evenly balanced, but also just a short moment after end of charging sagged a little bit, enough to take a regen charge straight away.Just talked to a '12 iMiEV owner who had his car from new,he never noticed that phenomenon either (not looking for it,of course).

So this morning I drove off with the above mentioned 15 bars on the fuel gauge ; GoM 94 km; pack volt 357.80; high cell 4.08; low volt cell 4.04 .
Result after driving (no freeway but up and down a small slope) 42.4 km:
Pack v now 331.1 V; high cell 3.79 V;;low cell 3.75V and GoM 25km. Back on charger,wonder how many Ah/kWh it will accept from this position.

[ Edited Coulomb: fixed quoting. ]

[jonescg]

I wonder if the balance will improve with more short (10 to 20 km) drives and recharges?

[dracekvo]

I mean there is problem with battery real state and SOC state in BMS.
If car menas that SOC is 100%. Car wIll stop charge = ballancing not completr and battery is not really on 100%.

I ask again. Can you try run battery calibration process via diagnostics software?

[coulomb]

So my deduction is that if the drive is on,it bleeds a small amount from the cell,allowing the lower cells to catch up.

Good, that's the way that makes most sense.

Perhaps you need to discharge the battery to a point where roughly 50% of the cells are bypassing. All bypassing does no good (just wastes stored energy), and no or few bypassing does little good. Then just leaving the battery at about that state of charge (it might be necessary to top up or bleed down now and then) might allow better balancing while you do other things, even while you sleep. Of course, it would take at least weeks to balance it, but you might as well leave it at the ≈50% balancing point as much as possible.

Or perhaps it's mainly the two cells you singled out for mentioning. Perhaps those could be manually balanced. But of course, automatic balancing is waaay less effort, even if it's 200x slower.

[nuggetgalore]

I mean there is problem with battery real state and SOC state in BMS.
If car menas that SOC is 100%. Car wIll stop charge = ballancing not completr and battery is not really on 100%.

I ask again. Can you try run battery calibration process via diagnostics software?

No, the i909 cannot do that. As I posted above, I either need to borrow a MUT3 or pay Mitsubishi to do it. I talked to the local Mitsi dealer, but am not confident that they would be knowledgeable enough to do it, considering it is NOT a standard LEV50 pack. There are just not enough iMiEVs around for the mechanics to gain experience.
Anyway, since the cells were perfectly balanced at install,and got out of balance with the first charge, I doubt that smoothing and calibration would fix it. I can of course be wrong, if the charging and associated learning improves with a number of charges, then I will eat humble pie

[dracekvo]

Yes I understand.
I have Mitsu clone Peugeot ION. There working diagnostics PSA Lexia. I buy in aliexpress clone PSA diagnostic tool. Working well.
I mean this is not good idea, try replace battery without original diagnostic software.

Is not possible buy in china clone mitsubishi diagnostics?

[nuggetgalore]

Yes I understand.
I have Mitsu clone Peugeot ION. There working diagnostics PSA Lexia. I buy in aliexpress clone PSA diagnostic tool. Working well.
I mean this is not good idea, try replace battery without original diagnostic software.

Is not possible buy in china clone mitsubishi diagnostics?

You are correct in saying that without the proper tools it is not a good idea to replace cells.
I looked at Chinese MUT III clones, but am too stingy to buy something others people have had negative experience with and I possibly may not know how to run it with only Chinglish instructions. I am still hoping I may be able to hire or borrow a unit and its owner who knows how to use it.

[nuggetgalore]

I wonder if the balance will improve with more short (10 to 20 km) drives and recharges?

Yes that is one other method I will try.
Thinking back, I first paralled all the cells and got them to within 20mv or better. Then I series connected the 8 cells in 11 modules and paralleled and charged them. I soon noticed that while the 8 cell units them self were all charging to the exact same total voltage, there were always some individual cells high and others low inside the modules(200 or up to 400 mv). At least three times I painstakingly balanced all the cells, every time by just charging at small steps (varied amperage from as low as 5A up to 20 A max from just under 30 VDC to about 31.5 VDC) the discrepancy was back. Interestingly though, the voltage of cells at rest was very stable (over weeks or months).

A question I should have asked a long time ago: is charging at a very low rate more prone to exacerbate the imbalance problem ? My iMiEV charges at about 12.5 A off the wall 230 VAC, EvBatMon says 8 A but the i909 reads it as 7A @ what ever DC volt the pack is.

[coulomb]

is charging at a very low rate more prone to exacerbate the imbalance problem ?

I can't see why that would be the case.

[nuggetgalore]

is charging at a very low rate more prone to exacerbate the imbalance problem ?

I can't see why that would be the case.

Neither can I, but I have been wrong too often to dismiss the question.
Ok, so I charged the car yesterday afternoon (after driving 42.4 km) and pumped 7.47kWh off the wall into the tank. Like the previous charges, the charger kept going on for a very long time at 1 A (i909 reading, 260 W on the wall plug), but making very little change on the SoC % resp the cell voltages. Pack voltage appears to not go above 359 V with top cell @ 4.11 V. This time fuel gauge 15 bar,RR 86km. SoC (control) 90.50% ,SoC (display) 90%. That looks a lot better.
I left it untouched until this morning, amazingly the pack and the max and min cell volts; fuel gauge and RR have not sagged or changed one iota.
I took it for a short spin and this time it behaved a lot more normal,RR reduced to 85 after 200 meters and top cell volt dropped to 4.10 after .5 km ; to 4.08 V after 1.5km with RR 83; SoC 88% resp 87.5%.
After brekky I take it for a 10km drive (50% freeway) and charge it again as per Chris's suggestion.

[jonescg]

Hopefully these trends continue!

[nuggetgalore]

Hopefully these trends continue!

That above mentioned 10 km trip ended up being 20.7 km (~5km at 100km/h freeway)
RR dropped to 53km;SoC (control) 65.50%;SoC(display) 65.50% (equal!);3.93 V and 3.90 V with pack 344.5 V .
Charge to full again. 3.64 kWh off the wall. SoC (control) 91.50%; SoC (display) 102.5% !
RR 86; fuel gauge 16 bars.
No change of end of charge voltages (4.11;4.07;359.70)

Next trip 28.5km (~ 10 km @100km/h): RR 50;fuel gauge half (8bars); SoC(c)44%; SoC(d) 55% (strange, these values all over the place); 3.92; 3.89 (0.03 diff, best I've seen so far) and 343.1 V
Full charge next morning:4.63 kWh; SoC (control) 87.5%; SoC (display) 85.5%
RR 82; fuel gauge 14 bars.
No change of end of charge voltages (4.11;4.07;359.40).

Now a 10.7km trip:(50km zone) RR 61km;fuel gauge 12 bars; 4.00;3.98;350.9 V
2.22 kWh to fill the tank ; 4.11;4.07 ;359.8 V.
A further note: All above charge cycles were until the charger switched off by itself.
Charging is at full amps until about 1kWh before full when it starts to taper off for about one hour to fill the last kWh (when the final cell and pack voltage are reached). Then (I assume) it keeps trying to balance resp fill the lagging cells for many hours (up to five) , basically wasting energy (more than half a kWh in some instances).

[nuggetgalore]

Perhaps you need to discharge the battery to a point where roughly 50% of the cells are bypassing. All bypassing does no good (just wastes stored energy), and no or few bypassing does little good.

That appears to be the big problem.
At any stage of battery fill, very few balance drives are off, or cycle on /off, reducing to one (in the section from cell CMU01 A to cells CMU09 H accessible with the i909) at about when the charge tapering down starts.

Then just leaving the battery at about that state of charge (it might be necessary to top up or bleed down now and then) might allow better balancing while you do other things, even while you sleep. Of course, it would take at least weeks to balance it, but you might as well leave it at the ≈50% balancing point as much as possible.

I don't quite get this. Do you mean just leaving the battery at rest without being charged? All the balance drives are OFF as soon as the charging stops, irrespective whether the EVSE cable is plugged in or not. As far as if understand, no balancing happens except if charging.

Or perhaps it's mainly the two cells you singled out for mentioning. Perhaps those could be manually balanced. But of course, automatic balancing is waaay less effort, even if it's 200x slower.

During precharging I balanced wayward cells multiple times.Unfortunately I failed to record which cell did what.But I know the unbalance happened very quickly. Now I know a bit more,high and low cells are displayed and the ones in between I can read with the i909.
How can individual cells be manually balanced without taking the pack out?

[coulomb]

I don't quite get this. Do you mean just leaving the battery at rest without being charged? All the balance drives are OFF as soon as the charging stops, irrespective whether the EVSE cable is plugged in or not. As far as if understand, no balancing happens except if charging.

Oops. I was thinking continuous bypassing like with the Nissan Leaf.

How can individual cells be manually balanced without taking the pack out?

I believe it can be done with 0x3C3 CAN messages (on the bus between the BMU and the CMUs, I don't know if that can be accessed from the OBD II port). I've seen posts on DiyElectriccar indicating that it can be done, but the poster didn't give details. I might take another look at the CMU firmware to see if I can figure out what the flags in bytes 3 and 4 actually do. Edit: bytes 1 and 2 are known to contain the threshold voltage, above which the bypass FETs turn or pulse on.