Precautions to take with LiFePO4 batteries

karrak · April 2013

I have just installed a bank of 8x4 90AH Winston LIFePO4 cells to replace my aging NICd batteries in my 24 volt off grid system.

As discussed in much literature and on many forum discussions, the one big flaw with the LiFePO4 batteries is they do not take kindly to being overchaged or over discharged. With Lead Acid batteries or NiCds the occassional overcharge or over discharge is not likely to severly damage or destroy them, which does not seem to be the case with LiFePO4 cells.

IMHO one is taking a big risk by just letting the solar controller and inverter monitor the whole battery string voltage and use this to shut down charging or disconnect the load for any system with a voltage over 12 volts.

If we take a 24 volt string (8x3.2 volts, nominal voltage 25.6v) and set our solar controller to stop charging at 3.5v per cell (28 volts for 8 cell string), if one cell in the string has probably only a few percent more charge we could get a situation where we have 7 cells at 3.3 volts and the overchaged cell being at 4.9v (7*3.3+4.9=28)! This would not have been a problem for large enough Lead acid batteries or NiCds as they whould have gassed and balanced out, but is a big problem for LIFePO4 batteries.

On the other end of the scale the maximum voltage my inverter will disconnect the load at is 23 volts, (2.875 volts per cell) which sounds resonable but we could get a situation where we have 7 cells at 3.1 volts with the remaining one at a disasterous 1.3 volts(7*3.1+1.3=23).

There is also the problem of a fault in the solar controller causing it to continue charging the cells even after it has reached a safe cutoff voltage, a BMS system that only does cell balancing will not help in this case. Again not much of a problem for large enough lead acid or NiCd batteries.

To protect the batteries from overcharging and over discharging i am using two Cell Logger 8s to monitor the individual cells. Not ony does this display the individual cell voltages but one will provide a hi voltage level alarm and the other a low voltage alarm. The high voltage alarm will be used to toggle a latching relay to disconnect the solar panels from the solar controller and the low level alarm will be used to disconnect the load by simulating an "earth fault" on the AC side of the inverter and trip the RCD (Residual Current Detector) breaker.

Having said the overcharging and over discharging is a serious issue with the LiFEPO4, i think this fairly simple and cheap option i have taken should make it very unlikely that this problem will occur.

Any comments or suggestions would be appreciated

Simon

BB. · April 2013

Re: Precautions to take with LiFePO4 batteries

Simon,

I agree with you--Per Cell monitoring/logging/alarming/independent cutoff is really important. I think you are taking seriously both the safety and potential damage to a very expensive battery bank.

I would also suggest for >48 volt battery banks, that per cell (or per group of 3-6 cells) is something that may be needed too... A 2 volt cell failure out of 12 volts is obvious. A 2 volt cell failure out of 48 volts is barely detectable by somebody paying attention with standard banked monitoring.

-Bill

karrak · April 2013

Re: Precautions to take with LiFePO4 batteries

The precautions have paid off! When i went to bed last night there was about 100mV difference between one block of cells and the others. I was awakened at 4:15am by the cell logger 8 alarm. On inspection i found that seven of the 8 blocks of cells were around 3.16 volts while one block was at 2.8 volts. My controller has logged the overall battery voltage at that time as 24.8 volts, about 1.8 volts higher than the inverter cutoff voltage. So the cell logger 8 saved the day and maybe four batteries.

On further investigation i don't think the cells were all full and balanced when i assembled the battery pack. I paid the battery supplier extra to balance the cells but i think one must have been missed. Unfortunately I didn't check the cell voltages before i put the battery pack together as i assumed they had been checked by the battery supplier, mind you i might have missed the problem as i am new to these batteries and the difference between these low cells when they were assembled into the pack and the rest was only 2mV when they were in the middle of thier charge range.

So ironically i have had to charge and balance these low cells using the old dying NiCd batteries that they replaced. The block of four cells has take about 25ah to bring it in line with the other cells. Lets hope for some sun tomorrow!

Simon

westbranch · April 2013

Re: Precautions to take with LiFePO4 batteries

I hope you will not just retire those NiCd's, they usually can be revived with electrolyte replacement, if you can empty them easily.

hth

mike95490 · April 2013

Re: Precautions to take with LiFePO4 batteries

I guess you need to make up a BMS system. The banks will never stay balanced, over time as cells age, they will begin to vary, and you will have to switch out a bank, and individually charge the low cells. Or get a set of Battery Balancers to do it automatically. At least you are monitoring them, and can catch them before they go boom.

john p · April 2013

Re: Precautions to take with LiFePO4 batteries

A home made BMS is very easy to implement if you only use the 10a cells. It gets a lot harder if you use the100a or 200a cell packs.
I have now made up many small systems using the 10a cells and some now been in use about 5 yrs. Never a problem.

Other alternate is to simply buy a battery pack in the voltage you need 24,48,144 etc. This way the BMSis built in . very easy to use ,with no balancing needed if parallel connecting.

karrak · April 2013

Re: Precautions to take with LiFePO4 batteries

westbranch wrote: »

they usually can be revived with electrolyte replacement,

hth

Unfortunately we live at the end of the universe, nobody locally stocks the LiOH in the grade i want, and as it is a HAZMAT it costs more to ship it than its value!

On top of the cost of the replacement electrolyte, most of the cells i have are in blocks of 6 cells, each block weighing around 100kg, with maybe one dodgy cell in each block. Then there is the old electrolyte to be got rid of.

I also have the money to replace them, and from an operational point of view on paper the LiFePO4 are so superior to lead acid or NiCd batteries. Only time will tell if this is indeed the case.

Simon

westbranch · April 2013

Re: Precautions to take with LiFePO4 batteries

Hmmm,, LiOH... Mine are SAFT and use KOH which is readily available, too bad as used KOH is also easily disposed of, when carefully handled. That weight is a problem too.
What brand are yours?

karrak · April 2013

Re: Precautions to take with LiFePO4 batteries

mike95490 wrote: »

I guess you need to make up a BMS system. The banks will never stay balanced

I would have agreed with you except that the crowd where i got the batteries from, who have been converting cars to Electric Vehicles for over five years have decided it is not worth doing automatic battery balancing. I will keep an eye on the batteries and design something up if need be, it will be an adjunct to my solar controller, i think it is better to have the charging process control the cell balancing.

john p wrote: »

A home made BMS is very easy to implement if you only use the 10a cells. It gets a lot harder if you use the100a or 200a cell packs.
I have now made up many small systems using the 10a cells and some now been in use about 5 yrs. Never a problem.

True, the % of charge that can be easily dumped by the cell balancing circuit gets less with larger cells. I am concerned that unless the BMS has some control over the charging process you can end up with a situation where the cell balancing circuit cannot dump enough current from the charger to save a cell that is badly out of balance. Does your BMS have an alarm that will shut down the charging if an individual cells goes overvoltage or disconnect the load if a cell goes undervoltage?

john p wrote: »

Other alternate is to simply buy a battery pack in the voltage you need 24,48,144 etc. This way the BMSis built in . very easy to use ,with no balancing needed if parallel connecting.

From my experience over many years with lead acid and NiCd batteries it is often one, or a couple of cells in a battery stack that will die and kill the whole battery, even though there might be quite a bit of life left in the other cells. If you have individual cells you can just swap out the dead one. If you have banks with at least two cells in parallel you can, as they get near the end of their life swap cells around to get the most life out of them.

I certainly wouldn't use multiple cell in series LiFePO4 battery packs if they do not have an alarm output.

Simon

karrak · April 2013

Re: Precautions to take with LiFePO4 batteries

westbranch wrote: »

Hmmm,, LiOH... Mine are SAFT and use KOH which is readily available, too bad as used KOH is also easily disposed of, when carefully handled. That weight is a problem too.
What brand are yours?

The bulk of the electrolyte is KOH but NiCd batteries need some LIOH as well to stop erosion of either the cathode or anode, i can't remember which. If you want more details there is a very good book called "Alkaline Storage Batteries" by Falk & Salkind, SBN 471 25362 6, now out of print but available from second hand book dealers. Also some very good articles in Home Power Magazine" from the 1990s, this one in particular Home Power Magazine - Issue 015 - 1990-02-03.

The OH can be neutralised with sulphuric acid, the thing that concerns me is how much cadmium there is in the old electrolyte, i am sure there would be some, how much i don't know.

I have a whole lot of 6 cell NiCds made by NiFe who were bought out by SAFT and 10 dual cell SAFT batteries. If anyone in SW Australia is interested in them let me know.

Simon

Precautions to take with LiFePO4 batteries

Comments

Categories