Lindqvist method : DIY BMU Battery Reset

[michalsarna]

Who would forbid them ? :?:

Hmmm.. let's think about it. :?:
Mitsubishi Japan ?
As they did !

And I know that there is one leak about it, but I can tell You that my local Authorized Mitsubishi Service Center couldn't set me up since they don't have this procedure but ware able to measure my battery capacity.
And I know from people who had they PHEVs serviced there they did triple for them.
I've even wrote to Mitsubishi Poland to ask about this procedure, and they responded that there is no such procedure and battery condition will be evaluated by service center and only thing they can do is to exchange battery for new one.

So hmmm ....

 

[user 4407]

Has anyone successfully Lindqvisted a MY2018 in the USA? I have tried three times, increasing the 12V battery disconnect and charge times with each cycle. Last attempt, I had 12V battery disconnected for 18 hours and then after connecting for 3 seconds to cycle the charge relays, I left it on the 6Amp charge rate for about 26 hours. Still no joy here.

 

[jaapv]

Who would forbid them ? :?:

Hmmm.. let's think about it. :?:
Mitsubishi Japan ?
As they did !

And I know that there is one leak about it, but I can tell You that my local Authorized Mitsubishi Service Center couldn't set me up since they don't have this procedure but ware able to measure my battery capacity.
And I know from people who had they PHEVs serviced there they did triple for them.
I've even wrote to Mitsubishi Poland to ask about this procedure, and they responded that there is no such procedure and battery condition will be evaluated by service center and only thing they can do is to exchange battery for new one.

So hmmm ....

Mitsubishi Japan WRITES the firmware. Whatever your local importer does will certainly not influence the company policy. At any rate, they have been selling tens of thousands of these vehicles over here since 2013 without such shenanigans.

 

[MadTechNutter]

* According to the specs the LEV40 cells hold for 5500 full cycles until 80% SoH, that's 247 500km with an average range of 45km per charge.
5500 is FULL cycles mind's you, from 3V to 4.1V. This is Never reached with EV packs, so the expected life until 80% SoH is far more the 5500 cycles, estimated I would say it's 3x that much.

Please provide proof for that claim of 5500 full cycles from the official website:
https://www.yuasabatteries.com/
and NOT some data sheet that somebody could had made up and is now circling on the internet, which I have seen already, possibly by somebody who is trying to sell used batteries.

So far I could only find on a technical report from Yuasa that a similar battery, the LEV50, degrades to 90% after only 700 100% DoD cycles, where the wording might be unclear if this is related to a discharge under a 4C condition.

In general the degradation using a discharge from 100% down to 0% in comparison to 100% down to 30% is not going to be very different.
Generally for Li-Ion batteries you will get roughly 30% more of these 70% DoD discharge cycles than the full 100% but overall degradation will be roughly the same and certainly not 3x as many cycles.
Please provide evidence of that claim for the LEV40 or any other battery for that matter from their official manufacturers websites.

According to https://www.secondlife-evbatteries.com/pdf/Datasheet-LEV40-Module.pdf you can pick up the information from https://www.gs-yuasa.com/en/technic/archive.php.
If you are motivated enough to translate and search for it your self then you have what you are asking for.

That is exactly where I have researched as I mentioned above and find for the LEV50 only 700 cycles until 90%
The Archive goes only to 2006

Edit:
Found you a very good proof of cycle count when a cell is managed correctly.
https://en.m.wikipedia.org/wiki/Tesla_Powerwall.

.

from that link:

The 10 kWh battery as originally announced has a nickel-cobalt-aluminum cathode,[9] like the Tesla Model S,[10] which was projected to function as a backup/uninterruptible power supply, and had a projected cycle life of 1000–1500 cycles.[11][12][13][14]

Sounds more like it !!!

, where do you get your information from?

15 years servicing laptops and power tools with Li-Ion batteries

 

[MadTechNutter]

@michalsarna,
do you have the same charger as mine in the image above?

Nope.

Thank you. That is very valuable information

This might confirm my theory that different chargers may be the reason why this method works only for some, based on the time it takes for the charger to commence charge.
As mentioned earlier, my charger takes so long that in the car 6 relays click on before it starts but your audio sample shows 4 relay clicks.
So you might have managed to get the car in an undefined state by being able to remove the 12V before the last 2 relays go on but got it to start charging.

I will know more once I have built my own charger, that I will program so that it can start charge almost immediately after the correct +6V signal level of the pilot is detected.

 

[Lon12]

Has anyone successfully Lindqvisted a MY2018 in the USA? I have tried three times, increasing the 12V battery disconnect and charge times with each cycle. Last attempt, I had 12V battery disconnected for 18 hours and then after connecting for 3 seconds to cycle the charge relays, I left it on the 6Amp charge rate for about 26 hours. Still no joy here.

I have a North American GT model and have not been able to get this to work. I have the same question as you. If someone is successful please let us know.

 

[elm70]

According to https://www.secondlife-evbatteries.com/pdf/Datasheet-LEV40-Module.pdf you can pick up the information from https://www.gs-yuasa.com/en/technic/archive.php.
If you are motivated enough to translate and search for it your self then you have what you are asking for.

I for one trust a company that puts their reputation and business at stake when providing that information.
If you don't trust them then that's a problem you need to work with and checking the technical archive is a method to do just that.

YUASA LEV40 Cell
Nominal voltage 3.65V
Operating voltage range 2.65V to 4.0V
1-hr rate typical Capacity 25C 40Ah

This is shocking to me

3,65v nominal voltage ? In Yuasa official pages the LEV50 has much higher nominal voltage

Here SecondLifeEVBattery suggest to use a much different voltage range :
Mitsubishi force an usage between 4.1v down to 3.8v
SecondLife suggest between 2.65v to 4.0v

Opposite point of view

Can be that a little company know better then Mitsubishi ?

 

[elm70]

Has anyone successfully Lindqvisted a MY2018 in the USA? I have tried three times, increasing the 12V battery disconnect and charge times with each cycle. Last attempt, I had 12V battery disconnected for 18 hours and then after connecting for 3 seconds to cycle the charge relays, I left it on the 6Amp charge rate for about 26 hours. Still no joy here.

I did tried few times now ... still no joy ... mine is a MY2013 from NL .. which should accept the Lindqvist method

Either is not working for some firmware update, or the procedure is very tricky to be repeated

 

[Pomst]

Please provide proof for that claim of 5500 full cycles from the official website:
https://www.yuasabatteries.com/
and NOT some data sheet that somebody could had made up and is now circling on the internet, which I have seen already, possibly by somebody who is trying to sell used batteries.

So far I could only find on a technical report from Yuasa that a similar battery, the LEV50, degrades to 90% after only 700 100% DoD cycles, where the wording might be unclear if this is related to a discharge under a 4C condition.

In general the degradation using a discharge from 100% down to 0% in comparison to 100% down to 30% is not going to be very different.
Generally for Li-Ion batteries you will get roughly 30% more of these 70% DoD discharge cycles than the full 100% but overall degradation will be roughly the same and certainly not 3x as many cycles.
Please provide evidence of that claim for the LEV40 or any other battery for that matter from their official manufacturers websites.

According to https://www.secondlife-evbatteries.com/pdf/Datasheet-LEV40-Module.pdf you can pick up the information from https://www.gs-yuasa.com/en/technic/archive.php.
If you are motivated enough to translate and search for it your self then you have what you are asking for.

That is exactly where I have researched as I mentioned above and find for the LEV50 only 700 cycles until 90%
The Archive goes only to 2006

Edit:
Found you a very good proof of cycle count when a cell is managed correctly.
https://en.m.wikipedia.org/wiki/Tesla_Powerwall.

.

from that link:

The 10 kWh battery as originally announced has a nickel-cobalt-aluminum cathode,[9] like the Tesla Model S,[10] which was projected to function as a backup/uninterruptible power supply, and had a projected cycle life of 1000–1500 cycles.[11][12][13][14]

Sounds more like it !!!

, where do you get your information from?

15 years servicing laptops and power tools with Li-Ion batteries

This is getting tedious.
First, you didn't read the Wiki. If you had you would have clearly seen this:

Secondly, you clearly did not see the specification for the LEV50 I attached.
It clearly says 5500 cycles for LEV50.

Also, with some searching I was able to find this.
http://news.yuasa.co.uk/wp-content/uploads/2016/09/Yuasa_Lithium_Power_Cabinets_Web.pdf
This shows even higher cycle count for these cells.
LIM50 (LEV50) does 11000
LIM25 does 20000
LIM40 (LEV40) should do something between if managed correctly.

As you might know. Mitsubishi does NOT purchase the cells from GS Yuasa directly.
Instead, the cells are re-branded and re-sold via a different company called Lithium Energy Japan, these are re-branded LIM X cells sold as LEV X cells, thus a specification directly from GS Yuasa is impossible.
This is probably done for several purposes, one of them is protection against potential legal action about issues like the one experienced on this vehicle.

These cells are VERY old, pre-dating 2007 and as proof of that see the below links.
https://web.archive.org/web/20071214090638/http://lithiumenergy.jp/en/products/index.html
https://www.lithiumenergy.jp/en/products/index.html

All of this supports what I am saying about the cause of this issue.
All of the specifications of the cells is supporting my claim that this is a Software or Hardware (electronic) issue.

I think that's about it with the rebuttals and I hope we can move on with the discussion about the issues from here now that the facts are clear.

 

[Pomst]

According to https://www.secondlife-evbatteries.com/pdf/Datasheet-LEV40-Module.pdf you can pick up the information from https://www.gs-yuasa.com/en/technic/archive.php.
If you are motivated enough to translate and search for it your self then you have what you are asking for.

I for one trust a company that puts their reputation and business at stake when providing that information.
If you don't trust them then that's a problem you need to work with and checking the technical archive is a method to do just that.

YUASA LEV40 Cell
Nominal voltage 3.65V
Operating voltage range 2.65V to 4.0V
1-hr rate typical Capacity 25C 40Ah

This is shocking to me

3,65v nominal voltage ? In Yuasa official pages the LEV50 has much higher nominal voltage

Here SecondLifeEVBattery suggest to use a much different voltage range :
Mitsubishi force an usage between 4.1v down to 3.8v
SecondLife suggest between 2.65v to 4.0v

Opposite point of view

Can be that a little company know better then Mitsubishi ?

This could be as easy as the use case for the cells.
Different use case require different specifications.

Batteries are not hardware, they are living chemically and depending on how you treat them they will act differently.
I think Mitsubishi is treating the batteries as good as they can so they will live as long as possible and provide Mitsubishi with great user stories, thus a specification for their specific use case.
SecondLifeEVBattery doesn't care since they only sell cells and can care less if you abuse them. They only care about the cells not catching fire and with those specs they wont.

The only way you will get a non-biased specification is to go directly to the producer and ask.
Which we cant, since the cells are re-branded and very old.

I have, however, found additional information on the cycle count and other data in the post above.

 

[mellwaters100]

Recently started looking into this topic and read this thread with interest - am going to run the procedure on mine in a couple of weeks. Current SoH = 32Ah / 84.2% @ 169,522 kms - 14 plate UK car.

I have watched a couple of the Unplugged EV videos (the one about the BMU writing values, etc) and, as an occasional programmer, makes sense. A number of things are going through my mind:

1) Looking at my 'dog', the battery SoH drops 0.1Ah approximately every 4,000kms, give or take a couple of hundred kms.
2) I have about a years worth of data in the 'dog' which shows that I am getting the same performance from the battery (@ 2km/Ah).
3) Currently, I am getting around 40km pure EV - GOM reports 49km when first switched on, although these current trips include approx 650m descent and 650m ascent (roundtrip!).
4) On the last measured trip (39km / 100% EV), cell voltage was 3.742 : 4.11.

All this suggests to me that Mitsubishi have programmed in a 'protection' process into the BMU, to reduce the capacity of the battery to protect it, rather than the reported SoH being a true reflection of the actual SoH. My main question with this is: if the BMU requires the 12v battery to maintain the 'learnt' SoH and, by disconnecting the 12v the BMU gets wiped, what stops / controls the charging process during this procedure?

Charging through a ChAdeMO, the car stops charging at 80%, yet you can reinitiate the charge and the charging rate drops. Isn't this the BMU controlling the rate of received charge? So, if it is wiped during this process, what is controlling the charging rate during this Lindqvist procedure?

Irrespective, it does seem as if taking the BMU out of the loop - and any artificial capacity value that it has stored - this procedure reverts to the battery hardware to determine it's true capacity.

Will certainly post my experience to try and help others (should be mid-August when I can get to do it).

 

[Pomst]

Recently started looking into this topic and read this thread with interest - am going to run the procedure on mine in a couple of weeks. Current SoH = 32Ah / 84.2% @ 169,522 kms - 14 plate UK car.

I have watched a couple of the Unplugged EV videos (the one about the BMU writing values, etc) and, as an occasional programmer, makes sense. A number of things are going through my mind:

1) Looking at my 'dog', the battery SoH drops 0.1Ah approximately every 4,000kms, give or take a couple of hundred kms.
2) I have about a years worth of data in the 'dog' which shows that I am getting the same performance from the battery (@ 2km/Ah).
3) Currently, I am getting around 40km pure EV - GOM reports 49km when first switched on, although these current trips include approx 650m descent and 650m ascent (roundtrip!).
4) On the last measured trip (39km / 100% EV), cell voltage was 3.742 : 4.11.

All this suggests to me that Mitsubishi have programmed in a 'protection' process into the BMU, to reduce the capacity of the battery to protect it, rather than the reported SoH being a true reflection of the actual SoH. My main question with this is: if the BMU requires the 12v battery to maintain the 'learnt' SoH and, by disconnecting the 12v the BMU gets wiped, what stops / controls the charging process during this procedure?

Charging through a ChAdeMO, the car stops charging at 80%, yet you can reinitiate the charge and the charging rate drops. Isn't this the BMU controlling the rate of received charge? So, if it is wiped during this process, what is controlling the charging rate during this Lindqvist procedure?

Irrespective, it does seem as if taking the BMU out of the loop - and any artificial capacity value that it has stored - this procedure reverts to the battery hardware to determine it's true capacity.

Will certainly post my experience to try and help others (should be mid-August when I can get to do it).

That's an astute observation.
This is my entire point.
Cutting the BMU out of the loop would not reset any kind of capacity loss what so ever if it was the Actual cells that were degraded.
Cutting the BMU out might actually allow for any bug (or feature or planned obsolescence) in the firmware to be bypassed and allow the cells to charge to full once again.

And yes, according to the Chademo standard the car tells the charger what amperage to allow.
What is doing the actual logic to determine what charge to allow I can only guess, but I suspect the BMU is telling something (probably temperature and SoC) to something that in turn is telling the charger.

If you are going to get this procedure done you should ultimately have a CANbus monitor installed to monitor and record all commands they send.
This could in turn be used to replicate the process.

 

[Darkflow]

So, if it is wiped during this process, what is controlling the charging rate during this Lindqvist procedure?

I suspect that disconnecting the 12V battery does not wipe the BMU values - that would actually be a pretty crappy design, as the 12V battery can go flat if the car is in storage for a long time.

The BMU does have power during charging with the 12V battery disconnected, via the 14V DC-DC converter. However, at the end of charging, the charger disconnects power, which powers down the converter. The BMU is still running up until this point and suddenly loses power. This could well lead to undefined states and corrupted data being stored - which could prompt a reset the next time the BMU is powered up.

 

[MadTechNutter]

Please provide proof for that claim of 5500 full cycles from the official website:
https://www.yuasabatteries.com/
and NOT some data sheet that somebody could had made up and is now circling on the internet, which I have seen already, possibly by somebody who is trying to sell used batteries.

This is getting tedious.
First, you didn't read the Wiki. If you had you would have clearly seen this:

Secondly, you clearly did not see the specification for the LEV50 I attached.
It clearly says 5500 cycles for LEV50.

Also, with some searching I was able to find this.
http://news.yuasa.co.uk/wp-content/uploads/2016/09/Yuasa_Lithium_Power_Cabinets_Web.pdf
This shows even higher cycle count for these cells.
LIM50 (LEV50) does 11000
LIM25 does 20000
LIM40 (LEV40) should do something between if managed correctly.

As you might know. Mitsubishi does NOT purchase the cells from GS Yuasa directly.
Instead, the cells are re-branded and re-sold via a different company called Lithium Energy Japan, these are re-branded LIM X cells sold as LEV X cells, thus a specification directly from GS Yuasa is impossible.
This is probably done for several purposes, one of them is protection against potential legal action about issues like the one experienced on this vehicle.

These cells are VERY old, pre-dating 2007 and as proof of that see the below links.
https://web.archive.org/web/20071214090638/http://lithiumenergy.jp/en/products/index.html
https://www.lithiumenergy.jp/en/products/index.html

All of this supports what I am saying about the cause of this issue.
All of the specifications of the cells is supporting my claim that this is a Software or Hardware (electronic) issue.

I think that's about it with the rebuttals and I hope we can move on with the discussion about the issues from here now that the facts are clear.

You are not reading my posts.
I said find please evidence from Yuasa the manufacturer, that the LEV40 has 5500 cycles!
Secondlife, is NOT the manufacturer and can make up anything they want.
They are selling USED batteries !!!

Then you go on about some newer batteries that are not used in the PHEV, although I did initially ask for it.
As I said and you confirmed, the LEV40 is very old. There is no indication from the official sites that it can have 5500 cycles.

Claiming 5500 cycles is NOT a clear fact here!

It can only be a clear fact if Yuasa, the distributer or Mitsubishi themselves claim that.

A company with the magnitude of Mitsubishi allowing to display false OBD-II data, in this case the battery cell voltages, seems very questionable.
I have to assume the Watchdog software is displaying the actual data it is receiving because that corresponds with the cars performance.

What ever these magical resets are doing, nobody has reported that they last very long.

I will of course continue to try because my battery seems to be toast at 75% for 60,000km (1500cycles if pure EV) anyhow.

EDit: ...and why this is so important that I am not leaving this miss-information persist ?
If I were to go to a Lawyer with legal evidence that the batteries have a 5500 cycle life down to 80%, I may be able to force Mitsubishi to replace it under Statuary Warranty conditions here.
Currently there seems to be no replacement warranty in Australia unless individual cells have completely failed and even this warranty expires after 5 years.

 

[Pomst]

You are not reading my posts.
I said find please evidence from Yuasa the manufacturer, that the LEV40 has 5500 cycles!
Secondlife, is NOT the manufacturer and can make up anything they want.
They are selling USED batteries !!!

Then you go on about some newer batteries that are not used in the PHEV, although I did initially ask for it.
As I said and you confirmed, the LEV40 is very old. There is no indication from the official sites that it can have 5500 cycles.

Claiming 5500 cycles is NOT a clear fact here!

It can only be a clear fact if Yuasa, the distributer or Mitsubishi themselves claim that.

A company with the magnitude of Mitsubishi allowing to display false OBD-II data, in this case the battery cell voltages, seems very questionable.
I have to assume the Watchdog software is displaying the actual data it is receiving because that corresponds with the cars performance.

What ever these magical resets are doing, nobody has reported that they last very long.

I will of course continue to try because my battery seems to be toast at 75% for 60,000km (1500cycles if pure EV) anyhow.

EDit: ...and why this is so important that I am not leaving this miss-information persist ?
If I were to go to a Lawyer with legal evidence that the batteries have a 5500 cycle life down to 80%, I may be able to force Mitsubishi to replace it under Statuary Warranty conditions here.
Currently there seems to be no replacement warranty in Australia unless individual cells have completely failed and even this warranty expires after 5 years.

You will find it harder then a Google search or a forum whine to get specifications for cells older then 2007.
What I found is what is available from 5 minutes of searching to prov a point about cycles on batteries and my calculation on how long these cells Should last.
If its not acceptable to get you to understand the potential cycles of these cells go get more information your self.
Don't whine about the Free information provided gathered on my spare time to help people figure the issue out.

Regards misinformation or false data.
VW at 330% larger then the entirety of Mitsubishi did it.
Make cells degrade to sell more vehicles or packs over a 10 year period would definitively be something up that street.

Regarding GS Yuasa and Mitsubishi, you need to get your facts right.
https://www.lithiumenergy.jp/en/corporate/index.html
https://www.mitsubishicorp.com/jp/en/pr/archive/2007/files/0000002421_file1.pdf
Mitsubishi gets the cells delivered from Lithium Energy Japan.
Asking for information that does not exist is counter productive.
If Lithium Energy Japan sold the cell then GS Yuasa will not have a specification for it.
They will, however, have a specification for the original cells name which is LIM40 not LEV40. First produces some where pre 2007.

Like I said, this is probably a move to save their ass if some litigation come up, it is common practice to have a bouncer company.

 

[greendwarf]

Edit:
Found you a very good proof of cycle count when a cell is managed correctly.
https://en.m.wikipedia.org/wiki/Tesla_Powerwall. If that's not enough for you then I'm not sure anything will be.

Unfortunately - quite the opposite of "proof". Wikipedia articles are not allowed as "evidence" in most legal jurisdictions, as they can be easily edited to say what ever you want them to. The only time they should be used is as a research tool is to find the authoritative sources given in the references. :lol:

 

[Pomst]

Edit:
Found you a very good proof of cycle count when a cell is managed correctly.
https://en.m.wikipedia.org/wiki/Tesla_Powerwall. If that's not enough for you then I'm not sure anything will be.

Unfortunately - quite the opposite of "proof". Wikipedia articles are not allowed as "evidence" in most legal jurisdictions, as they can be easily edited to say what ever you want them to. The only time they should be used is as a research tool is to find the authoritative sources given in the references. :lol:

That wasn't the question.

The question was:

Please provide evidence of that claim for the LEV40 or any other battery for that matter from their official manufacturers websites.

Which was provided.

 

[greendwarf]

Edit:
Found you a very good proof of cycle count when a cell is managed correctly.
https://en.m.wikipedia.org/wiki/Tesla_Powerwall. If that's not enough for you then I'm not sure anything will be.

Unfortunately - quite the opposite of "proof". Wikipedia articles are not allowed as "evidence" in most legal jurisdictions, as they can be easily edited to say what ever you want them to. The only time they should be used is as a research tool is to find the authoritative sources given in the references. :lol:

That wasn't the question.

The question was:

Please provide evidence of that claim for the LEV40 or any other battery for that matter from their official manufacturers websites.

Which was provided.

But that was YOUR answer - "a very good proof" - which it is not. You then compound the error by quoting directly from it rather than an authoritative source. I'm not saying its not correct, I'm merely pointing out that wiki articles are well known to be subject to falsification and therefore can never be cited as "proof". :idea:

 

[Pomst]

Unfortunately - quite the opposite of "proof". Wikipedia articles are not allowed as "evidence" in most legal jurisdictions, as they can be easily edited to say what ever you want them to. The only time they should be used is as a research tool is to find the authoritative sources given in the references. :lol:

That wasn't the question.

The question was:

Please provide evidence of that claim for the LEV40 or any other battery for that matter from their official manufacturers websites.

Which was provided.

But that was YOUR answer - "a very good proof" - which it is not. You then compound the error by quoting directly from it rather than an authoritative source. I'm not saying its not correct, I'm merely pointing out that wiki articles are well known to be subject to falsification and therefore can never be cited as "proof". :idea:

Yes, you are correct.
The wiki was the best effort in a 10 second search for "any other battery".
For the amount of time it took me it was "a very good proof". That doesn't mean its the best proof and I was not claiming it was the best either.

I did, however, provide other documentation as proof of cycle count as well.
Of which you seem to not have seen.

This document clearly says that similar cells from Yuasa can, according to them, do many thousands of cycles (11000 for 50Ah cell to be exact).
http://news.yuasa.co.uk/wp-content/uploads/2016/09/Yuasa_Lithium_Power_Cabinets_Web.pdf
This backs my claim of these cell probably lasting for vastly more then 5500 cycles if managed correctly.
Thus supporting the viability of the Lindqvist method as an intermediate solution until something better comes along.

 

[elm70]

Guys ...

Let's go back to the main topic ...

I'm still trying to reproduce this Lindqvist method on my PHEV

Even trying right now .. this time I will have the 12v battery disconnected for over 24h before start the charging process

Not sure it will make any difference ... but I don't have to use my PHEV today

It will be nice to know how much success we can have in this procedure .. based on a similar topic in FB, the success rate is around 30/40%

More then the success rate, it will be nice to understand which is the critical part of the process that cause this magic reset ... and for this the experience of who did initially fail and then made it working it is very important.