Speak EV - Electric Car Forums banner

1 - 20 of 39 Posts

·
Banned
Joined
·
1,827 Posts
Discussion Starter #1
Is there anyone on this forum with expertise who can facilitate cell replacement on old I-Miev cars, that are fully working but have lost battery?

I feel I would not get any fair value to sell or trade in my I-Miev but other than reduced battery it is working perfectly fine, and you get to know your car, and I have a hunch it could run nicely for years to come. Replacing the lost cells could be a good cost effective option for me going forward.

Bit niche and specialist this though isn't it? I am willing to travel, I can get my I-Miev to anywhere in the UK and drove it home from Nottingham to Scotland last year.
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
I too am looking into this. I am considering a project to extend the range of an i-Miev with newer battery technology so we'll see how that goes.
I will post any info here and also, hopefully, I will be able to get some help with it from the community.
 

·
Registered
Joined
·
3,762 Posts
Is there anyone on this forum with expertise who can facilitate cell replacement on old I-Miev cars, that are fully working but have lost battery?

I feel I would not get any fair value to sell or trade in my I-Miev but other than reduced battery it is working perfectly fine, and you get to know your car, and I have a hunch it could run nicely for years to come. Replacing the lost cells could be a good cost effective option for me going forward.

Bit niche and specialist this though isn't it? I am willing to travel, I can get my I-Miev to anywhere in the UK and drove it home from Nottingham to Scotland last year.
I've successfully done a 4 cell swap on my Ion. It was a lot of work but doable and I did it by myself over a several day period. I have no intention of doing any more though, I have enough renovating and DIY already to keep me busy. ;)

However I fear that the original premise here may be flawed. I interpret "Replacing lost cells" as you thinking that you have a few bad cells and that replacing them will restore significant range to your car. Unfortunately this is very unlikely to be the case, based on my own experience.

My car has had more degradation than expected and I had what appeared to be three clearly "weak" cells, so I decided to replace the four worst cells in the pack to see if the usable capacity increased. It did, but only by less than 2Ah, so really not worth the effort. (Although to be fair, rapid charging speeds - which were a bit slower with the weaker, high resistance cells returned to normal so in that sense it was a success) In the months since then some of that gain has ticked away already, although that might be the BMS being pessimistic as it is following a programmed in degradation model most of the time.

Now that those cells are replaced there are other cells that are now the weakest cells that are limiting the capacity - you could go on like this forever until you replaced the entire pack. Just not worth it.

Yes if you have a clearly faulty cell that is extremely at odds with other cells and the range is say < 1/4 of what it should be for its mileage then it may be worth trying to swap that one cell and there have been successful repairs like this with a single clearly very faulty cell.

However you are probably just seeing the same I did - a spread in degradation between cells (probably) as a result of frequent deep discharges. What look like weak/faulty cells are just cells that are slightly ahead of the rest in the decay curve, but as soon as you replace those there will be more that will follow suit soon after.

Sorry to be a buzzkill but after putting in all the hours I just don't think individual cell swaps are worth trying on a significantly degraded pack unless the difference between the "faulty" cell and the rest of the cells is very, very large and obvious. And if multiple cells seem to be "faulty" it's highly unlikely they are actually faulty, just that there are a few cells that have degraded a little faster.

As cells degrade there is an initial shallow degradation rate - once they reach a certain point in their life (around 1000 cycles on these cells it seems) the degradation rate dramatically increases - the cells are still usable but their degradation progresses at a faster rate. Each different cell reaches this "knee point" at a slightly different time so as they all cross that knee point it can exaggerate differences between cells.

A better bet would be replacing the entire pack's worth of cells with after market replacements, but that's costly and a whole project in itself to make the BMS happy and be sure that it is treating the cells right. (For example the LEV50's in these can take higher charge rates than a parallel group of 18650's of the same Ah capacity, so the BMS would try to charge 18650's too fast unless you re-programmed it)
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
I was actually thinking more along the lines of completely replacing or re-engineering a battery pack. Possibly with new cells that use newer technology.
This it seems would possibly mean ditching the existing BMS and building a custom Arduino setup. Who knows. certainly not me...Yet.

This is going to be a project (because I like to do these things) rather than a necessity at this point. I have an 8 mile a day total commute so I will keep using the car until it's no longer viable for me and then dig into it. In the mean time I have a lot to learn and appreciate yours and everyone else's help and advice.
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
So I'm looking at 18650 3.7v 5000mAh Li-ion cells. Using these would of course mean assembling my own battery packs from scratch BUT... To purchase just over the required replacement capacity for the standard I-miev, the batteries I have found so far would only cost £660 in this format. That sounds too good to be true.
Am I calculating the capacity correctly? Is 5000mAh = 5Ah? Is there another factor I'm not considering?

EDIT: Never mind. I'm likely looking at fake cells. smh
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
Cheapest cost of new cells with the equivalent storage so far £4250 - Using Molicel INR-21700-P42A.This is the cost for 1100 individual cells and doesn't include wiring, sub assembly, custom BMS etc.

I think, if I'm going to do this, I will be making a single LEV50 replacement pack first and see if I can get that to work.
 

·
Registered
Joined
·
3,762 Posts
Yeah this is the problem with this sort of after market replacement - it's not cheap, as buying the cells in low quantities and cobbling them together is a lot more expensive than what the manufacturer could do.

And without being able to reverse engineer and/or reprogram the BMU, you could easily spend a lot of money on cells only to find that the BMU is unhappy with the measured performance of the cells, or can't be programmed to treat the cells properly. (Charge rate vs temperature for example - which is pre-programmed in the ECU for the specific characteristics of the LEV50)
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
So far I've found a supply of branded, new 21700 5Ah cells at £2.77 each. I can fit 20 into the LEV50 envelope effectively doubling the capacity of that form to 100Ah
Replacing all 88 cells would cost £4875 in just cells but would be a 32kwh unit. Alternatively half the cells, maintaining the original range, would cost £2438, so getting a little closer cost wise.
Still much research to be directed to battery management/thermal management and the interface with the existing electronic controls built in to the car.

128823
 

·
Registered
30KW Tekna (2017)
Joined
·
218 Posts
I find this discussion so interesting as I nearly bought a triplet (ion) before I opted for my leaf. I’ve got battery problems myself and my understanding and knowledge about EV’s has increased considerably sine my purchase.

The more I learn about not only battery degradation but also weak individual cells in my opinion is going to be a massive block to ev uptake by the population. As EVs become more mainstream, unless manufacturers allow what you are trying to accomplish by supplying new cells or packs for a reasonable price or opening up the bms for reprogramming it will not be long before the scandals hit the headlines.

Green cars that are basically disposable after 10 years due to degradation.

We are all early adopters and enthusiasts so maybe accept these facts more but the motoring population as a whole will not, I don’t think.

It would be so simple to program in some hidden back door so owners could reset the bms to a different kw or have an option of reset bms to new battery pack to help people keep thir cars on the road longer.

How can manufacturers claim evs are so much more reliable as so few moving parts but conveniently forget to mention the big chemical reaction under the car is on borrowed time!
 

·
Registered
30KW Tekna (2017)
Joined
·
218 Posts
I find this discussion so interesting as I nearly bought a triplet (ion) before I opted for my leaf. I’ve got battery problems myself and my understanding and knowledge about EV’s has increased considerably sine my purchase.

The more I learn about not only battery degradation but also weak individual cells in my opinion is going to be a massive block to ev uptake by the population. As EVs become more mainstream, unless manufacturers allow what you are trying to accomplish by supplying new cells or packs for a reasonable price or opening up the bms for reprogramming it will not be long before the scandals hit the headlines.

Green cars that are basically disposable after 10 years due to degradation.

We are all early adopters and enthusiasts so maybe accept these facts more but the motoring population as a whole will not, I don’t think.

It would be so simple to program in some hidden back door so owners could reset the bms to a different kw or have an option of reset bms to new battery pack to help people keep thir cars on the road longer.

How can manufacturers claim evs are so much more reliable as so few moving parts but conveniently forget to mention the big chemical reaction under the car is on borrowed time!
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
20 x 21700 LG M-50 cells ordered and on their way for initial testing.
This is enough for a 100 Ah cell, double the capacity of the LEV-50 used in the car currently, lighter and in the same form factor.
If I can get it to work, a full pack (at the price I paid for 20) would cost £5200 just in cells but give the car a real world range of about 170 miles.

Could be a fairly expensive project...
 

·
Registered
Joined
·
3,762 Posts
20 x 21700 LG M-50 cells ordered and on their way for initial testing.
This is enough for a 100 Ah cell, double the capacity of the LEV-50 used in the car currently, lighter and in the same form factor.
If I can get it to work, a full pack (at the price I paid for 20) would cost £5200 just in cells but give the car a real world range of about 170 miles.

Could be a fairly expensive project...
Do you have access to a MUT-III ? (for i-Miev......Diagbox for C-Zero or Ion)

Without one you're not going to be able to perform a full battery capacity re-calibration to make the BMU aware of a 2x increase in available capacity. If the BMU does not believe there is increased cell capacity it will not let you use the extra capacity. (It will still go into turtle mode when 90% of the current recorded Ah capacity is used)

While the normal learning algorithm will gradually make increases to the Ah figure if it measures a significant increase in capacity each charge cycle it is very slow to adapt, and rightly so, as batteries don't normally increase in capacity on their own. It adapts at something like a 1Ah increase per 1000 miles. An owner over on myimev who did a full pack swap from a pack with about 26Ah remaining to a second hand one with a presumed capacity in the mid 30's without access to a MUT-III has watched his Ah figure creep up painfully slowly over many many months and a lot of driving, and I think it's still only up to about 32Ah.

This would not be feasible if you were fitting cells with twice the capacity and wanted to be able to use that capacity any time soon. The calibration process (I've done it a couple of times on my Ion using Diagbox, once after I did a 4 cell swap) basically involves discharging the pack until the lowest cell is around 3.75 volts, then doing a 100% Level 2 charge under the supervision of the diagnostic tool. At the end of the process the recorded Ah figure is immediately updated to the new measurement without any delay or gradual learning process.

However there is a huge unknown in that nobody knows what the highest Ah figure is that the BMU or calibration process will accept. In other words does it perform any kind of sanity check that says if measured capacity is above 50Ah, don't believe it ? (Which would require tweaking of the BMU firmware to get around) Or would it happily calibrate to 100Ah and unlock all that range ?

Nobody knows as far as I'm aware, and if you did go through with building a 100Ah pack you might be the first to find out... :)

I definitely wouldn't go ahead with building a high capacity pack without access to a MUT-III to perform the battery calibration though - you're guaranteed to not be able to access that extra capacity without being able to perform the calibration.
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
Do you have access to a MUT-III ? (for i-Miev......Diagbox for C-Zero or Ion)

I definitely wouldn't go ahead with building a high capacity pack without access to a MUT-III to perform the battery calibration though - you're guaranteed to not be able to access that extra capacity without being able to perform the calibration.
I do of course have to figure out the control interface side of this project and I thank you for the information you've posted RE: MUT-III. I will definitely look at that next.

It has become clear pretty early on that nobody (So far... (Eternally optimistic)) outside of the OEM knows about the inner workings of the BMS and ECU and the interface between the two. I'm determined to take that on one way or another.

Thanks again. All constructive input and criticism is greatly appreciated.(y) (I'll be filtering out the blatant cynicism though.;))
.

[EDIT] Wow. MUT-3 is a whole new area to delve in to. Full of units with slightly different model numbers, multiple software platforms, innumerable software installation problems and no reference so far to the I-Miev directly... ONWARDS!! 😂👉
 

·
Registered
Joined
·
3,762 Posts
Sorry if you're detecting any cynicism, I don't intend it that way. :)

However having been involved in diagnosing bad cells on my own car, replacing them, going through the process of re-calibrating the BMU, (having access to the Diagbox/Lexia 3 dealer diagnostic tool the C-Zero/Ion versions of the car use) making some mistakes along the way (not getting the SoC balance quite right of the replacement cells requiring a couple of months worth of charging cycles to equalise) I have some insight into the battery system on these cars.

So consider my posts a heads up for some of the challenges you may face that will need to be overcome, not cynicism. As I said, a diagnostic tool that can force a re-calibration of the Ah figure would seem to be essential if you're trying to double the capacity of the original battery because unless you can get that reported Ah figure to budge, the BMU will refuse to let you use the extra capacity - I know that much for sure from actual testing.

There are two things that put the car into tortoise mode - calculated SoC going below about 10% or individual cell voltages dropping below 3.0 volts while driving. The SoC is always referenced to a full charge or 4.1 volts per cell.

So if the BMU thinks it has 45Ah available but you really have 100Ah available, as the SoC drops from 100% the SoC scale will be based on this assumption of 45Ah, and it will go into tortoise mode after approx 41Ah has been discharged, even though you have nearly 60Ah left and the cell voltages will be a lot higher than it would have expected for a 10% SoC.

You would think that it would notice the cell voltages are still high and let you keep driving but it doesn't! Also it takes many charge cycles for it to "believe" there is an increase in capacity and it will only increment it by 1Ah at a time. I do about 1000 miles a month and mine will only increase 1Ah after about 2 weeks of driving.

The low voltage tortoise mode is a bit different - if any cell dips below about 3.0 volts during driving it will instantly bring on the tortoise icon and start to reduce power output, however if you ease off on the power and the voltage goes up the tortoise mode will go off again.

In my testing it seems you would normally never hit this low voltage threshold (as 0% SoC is about 3.6 volts) unless you either have a faulty cell or the Ah figure is way higher than actual state of the battery - for example I had it happen to me when I had reset the BMU (Ah 45.8Ah) but my pack is only about 33Ah. It let me keep driving well past the point where the tortoise mode would have come on until I had a very sudden tortoise mode at 3.0 volts per cell and the car limited the power and basically shut down within a few hundred metres of driving.

So getting the BMU to accept the new higher capacity of the replacement cells is critical to be able to utilise it. Hopefully it will be willing to learn a higher figure when the calibration is run, but I'm not aware of a single person who has tried yet. Someone will have to take the chance and go first I suppose. If it's any consolation, there was a 20kWh prototype of the car tested before it released as 16kWh, so there might be some extra wiggle room still there in the ECU code....

Another interesting problem to solve with replacement cells is the temperature sensor for the cells. 66 of the 88 cells are monitored by temperature sensors on the CMU boards, but despite having had the modules apart, replaced cells, taking high res photos of the CMU boards, I'll be darned if I can positively identify the temperature sensors... I was expecting to see sensors poking out the bottom of the CMU board to make physical contact with the flat flange of one of the terminals, but there is no mechanical contact. My guess - and it's only a guess - is that they are using an optical infrared sensor pointing at the cell to measure the temperature using IR emissivity. Whether it points at the rough steel plate on the top of the cell or the cell body I'm not sure. I was in a bit of a hurry to get my pack back together and didn't investigate further.

The temperature readings primarily affect rapid charging. If any cells are below about 11C, the charge rate is limited to 22kW. (and below about 0C I think it drops again to 11kW) If the temperature is between about 12C and 40C the charge rate is allowed to go to the full 43kW. (Although the original cells can only sustain this high charge rate below about 40% SoC) Above 40C it throttles back again to 22kW and lower. Highest I've ever seen my cells on extreme testing was 46C.

It also controls the cooling of the pack. If all cells are below 20C the pack is not cooled at all. If any cells are between 20C and 30C the fan will blow ambient air through the pack. If any cells are above 30C the air con compressor comes on as well to chill the air, (you'll see the AC light come on) and as you go higher it gets more urgent, by 40C the air con is working pretty much on full. It looks like the BMU tries to target 30-40C for rapid charging, which is where it will charge the fastest.

Figuring out how the temperature sensors on the CMU work and allowing them to sample at least one cell in each parallel bundle (of the 66 cell slots that are monitored) correctly is probably pretty important from a safety perspective - you don't want to overcharge hot cells (fire risk) or very cold cells (dendrite growth risk resulting in eventual short circuits)
 

·
Registered
Joined
·
3,203 Posts
Hi all, there is a post on the AEVA forum dealing with fitting large capacity packs to the iMiev and clones. The consensus seems to be that the BMU is hard programmed for a max capacity of 60Ahr.

Workarounds have been suggested such a two parallel current shunts to trick the BMS into thinking it is only delivering 1/2 the actual current.........
 

·
Registered
Joined
·
3,762 Posts
Hi all, there is a post on the AEVA forum dealing with fitting large capacity packs to the iMiev and clones. The consensus seems to be that the BMU is hard programmed for a max capacity of 60Ahr.
Ouch! That's the biggest worry about any sort of capacity increase project like this - a lot of money could be sunk into cells only to have the BMU to refuse to play ball. Do you have a link to the discussion ?
Workarounds have been suggested such a two parallel current shunts to trick the BMS into thinking it is only delivering 1/2 the actual current.........
I don't like the sound of that. Surely that's going to confuse it when the current reported from the battery pack doesn't match the current drawn by the MCU ? And wouldn't that cause it to try to charge the battery twice as fast when rapid charging if charge current is under reported ?

I wonder what the feasibility is of getting an ECU reprogramming/unlocking firm to have a look at the firmware to see if it could be tweaked ? Well above my skill level I'm afraid, even though I work in computers..
 

·
Registered
Mitsubishi i-Miev
Joined
·
24 Posts
Another interesting problem to solve with replacement cells is the temperature sensor for the cells. 66 of the 88 cells are monitored by temperature sensors on the CMU boards, but despite having had the modules apart, replaced cells, taking high res photos of the CMU boards, I'll be darned if I can positively identify the temperature sensors... I was expecting to see sensors poking out the bottom of the CMU board to make physical contact with the flat flange of one of the terminals, but there is no mechanical contact. My guess - and it's only a guess - is that they are using an optical infrared sensor pointing at the cell to measure the temperature using IR emissivity. Whether it points at the rough steel plate on the top of the cell or the cell body I'm not sure. I was in a bit of a hurry to get my pack back together and didn't investigate further.
129057


Regarding thermal management / measurement, I still do not know yet either but my VERY preliminary impression (which I haven't investigated yet) was that changes in temperature would alter the material properties of the thin metallic strips circled in the picture, either resistivity or physical dimensions, and that the resistance or relative proximity would be measured and referenced as a temperature guide by the BMS...

What do you think?
 

·
Registered
Joined
·
3,762 Posts
View attachment 129057

Regarding thermal management / measurement, I still do not know yet either but my VERY preliminary impression (which I haven't investigated yet) was that changes in temperature would alter the material properties of the thin metallic strips circled in the picture, either resistivity or physical dimensions, and that the resistance or relative proximity would be measured and referenced as a temperature guide by the BMS...

What do you think?
I'm not sure - I think the black bits circled are just a flexible connector essentially. The screw hole on the floating "island" board is a small screw which screws onto the plate that the battery post attaches to. It gives an electrical connection from the cell to the LTC chip which does both voltage monitoring and balancing. The board being an island with a flexible connector is probably just to allow proper alignment with the cell as the cells aren't tightly packed. (They have a sticky foam strip between each cell about 5mm thick so there is a little bit of movement possible between cells)

To the top left of each screw hole on the bottom left two island boards you'll notice two tiny soldered component legs with traces going to them, (near the yellow blobs) this is probably the sensor.
 
1 - 20 of 39 Posts
Top