DBMandrake
Posts: 275
Joined: Mon Mar 20, 2017 4:57 am
Location: Scotland

Re: Fuel gauge vs tortoise

Sat Feb 22, 2020 12:35 pm

I obtained my 4 second hand cells from the following UK company:

https://www.secondlife-evbatteries.com/ ... lev50.html

All of them were about 43Ah when discharge tested from 4.1 volts down to 3.0 volts, so better than even the best cells in my pack.

I see they are currently out of stock however I think they get stock in regularly (from scrapped cars) so it may be worth contacting them. Not sure where you are located but I suspect they may also ship to Europe.

There is also a guy in the Netherlands who strips and sells LEV50 cells (and other types) I can find his contact details if needed.
- Simon

EV: 2011 Peugeot Ion
ICE: 1997 Citroen Xantia V6

vh2q
Posts: 37
Joined: Fri Dec 18, 2015 12:39 pm

Re: Fuel gauge vs tortoise

Sat Feb 22, 2020 3:13 pm

Car is in Bahamas .. still don't get why one weak cell will suddenly prevent pack from charging more than 25% though.

DBMandrake
Posts: 275
Joined: Mon Mar 20, 2017 4:57 am
Location: Scotland

Re: Fuel gauge vs tortoise

Sat Feb 22, 2020 3:34 pm

vh2q wrote:Car is in Bahamas .. still don't get why one weak cell will suddenly prevent pack from charging more than 25% though.

Because all the cells are in series, so the usable capacity of the pack as a whole is limited to that of the weakest cell.

Weakest link in the chain. It only takes one faulty cell to decimate the usable capacity and range of the pack.
- Simon

EV: 2011 Peugeot Ion
ICE: 1997 Citroen Xantia V6

vh2q
Posts: 37
Joined: Fri Dec 18, 2015 12:39 pm

Re: Fuel gauge vs tortoise

Sun Feb 23, 2020 8:33 am

So the problem is the weak cell gets down to 3V pretty quick and the computer then shuts down the car to protect it? Can one short out the bad cell?

DBMandrake
Posts: 275
Joined: Mon Mar 20, 2017 4:57 am
Location: Scotland

Re: Fuel gauge vs tortoise

Sun Feb 23, 2020 9:28 am

vh2q wrote:So the problem is the weak cell gets down to 3V pretty quick and the computer then shuts down the car to protect it?

Essentially yes. Although it won't let it go down as low as 3.0 volts. It will only go down to about 3.6 volts before turtle mode will come on.
Can one short out the bad cell?

Definitely not. (I presume you mean remove and bypass, not literally short it out. :shock: )

The Battery management system monitors every cell individually if there is a problem with any cell the car will not work properly.

A cell swap with a good cell would be the only option.
- Simon

EV: 2011 Peugeot Ion
ICE: 1997 Citroen Xantia V6

vh2q
Posts: 37
Joined: Fri Dec 18, 2015 12:39 pm

Re: Fuel gauge vs tortoise

Sun Feb 23, 2020 11:14 am

I didn't mention it earlier but I got these two errors when I first scanned car:
P1AE6 EEPROM error
P1A51 EEPROM abnormal
I cleared them and they did not come back. The other two errors cannot be cleared
P1A4B cell voltage abnormal
P1A2F cell error

And this all happened after aux battery was out of car for some months

kiev
Posts: 1122
Joined: Sun May 03, 2015 7:15 am
Location: The Heart o' Dixie
Contact: Website

Re: Fuel gauge vs tortoise

Sun Feb 23, 2020 6:19 pm

It's unfortunate that you had a cell go bad,

the first two DTC about the eeprom are likely due to the lack of a 12V aux battery for a time, and the ECU was confused at first when coming back up with power.

The last two are due to the failed cell.
kiev = kenny's innovative electric vehicle

vh2q
Posts: 37
Joined: Fri Dec 18, 2015 12:39 pm

Re: Fuel gauge vs tortoise

Tue Feb 25, 2020 12:16 pm

For what it's worth, I ran battery down to 0 bars and when the turtle came on the bad cell was down to 3.48V, the rest were still at 4.0V. So it seems that the bmu software sets the charge gauge based on the weakest cell, but controls the charger based on the strongest cell.

After two charging cycles, weakest cell still does not get above 3.9V when the rest are at 4.1

Hobdrive was useful in diagnosing this situation, but there are an awful lot of meaningless screens and data. A lot of the fields come up as zero and you can't even get a current draw off the aux or the dc to dc converter. You can get a draw off the MCU. Many of the temps default to 32F. You can turn off the trouble codes but the car will regenerate them so the serious ones just pop up again. The cell smoothing function does not work. There are a bunch of screens that display CMU voltage. These appear to be clusters of 8 battery cells. This data is pretty useless.

And finally, a battery pack that is only as good as its weakest cell is an abomination. Especially when, all too often it seems, one cell goes bad.

DBMandrake
Posts: 275
Joined: Mon Mar 20, 2017 4:57 am
Location: Scotland

Re: Fuel gauge vs tortoise

Wed Feb 26, 2020 2:59 am

vh2q wrote:For what it's worth, I ran battery down to 0 bars and when the turtle came on the bad cell was down to 3.48V, the rest were still at 4.0V. So it seems that the bmu software sets the charge gauge based on the weakest cell, but controls the charger based on the strongest cell.

This is essentially correct. The BMU will not allow any cell to charge above 4.1 volts or below a minimum limit which seems to be a bit more flexible - typically between 3.3 volts and 3.6 volts.

When there is a weak cell these limits will be reached much sooner and the effective usable capacity is reduced.
After two charging cycles, weakest cell still does not get above 3.9V when the rest are at 4.1

Two charging cycles is not enough to do anything of significance. It took nearly 40 cycles to correct a 50mV error on my car.
Hobdrive was useful in diagnosing this situation, but there are an awful lot of meaningless screens and data. A lot of the fields come up as zero and you can't even get a current draw off the aux or the dc to dc converter. You can get a draw off the MCU. Many of the temps default to 32F. You can turn off the trouble codes but the car will regenerate them so the serious ones just pop up again. The cell smoothing function does not work. There are a bunch of screens that display CMU voltage. These appear to be clusters of 8 battery cells. This data is pretty useless.

The user interface of Hobdrive is pretty poor, I agree. However it can only provide information that the ECU's in the car make available. Under many conditions some of those fields are not applicable.

Resetting a fault code is not a magic cure - if the fault returns its because the fault is still there! Not the problem of the diagnostic tool.

The cell smoothing process is just a diagnostic process that can be initiated in the BMU ECU - I have an official Lexia 3 diagnostic tool for my Ion (the same tool used by dealers) and the cell smoothing function in that also did not work for me.

For reading cell voltages the best tool to use by far is Canion. If you have an STN11xx based adaptor like an OBDLink LX you can use Canion. Canion is read only and cannot reset codes or make any changes but I found it the best tool to diagnose weak cells on my car.
And finally, a battery pack that is only as good as its weakest cell is an abomination. Especially when, all too often it seems, one cell goes bad.

Unfortunately that's just the way it is - with nearly all EV's! Most EV's use either a single series string or more commonly a xxs2p configuration where cells are connected in parallel in pairs then the pairs are connected in series. The Leaf uses this paired approach.

However even in a pared approach there will be a large drop in capacity of the pair of cells if one cells go bad and the same issue will occur. Plenty of Leaf's have had cell failures as well.

About the only cars not affected by individual cell failures are Tesla's, which use massively parallel groups of small cells which are independently fused. If one cell in a parallel cluster of >20 cells goes faulty the capacity loss is minimal and the usable capacity of the pack will only be affected a small amount. However with a large number of cells the likelyhood of failure of a given cell is higher of course.

At least in the i-Miev a faulty cell can be replaced somewhat easily - I've done it on mine. The hardest part is getting the pack out of the car and back in again. Once the pack is down on a trolley the cell swap is relatively easy if you are an experienced DIY'er and have some electrical knowledge. Cells and modules in a Tesla pack are not DIY serviceable.
- Simon

EV: 2011 Peugeot Ion
ICE: 1997 Citroen Xantia V6

vh2q
Posts: 37
Joined: Fri Dec 18, 2015 12:39 pm

Re: Fuel gauge vs tortoise

Wed Feb 26, 2020 7:52 am

That is helpful thanks.

What tools do I need to pull the pack? Am on the proverbial desert island here. Instructions?

Return to “Batteries and Battery Management”