Aftermarket iMiEV battery upgrade project

Mitsubishi EV Interest Group
rhills
Site Admin
Posts: 480
Joined: Fri, 25 Jul 2008, 01:57
Real Name: Rob Hills
Location: Waikiki, WA

Re: Aftermarket iMiEV battery upgrade project

Post by rhills »

What would locktite do to the resistance in the connection between the stud and the thread? Would that matter? I guess the current will mainly pass through the contact between the lug on the end of the cable and the metal plate around the stud?
Rob Hills
AEVA Webmaster
2014 Mitsubishi Outlander Aspire PHEV
Petrol Usage to last refill: Jul 2014 - Oct 2020
Total Petrol: 756.3L
ODO: 67508
Av Consumption: 1.12 L/100km
nuggetgalore
Groupie
Posts: 251
Joined: Sun, 11 Dec 2016, 18:02
Real Name: Andreas Tobler
Location: Rowville 3178 V
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by nuggetgalore »

rhills wrote: Wed, 02 Sep 2020, 11:56 What would locktite do to the resistance in the connection between the stud and the thread? Would that matter? I guess the current will mainly pass through the contact between the lug on the end of the cable and the metal plate around the stud?
The alu contact face on the cells is about 18 x 37 mm. That is 15 times more area than the bus bar cross section. It would have to be a very poor fit to be not at least several times more than the 40mm2 of the busbar. And don't forget the ...
"Alminox contains sharp, conductive zinc granules suspended in a viscous petroleum base. Under pressure these granules make high pressure contact points with the parent metal to effect a sound electrical connection, whilst the base material seals the joint to prevent further corrosion."
I am (was) actually more worried about the galvanic issue (SS v Al) than the conductivity between the stud and the parents.
nuggetgalore
Groupie
Posts: 251
Joined: Sun, 11 Dec 2016, 18:02
Real Name: Andreas Tobler
Location: Rowville 3178 V
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by nuggetgalore »

Something else I should have thought of a long time ago but only realised recently.
With the cells standing up rather than laying on their sides, the BMS boards are not protected any more from the cold cooling air by being hidden away between the modules.This may have several issues, one is the cooling air normally would hit the exposed cell body sides but not directly the BMS boards that would see a temperature more akin to an ambient pack temperature.
With the BMS boards connected to the adapter board practically sealing the modules, the cooling air will first cool the BMS boards before taking heat away from the cells (except a bit from the sides of the modules that has conducted through the poly carbon box).
So the temp sensors would see a much cooler module than it actually is.
User avatar
jonescg
Senior Member
Posts: 3618
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Aftermarket iMiEV battery upgrade project

Post by jonescg »

Yes, this is true. It would probably give lower readings, but given the climate it's operating it it shouldn't be an issue most of the time.
AEVA National Secretary, WA branch chair.
stealthhack
Noobie
Posts: 10
Joined: Sat, 04 Jul 2020, 16:18

Re: Aftermarket iMiEV battery upgrade project

Post by stealthhack »

suziauto wrote: Fri, 14 Aug 2020, 15:42 well they are working well , still locked in at 49Ah so we are getting standard range. Trying to work out how to release the other 45Ah but we are hopeful some software being written.. hopefully soon.
Any solution on unlocking the rest of the capacity? Shunt on the current sensor?
suziauto
Groupie
Posts: 109
Joined: Fri, 20 Mar 2009, 16:55
Real Name: graeme Manietta
Location: brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by suziauto »

Hi , yes did the reset and battery capacity routines several different ways with same result.. will only see 49Ah. We have a great company that made our scan tool that has created several " patches" to allow us to install a new max cell capacity but still no joy, plan C is mike and dave are going to look at hacking the BMS so we can install our number. The car goes great but only 130k range. And clearly the data we see confirms its only accessing 49Ah.. we will crack this nut oneday ..
suziauto
Groupie
Posts: 109
Joined: Fri, 20 Mar 2009, 16:55
Real Name: graeme Manietta
Location: brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by suziauto »

we reset and battery capacity routines several different ways with same result.. will only see 49Ah. We have a great company that made our scan tool that has created several " patches" to allow us to install a new max cell capacity but still no joy, plan C is mike and dave are going to look at hacking the BMS so we can install our number. The car goes great but only 130k range. And clearly the data we see confirms its only accessing 49Ah.. we will crack this nut oneday ..
mark_hetho
Groupie
Posts: 119
Joined: Sun, 05 Jul 2015, 01:18
Real Name: Mark Hetherington
Location: Tamworth, NSW

Re: Aftermarket iMiEV battery upgrade project

Post by mark_hetho »

130K range sounds great to me - especially if you get that range for many more years because the batteries are not being cycled as hard.

I'm curious about something. Which part of the pack state of charge is being used? Is it charging the cells all the way full, and then hitting turtle mode before the cells are at low voltage cut out? Or is it emptying the cells fully and only charging part way? Are the cells staying in balance within the capacity they are using?

A custom BMS sounds like a bit of work in getting it to talk to all the other components. But it should sort the problem out.
suziauto
Groupie
Posts: 109
Joined: Fri, 20 Mar 2009, 16:55
Real Name: graeme Manietta
Location: brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by suziauto »

I feel its using the lower part but I really need to examine it closer.. I agree the lifespan should be great even as it is now .. but much better too if we can get to the other 45Ah.
richi
Noobie
Posts: 20
Joined: Thu, 09 Jan 2020, 21:06
Real Name: Richi Jennings

Re: Aftermarket iMiEV battery upgrade project

Post by richi »

It'll be the top half, surely? You're still charging the cells up to 4.1 V and then Coulomb counting the discharge down. Right?

But I still don't understand what you're actually seeing, Graeme—is it that the car slows and stops after using 49 Ah? Or is it just that the GOM is pegged at 130 km, but you can still drive for, say, 200 km?
stealthhack
Noobie
Posts: 10
Joined: Sat, 04 Jul 2020, 16:18

Re: Aftermarket iMiEV battery upgrade project

Post by stealthhack »

I think that the car is doing coulomb counting and after 49Ah just stops the charge. And it watch for under-Voltage for battery cutoff, to protect the battery at low SOC.
richi
Noobie
Posts: 20
Joined: Thu, 09 Jan 2020, 21:06
Real Name: Richi Jennings

Re: Aftermarket iMiEV battery upgrade project

Post by richi »

stealthhack wrote: Tue, 08 Sep 2020, 16:24I think that the car is doing coulomb counting and after 49Ah just stops the charge.
You mean Coulomb counting upwards? Anything's possible, but that would be weird.

Charging is all about firing a constant current into the cells until the pack Voltage reaches 360.8 V, then finishing off with a constant-Voltage phase (probably with a cell-balance phase inserted). I guess it's possible the BMU is cutting off the charge session early, but I don't think I've heard Graeme say that.

What I'm hoping is that Graeme or someone else on the ground can say what the symptoms are. Or if I've missed a message in this crazy thread, could someone point to him saying what actually happens?

My semi-educated guessumption has always been: If the BMU caps the estimated SoH, it'll cap the GOM, and it might cause the turtle to illuminate after Coulomb counting down to zero, but the car would carry on driving fine—only slowing down when the lowest cell Voltage reached a critical threshold. If so, an owner might think that's not so bad.

Many owners already see this behaviour when the SoH estimate is pessimistic: The turtle lights too early, and they can carry on driving for a good distance before it finally slows down when the Voltage reality catches up with the SoC estimate.

Alternatively, if the BMU is being too clever by half, I think your suggestion of hacking the current sensor is a good one. In my book, it's often better to make a simple analogue hack, rather than a complicated digital one.
User avatar
coulomb
Site Admin
Posts: 4323
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by coulomb »

richi wrote: Tue, 08 Sep 2020, 16:50My semi-educated guessumption has always been: If the BMU caps the estimated SoH, it'll cap the GOM, and it might cause the turtle to illuminate after Coulomb counting down to zero, but the car would carry on driving fine—only slowing down when the lowest cell Voltage reached a critical threshold.
My understanding is that the turtle appears when the GOM reaches 10-12%, and stops the car when the estimated range goes to zero. It will also turtle when any cell reaches 3.0 V.

Edit: Reference: about ¾ through this post: http://myimiev.com/forum/viewtopic.php? ... =10#p39802
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.
stealthhack
Noobie
Posts: 10
Joined: Sat, 04 Jul 2020, 16:18

Re: Aftermarket iMiEV battery upgrade project

Post by stealthhack »

turtle at first comes when battery was new at 12%(when the car was at 5-6k. km) after 26k km. it was on 10% SOC. That means that the car is starting to unlock bottom charge for using(10% is emergency charge and the cells are topped up to about 4,105v(about 90-95% real SOC). Graeme can check what's going on, with OBDLink dongle and CanIon.
Edit: my car is european version of i-miev, Citroen C-Zero with 80 cells under the floor.
User avatar
coulomb
Site Admin
Posts: 4323
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by coulomb »

stealthhack wrote: Tue, 08 Sep 2020, 17:32 That means that the car is starting to unlock bottom charge for using(10% is emergency charge
Interesting. Battery University mentions this as well, but no details.
and the cells are topped up to about 4,105v(about 90-95% real SOC).
Are you saying that the OBC targets 4.105 x 88 instead of 4.100 x 88? That seems a minuscule difference.
Graeme can check what's going on, with OBDLink dongle and CanIon.
So you're saying that the OBC's target voltage appears on a CAN bus that is accessible through Canion, and these CAN IDs and their formatting are well known?
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.
stealthhack
Noobie
Posts: 10
Joined: Sat, 04 Jul 2020, 16:18

Re: Aftermarket iMiEV battery upgrade project

Post by stealthhack »

you can see everything on CanBus, voltage, available capacity, temperature between cells, max cell voltage, avg voltage deviation, and so on. The OBC have hard-coded the max voltage target, and everything over that voltage triggers bleeding resistor.
richi
Noobie
Posts: 20
Joined: Thu, 09 Jan 2020, 21:06
Real Name: Richi Jennings

Re: Aftermarket iMiEV battery upgrade project

Post by richi »

coulomb wrote: Tue, 08 Sep 2020, 17:19My understanding is …
Mike, to be clear, I'm interested in finding out what the re-celled car is doing, so we can make a guess at how the BMS is reacting to a higher capacity pack and then how best to hack around the problem. I hope we're all well aware of how a normal car behaves.
richi
Noobie
Posts: 20
Joined: Thu, 09 Jan 2020, 21:06
Real Name: Richi Jennings

Re: Aftermarket iMiEV battery upgrade project

Post by richi »

coulomb wrote: Tue, 08 Sep 2020, 17:47
and the cells are topped up to about 4,105v(about 90-95% real SOC).
Are you saying that the OBC targets 4.105 x 88 instead of 4.100 x 88?
I believe 0.005 V is the limit of accuracy (at least as broadcast in the PIDs).
nuggetgalore
Groupie
Posts: 251
Joined: Sun, 11 Dec 2016, 18:02
Real Name: Andreas Tobler
Location: Rowville 3178 V
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by nuggetgalore »

suziauto wrote: Tue, 08 Sep 2020, 13:10 I feel its using the lower part but I really need to examine it closer.. I agree the lifespan should be great even as it is now .. but much better too if we can get to the other 45Ah.
Can you check the cell volt shortly before it stops charging?
Or even more valuable would be to know if the cell balancer drives are switching on/off, in other words are the cells getting balanced.
On my 2010, it appears the balance drivers are working right from the start al low charge to the very end. But of course the cells are degraded to 32 Ah.
nuggetgalore
Groupie
Posts: 251
Joined: Sun, 11 Dec 2016, 18:02
Real Name: Andreas Tobler
Location: Rowville 3178 V
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by nuggetgalore »

Just to clarify, I use the i909 to check the pack( for some reason the Canion app while still correctly reads and displays some date, the individual cell volts and temperatures are locked.
The i909 is clumsy to read, but has much more info than Canion or EvBatMon, reads all pack and cell voltages, module temps and other trivia. The cell balance drivers are interesting, maybe one day I'm catching the moment when the charger goes to sleep for some minutes (usually about three quarters to full but irregularly). We assume it does an intermediate balance cycle and then continues to charge to full with max current until the taper towards full.I may program an alarm into the PLC that records the charge times and currents to alert me to the pause in the charge, so I can check what the drivers do.
mikedufty
Senior Member
Posts: 660
Joined: Tue, 02 Dec 2008, 00:15
Real Name: Michael Dufty
Location: Perth, Western Australia
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by mikedufty »

richi wrote: Tue, 08 Sep 2020, 18:51 I hope we're all well aware of how a normal car behaves.
Interestingly all the comments on turtle modes are at odds with my experience where I was able to drive to the point of the car refusing to drive anymore, with 10km range remaining showing and no turtle. Maybe there was something wrong with mine but it didn't do anything wrong in normal driving. The run flat drive was all uphill going out and downhill back which I think messed with the range estimation.
User avatar
coulomb
Site Admin
Posts: 4323
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Re: Aftermarket iMiEV battery upgrade project

Post by coulomb »

richi wrote: Tue, 08 Sep 2020, 18:51 Mike, to be clear, I'm interested in finding out what the re-celled car is doing, so we can make a guess at how the BMS is reacting to a higher capacity pack and then how best to hack around the problem.
Time for an update. There is currently one iMiEV Minicab fitted with the 93 Ah cells. These are NMC cells packaged in a case that was only used so far for LFP cells. Graeme asked the manufacturer to put NMC into these cases. So these cells are unique at present. If all goes well, the NMC might be put into a custom battery case that fits even better, and may fit ≈100 Ah into those cells. But the range problem really needs to be sorted first.

My investigations find that NMC cells have a lower voltage characteristic, at other than 100% SOC, than LMO [ edit: was NMO, sigh ] cells; it's just the way that the chemistry works. I overlaid the curves from a generic NMC cell (from a technical paper, not any manufacturer's datasheet) with the curve for LEV50 cells. I found that the initial decrease in terminal voltage versus SOC is more than twice as steep in NMC than the LEV50s (which seem to be between LMO and NMC, much closer to the LMO curve). The BMS is expecting LEV50 cells, of course. So I'm guessing that it decides "wow, these are pathetic LEV50s, I'll reduce the learned capacity". That's what we see; after the initial BMS reset command from a scanner which stores 48.0 Ah into the learned capacity variable, the learned capacity actually decreases. I saw 46.1 Ah for example. Graeme assures me that these cells have been independently tested twice, and both confirmed around 93 Ah capacity. So that means that the BMS will probably never learn a higher capacity than 48 Ah, and will probably settle at something lower, perhaps 45 Ah.

When I first looked at the Minicab, it was showing 30 km on the range remaining meter (RR, also known as the Guess O Meter, or GOM). The owner told me that the turtle often happens at about 20-30 km RR. The lowest cell voltage was 3.825 V, max 3.885 V (these cells could do with some more balancing). 3.825 V seems to be around 50% SOC on an LEV50 cell (depending on rate of discharge), and about 70% SOC on a generic NMC cell. Other sources indicate that the turtle happens at 10-12% estimated SOC: nearer 12% SOC when new, closer to 10% as the battery pack ages and the BMS "unlocks" a little extra capacity to compensate. The turtle is also activated if a single cell reaches 3.0 V (it's not clear if this is a "latched" state, or the motor power is reduced to prevent any cell going lower than 3.0 V). The scanner reported "Charge state (Control)" (SOC) as 35.5%, and "Charge state (Display)" as 36.5%. So perhaps this was fairly far from the turtle; 35.5 - 12 = 24.5%; initial RR is around 98 km, so that would be another ~24 km).

The "right" way to fix this would presumably to somehow rewrite or patch the BMS firmware. Unfortunately, I've had no success so far obtaining an image to analyse and patch.

I've started on a prototype gadget that should reduce the current sensor readings by between 10 and 50%, screwdriver adjustable. The final version should be smaller and cheaper. I'll start with 10% reduction, which should hopefully show a noticeable increase in range (so we know it's working), but also any other side effects (if any) should be noticeable. Hopefully, the other side effects at 10% change won't be enough to damage anything. It should tell us whether it is necessary or desirable to also add a digital "man in the middle" CAN bus device to correct the current readings for other systems. If so, we'd also add a correction to the RR meter, to make it better reflect the actual range available.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.
richi
Noobie
Posts: 20
Joined: Thu, 09 Jan 2020, 21:06
Real Name: Richi Jennings

Re: Aftermarket iMiEV battery upgrade project

Post by richi »

Ohhhh, I didn't realise the new cells weren't the same chemistry. In that case, I dare say hacking the current sensor won't be enough—it would compensate for the Coulomb counting downward that I assume is happening, but not of course for the trigger at critical low cell Voltage.

Not sure from your description, but it sounds like the new cells reach that Voltage at a much higher SoC, right? Likely you'd need to intercept and rewrite the messages from the CMUs to the BMU, and fake the Voltages.

I don't know if anyone has pulled out the BMU hardware to see if it's actually reflashable. The Outlander PHEV community's belief is that its BMU isn't (and I assume it's similar).

What a pain in the arse. Thank you, Mike and Graeme for blazing this trail. For what it's worth, I send my best wishes from t'other side of the globe.
User avatar
jonescg
Senior Member
Posts: 3618
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Aftermarket iMiEV battery upgrade project

Post by jonescg »

Should quickly add that the original LEV50 cells were LiMnO2 chemistry, not LFP. LiMnO2 cells maintain a higher operating voltage in the first 50% of the SOC, unlike NMC which tend to show a quick drop off to 4 V and fairly linear drop over the SOC.
https://pushevs.com/2015/11/04/gs-yuasa ... vs-lev50n/

NMC cell discharge curve:
Discharge test.JPG
Discharge test.JPG (27.57 KiB) Viewed 777 times
AEVA National Secretary, WA branch chair.
tonyw
Groupie
Posts: 46
Joined: Thu, 26 Jun 2008, 01:08
Real Name: Tony Wyatt
Location: Sydney

Re: Aftermarket iMiEV battery upgrade project

Post by tonyw »

I'm sure I'm telling you nothing you haven't already calculated, but...

I would be careful about "desensitising" the current sensing. The current sensor is probably also used to limit the current if Leadfoot Leroy floors the pedal on a steep hill or driveway. If the indicated current is lower, you risk overheating something, and there are electronic components in there that don't like to be overloaded. The motor can probably take a 20% overload briefly, but silicon can't.
cheers

tony

Post Reply