Is a BMS actually necessary on LifePo4 cells.

Chat about all things battery in here.
Need to know what type to use or size or capacity then again place your thoughts here
User avatar
Rory166
Posts: 116
Joined: Sat May 17, 2014 7:45 am
Location: Norwich, UK

Is a BMS actually necessary on LifePo4 cells.

Postby Rory166 » Sat Jul 19, 2014 4:58 pm

Hi All

Yesterday I was shown a vehicle with loads of Lifepo4 cells and no BMS. Apparently all the cells were bottom balanced on install and then pack charged until the lowest capacity cell reached 3.7v at which point the pack voltage was noted and used as a cut-off value for the charger.

Clearly while the cells remain in good condition all is well.

This does make me wonder if the typical top balancing BMS is doing perhaps more harm than good. I do think that any system which is going to be used by other than the builder needs to have some method of detecting a cell which is going into excessive low voltage. Also excessive high voltage I assume.

I do wonder if the mantra that lithium cells always need a BMS was conceived for a different chemistry from LiFePo4? The concept that over time some cells will lose more charge through self heating than others and re-balancing is required, does not seem to be valid, at least not with new cells.

Rory
Electric Seicento conversion, Leaf

User avatar
skooler
Posts: 129
Joined: Fri Oct 19, 2012 5:19 pm
Location: Worcestershire
Contact:

Re: Is a BMS actually necessary on LifePo4 cells.

Postby skooler » Sun Jul 20, 2014 1:45 pm

Rory166 wrote:Hi All

Yesterday I was shown a vehicle with loads of Lifepo4 cells and no BMS. Apparently all the cells were bottom balanced on install and then pack charged until the lowest capacity cell reached 3.7v at which point the pack voltage was noted and used as a cut-off value for the charger.

Clearly while the cells remain in good condition all is well.

This does make me wonder if the typical top balancing BMS is doing perhaps more harm than good. I do think that any system which is going to be used by other than the builder needs to have some method of detecting a cell which is going into excessive low voltage. Also excessive high voltage I assume.

I do wonder if the mantra that lithium cells always need a BMS was conceived for a different chemistry from LiFePo4? The concept that over time some cells will lose more charge through self heating than others and re-balancing is required, does not seem to be valid, at least not with new cells.

Rory

As background to everone else, it was my RX8 that Rory saw.

To be clear, I charge until the first cell hit 3.5v, not 3.7.

I wrote a lot on the subject of bottom balancing over on DIY electric car. the thread is probably worth a read :)

http://www.diyelectriccar.com/forums/sh ... hp?t=85458

I now have 10,000 miles on the 24kwh pack in the RX8. all cells are still i balance and the weakest cell as identified using the process on the link is still the weakest cell.
Posts by Mike Schooling
Director and tech lead at indra.co.uk

Mazda RX8, Soliton Jr, 24KWH Sinopoly LiFePO4, Kostov 11alpha
http://www.diyelectriccar.com/forums/sh ... 61556.html

Grumpy-b
Posts: 991
Joined: Sat Dec 01, 2007 12:06 pm

Re: Is a BMS actually necessary on LifePo4 cells.

Postby Grumpy-b » Sun Jul 20, 2014 6:13 pm

It depends upon what you think a BMS is.
Battery Monitoring System = A system that merely monitors the cells in the battery pack, it could give some form of warning when a cell or the pack goes outside certain parameters. And may record data. This can be useful in problem solving.
Battery Management system= A system that attempts to manage the charge in the cells. This can only realistically be done on voltage and only at upper voltage . It could possibly control the charger rate of charge as well.

Lifepo4 cells do have a degree of differing finish charge voltage both between different makes with slightly different cell construction, and between cells. Ie calb cells tend to be full at around 3.45v but a Thundersky will be pretty much full at 3.6v. Charging above this voltage tends to put only a small amount of extra AH into the cell. In a Thundersky being charged at 20amp it will go from around 3.6v to around 4v in a matter of minutes. But can readily go way beyond that voltage very quickly.

A BMS actively managing a pack can readily destroy it. Poor cells not holding as much stored energy tend to go high voltage early. To which a top balance bleeding syetem will then try to bleed that cell down to the rest of the pack. If the bleed circuit is high enough rated to be of any real use this will allow the rest of the pack to continue to charge. When finished they will all then be bled down to a similar cell volatge. When they are discharged the low capacity cell will then go low voltage early , as it doesnt hold as much charge as the rest of the pack. Result is that cell could go negative, in being discharged down below say about 2.5v under load the cell will bloat, (I have one that nearly 40mm wider than it started out at 70mm wide) This can crush the cells besides, or worse still the cells can get hot, and melt the case and its neighbour.
If the BMS (Top balance) malfunctions, reads voltages incorrectly etc then it can bleed down a perfectly good cell. On a recent car that had been in an accident, one Cell BMS lead had broken, the BMS took that as showing the rest of the pack was high voltage, so it bled the rest of the pack down, taking out a few cells on the way. Others I have seen have had cells taken down because they read too high (they wernt high voltage just the BMS was wrong) so it would bleed good cells down, result when the pack was in use and fairly well depleated these cells were destroyed by over discharge.
My own Smart has a set of older Blue Calb 180ah that were bottom balanced to 2.7v below which they hold very little AH, they were all charged at the same rate for the same time, and all came to around the same total AH capacity together, reaching a finish voltage or around 3.6v per cell. They all are balanced and with no unbalancing extra connections in the pack they all discharge at the same rate and all charge at the same rate. SO until a cell starts to fail, all will be OK.
MODEC used a Lifepo4 in some of the packs, and whilst it had a million miles of small cable in it, it only monitored the cells. Extreme effort was made in the building of the pack to very closely bottom balance the pack. It did not top balance.
An exreme attempt to top balance is taken by EVC in Guernsey, who charged the pack as a whole, then used a simple plate with studs in a circle with stepper motor to connect to any cell and top it up. They claimed to be able to top up on the move so being able to support a low cell. Others dont agree that it worked as well as claimed.
Bottom balancing takes a lot of time. You have to take the cells below the target voltage, let them recover and take them down nearer to the target volatge and so on getting closer and closer. When they are stable I mark the voltages, and leave them for a number of days, if they self discharge they get put aside if the bounce up they get taken down. Putting them in a big parallel string works OK to get them all to the same volatge, but you will miss any that self discharge.
Charging you need an accurate charger (CC Amps) and plenty of patience. The result can be a really well balanced pack. But its important to have the final charge voltage correct. On Thunderskys there is absolutely no point in taking them above 3.8v As you will add very little range to the pack , but beyond that will take them to a region where they will start to diverge and may go high voltage. These cells will take the bulk of their charge below around 3.4to 3.5v and will then start to rise rapidly.
As with all of the EV issues its a matter of compromise, and what you wish to achieve.

Grumpy-b

User avatar
skooler
Posts: 129
Joined: Fri Oct 19, 2012 5:19 pm
Location: Worcestershire
Contact:

Re: Is a BMS actually necessary on LifePo4 cells.

Postby skooler » Sun Jul 20, 2014 7:34 pm

Grumpy, good explanation. I agree with everything you say apart from the cut off voltages but that's down to your own risk appetite :)

Thought I would add a bit more on the topic of battery monitoring systems.

Typically, this will be a method of sensing voltage attached to each cell. Usually a cable from a terminal on each single cell which goes off to a central location.

The single biggest issue with these is that the monitoring itself is a parasitic load on each cell and unless all cells are loaded 100% equally then there will be a drift in balance over time. I don't know of any voltage sensing method which can guarantee to load the cell exactly the same as an identical device.

It's also another point of failure.

10,000 miles on mine and I have NEVER heard of a bottom balanced LiFePO4 pack which has gone bad.
Posts by Mike Schooling
Director and tech lead at indra.co.uk

Mazda RX8, Soliton Jr, 24KWH Sinopoly LiFePO4, Kostov 11alpha
http://www.diyelectriccar.com/forums/sh ... 61556.html

Grumpy-b
Posts: 991
Joined: Sat Dec 01, 2007 12:06 pm

Re: Is a BMS actually necessary on LifePo4 cells.

Postby Grumpy-b » Mon Jul 21, 2014 9:06 am

The parasitic loards really arnt an issue on an externally powered device, such as most cars use. If its a small self powering pack such as a bike, then it can be a real issue owing to imbalance in the way the cells can be loaded. Same issue if you have a number of separate BMS units to cover the pack, they may work in slightly different ways.
Conceptually my preference is for an imbalance monitoring system, which gives a warning if either half of the pack is lower/higher in voltage to the other (by a given margin) this at least tells you you have a problem, which can then be investigated in more depth in the workshop.

The BMS in the C1 EVIe, do balance, but at only a couple of amps, so cant stop a wayward cell going wayward, they do go wrong, the small SMD devices can both cause the discharge resistors to stay bleeding and not bleed at all. All a pain but they do hold data, but once the data store is full it holds no more until its cleared. But having been cleared its usefull to be able to download and look at the data on charge and discharge.

Most important is to have a decent charger with good voltage control, and a way of being turned off if an external monitoring system gives an indication of a problem. The same for discharge, but thats a bit more complicated. The old CALB cells would keep giving their all below 2.7v for some time, but a thundersky at 2.7v under load will go downhill fast. SO its really important to understand what the cells can do, how they perform and utilise some system that recognises their issues.

Simple monitoring of AH in and AH out is a good alternative, but many devices (Hall especially ) dont perform as well in each direction. I believe the Leaf has such an issue which gave rise to the thought that the pack was going bad early, but it appears it was more associated to a drift in the Hall sensor accuracy as temperature rises so it appeared to have taken out more from the pack than was going in,so they must be losing capacity.

On the difference between the cells, an example if you charge Thunderskys at 20amp they will start to rise very rapidly above 3.6v, but take that down to 10amps or less and they will sit there for ages, take it up to 40amp and they would shoot up in voltage from around 3.5v. So simple voltage based control isnt simple.

This is one of the reasons why fast chargers dont go above 80% capacity, as the cells can get a bit unpredictable in their rate of voltage rise once they get into the 80 to 90% full range at high charge rates.
This gets a bit more unpredictable with increasing age of the cells, environmental temp, cell quality etc etc etc.

Long debate and plenty of alternatives, especially if you are in the business of making and selling BMS devices.

Grumpy-b

User avatar
Rory166
Posts: 116
Joined: Sat May 17, 2014 7:45 am
Location: Norwich, UK

Re: Is a BMS actually necessary on LifePo4 cells.

Postby Rory166 » Thu Jul 24, 2014 9:24 pm

Interesting about Hall effect devices. Where I used to work we sold a clamp on Hall device for 300 quid which was quite good beacause it had two Hall devices back to back thus eliminating the +/- non linearity problem. I am not sure if any of the cheap ones have this feature but one could use two ferrites back to back and invert one signal to achieve a similar result. I am not sure if the back to back devices help to solve the temperature drift issue.

Rory
Electric Seicento conversion, Leaf

User avatar
Rory166
Posts: 116
Joined: Sat May 17, 2014 7:45 am
Location: Norwich, UK

Re: Is a BMS actually necessary on LifePo4 cells.

Postby Rory166 » Sat Sep 13, 2014 8:11 pm

Still thinking about this. It seems to me that the self discharge of Lithium cells whilst low will not necessarily be identical between cells. This over a long period of time will lead to imbalance. If the cells are very well made to tight tolerances then it may be possible for a pack to retain balance over a good period of time. CAL-B seem quite good in this regard. For those who are not intending to regularly check their cell voltages then a BMS is a necessity. Unfortunately the BMS itself may cause imbalance by not loading all cells identically, thus making work for itself. On the matter of cell voltages it is really only meaningful as an indicator of state of charge when in the low range near to discharge or the high range near full charge so checking needs to be done at one of these points.

There are some small packs sold which have closely matched cells and no BMS. How this can be a practical manufacturing technique I do not know. To measure rate of discharge surely means leaving cells with an accurate state of charge for a longish period then measuring accurately how much energy is required to re-charge to 100%. This seems an impractically long process for a manufacturer. Added to which cells may change their character over time.

My proposed method for running a conversion car is as follows. Use a top balancing BMS. Run the vehicle, using an energy monitor, until the lowest capacity cell reaches a pr-determined low voltage. Record the energy usage and reset the energy meter to zero, recharge the battery and use the energy monitor as "fuel gauge". Apparently lithium battery energy retention is close to 100% making this practical.

Rory
Electric Seicento conversion, Leaf


Return to “All things battery related”

Who is online

Users browsing this forum: No registered users and 42 guests