Simple LiFePO4 battery bank for DIY?

DaveB

There is a lot of information about LiFePO4 battery banks spread out over numerous threads and websites.  I am trying to consolidate this into simple list of how to get the bank and equipment needed from a DIY perspective.  For the non-DIY person the easiest way to get a LiFePO4 is to simply purchase the Battle Born or SimplePhi batteries on NAWS. 

For a DIY person in the USA who wants to gather a list of "parts" and put it together, the questions and understandings I have are:
- Where are people getting new 3.2V LiFePO4 cells from?  My understanding is I should be looking for 1C or less cells.  Are some brands preferable then others for a renewable energy environment?

- Does it matter if the bank has aluminum strapping around it that I've seen on some banks that is supposed to prevent deformation if the battery gets hot?  I assumed in a RE situation this would not be an issue unless something is seriously wrong.

- There have been numerous mentions of different Battery Management Systems.  Is there currently a reliable BMS that is simple, and has all the safety features? (balancing, low voltage disconnect, high voltage disconnect) The "simplest" BMS I've seen and has been mentioned here is the one from evparts.com.au but requires importing from Australia.  Does anyone have experience using that one?

- There was concern about disconnecting the charge controller from the battery bank while PV is still connected to charge controller.  Is this an issue I need to be concerned about?  You already have a circuit breaker between the CC and battery bank that can be turned off manually even when PV is active.

- The evparts.com.au has a 240A latching relay to disconnect loads.  Would this not work for disconnecting an inverter?  I realize this is a fail-safe and the Outback inverter should be programmed to disconnect below a certain voltage already.

- With a LiFePO4 battery bank I should remove any temp sensors used by the system to control the voltage charge levels since the LiFePO4 is not need voltage levels adjusted based on temp.


My specific instance is for replacing a 250Ah 48V AGM battery bank which uses an Outback VFX3648, MX60, and Midnight Classic.

Estragon

I haven't found disconnecting batteries while still connected to pv to be a problem with my classics. What is a potential problem is removing both power sources, then applying pv power before battery power. You want to let the controller boot fully with battery power before closing the pv breaker. As long as the classic stays awake though, it hasn't been a problem for me.

An eclipse or really dark cloud could change that experience though

As I understand it, there should be no temp comp for these EXCEPT zero charging with freezing temps. I can't recall offhand if Outback has a temp comp vector - I think so, and the classic does. I'd set the vector to 0v/degree as well as removing RTS .

I'm no expert on LiFePo so FWIW.

BB.

Most charge controllers still have the internal temperature sensor--So not installing the remote battery temperature sensor does not stop "temperature based voltage offsets" from occurring.

Disabling the Temperature offset in the configuration screen is still needed (for various Li-Ion batteries).

I have not heard of a Charge Controller where you can program the minimum charging temperature (i.e., no charging at 32F/0C for Li batteries).

Just what is needed--More complexity/configurations to make things even more confusing.

-Bill

Dave Angelini

Did the OP look at the Simpliphi batteries in the store here? Very simple usage with any electronics and built-in BMS.

DaveB

Dave Angelini said:

Did the OP look at the Simpliphi batteries in the store here? Very simple usage with any electronics and built-in BMS.

With the SimplePhi you are looking at  $3679  * 3 =  $11K for ~200Ah 48v battery bank which still might not be covered by warranty because they state you have to size your bank based on inverter size.  A rough guess is that purchasing the cells yourself and adding protection would be about $7K (using 16 * $358/battery for 2v 200Ah + protection).  

Estragon

> @Dave Angelini said:
> Did the OP look at the Simpliphi batteries in the store here? Very simple usage with any electronics and built-in BMS.

OP specifically said they weren't looking at them. I agree they simplify things, but not the question as asked.

Dave Angelini

Estragon said:

> @Dave Angelini said:
> Did the OP look at the Simpliphi batteries in the store here? Very simple usage with any electronics and built-in BMS.

OP specifically said they weren't looking at them. I agree they simplify things, but not the question as asked.

I did not really get that from the post. I got the part about getting batteries "hot". Not everyone is really good at doing this right!
Just saying!

Estragon

@Daveb - assuming you could get 200ah at 48v for $7,000, and could use ~80% of that, you have 7000/160ah=~$43/ah. Using L16s you'd get ~350ah at 48v for around $2400, using 50% of that would be 2400/175=~$14/ah. Tough to make a case for LiFePo on money grounds unless you assume much better cycle life ( which may or may not turn out IRL), or have special needs like fast charging. I'm not against the tech, just sayin.

Raj174

@DaveB
There are about 5 forum members here who use DIY prismatic LFP batteries. A couple using 72AH cells in a series/parallel arrangement. One using 32 90AH cells. Another using 180AH cells in series and my bank of 16 195Ah cells in series. Forum member karrak, with the 90AH cells, uses a home made BMS. I have no BMS on my bank and it has been cycling daily for  over a year.

This is where I bought my batteries along with a couple of other members:
http://www.electriccarpartscompany.com/Prismatic-Lithium-Batteries_c_1.html

The better quality manufactures (better warranty) are CALB, SYNOPOLY, WINSTON, GLB AND HIPOWER, although HiPower went out of business in 2014. All of these manufacturers arose from the breakup of a company called Thundersky sometime around 2008. (Winston still retains the rights to the name Thundersky) These are the the cells the EV guys use and they work great for solar also.

A 51.2 volt 200AH bank (16 3.2V cells) would be about 4000.00 delivered

Cells capable of 2C and 3C are fine. It just means they have extremely low internal resistance which is a good thing.
I would look for 1 milliohm or less. And not to worry, you'll probably never draw over C/3 in a solar power setup anyway.
Top balancing the cells to 3.6 volts takes time and is well worth the care and effort to keep the bank balanced for over a years time without a BMS.
I installed high and low voltage disconnects on my CC to battery and inverter to battery cables for redundancy. These are controlled by a digital volt and amp hour meter.
I use a Midnite classic 150 and it works well with the LFP bank. Very configurable charge settings.
I would be happy to answer any questions you have about using LFP with solar power.

Rick    

Solray

If I were a non DIYer, and wanted lithium batteries, I think I'd go with Tesla for the warranty and ease of use, both important to the non DIYer.

karrak

DaveB said:

Where are people getting new 3.2V LiFePO4 cells from?  My understanding is I should be looking for 1C or less cells.  Are some brands preferable then others for a renewable energy environment?

I would add Headway to Raj's list of batteries with a track record with off-grid solar. If you can keep maximum discharge rates to less than 1C and charge rates to less than 0.5C with the Chinese prismatics you should get longer lifespans.

- Does it matter if the bank has aluminum strapping around it that I've seen on some banks that is supposed to prevent deformation if the battery gets hot?  I assumed in a RE situation this would not be an issue unless something is seriously wrong.

I agree that the strapping is only necessary for batteries used under severe conditions. There is no sign of any bulging batteries in the two 4 year old batteries that I have installed without strapping.

- There have been numerous mentions of different Battery Management Systems.  Is there currently a reliable BMS that is simple, and has all the safety features? (balancing, low voltage disconnect, high voltage disconnect) The "simplest" BMS I've seen and has been mentioned here is the one from evparts.com.au but requires importing from Australia.  Does anyone have experience using that one?

The main features that I think are important with a BMS are

• Programmable cell overvoltage and undervoltage alarm and disconnect values
• Suitable outputs to stop charge controller, inverter and others loads
  
• Programmable cell balancing criteria
• Current sensing for SOC calculation

Less important would be

• Remote access to data via computer, internet and phone
• Ability to upgrade BMS software
• Capacity to control loads

Here are some that I would look at:
Orion Jnr expensive but from the USA
123SmartBMS Only drawbacks are that it doesn't draw power equally so you would have to have to use the automatic battery balancing feature and it used hall effect sensors for the current monitoring that are not as accurate as using a current shunt.
Zeva BMS16  I have only seen this one recently but ticks most of the boxes. I am thinking of writing some software to hook my battery monitoring system up to this one.
Batrium Also from Australia

- There was concern about disconnecting the charge controller from the battery bank while PV is still connected to charge controller.  Is this an issue I need to be concerned about?  You already have a circuit breaker between the CC and battery bank that can be turned off manually even when PV is active.

I can see that depending on how well designed the controller is that if it was disconnected from the battery when charging at full power that the voltage output from the controller could go well above the maximum battery voltage which might damage other equipment connected to the controller. It would have to be very badly designed for this to damage the controller!

- The evparts.com.au has a 240A latching relay to disconnect loads.  Would this not work for disconnecting an inverter?  I realize this is a fail-safe and the Outback inverter should be programmed to disconnect below a certain voltage already.

Looking at your inverter's installation manual it looks like it has a remote inverter on/off that could be connected to the LVD output from a BMS which makes the 240A relay redundant.

- With a LiFePO4 battery bank I should remove any temp sensors used by the system to control the voltage charge levels since the LiFePO4 is not need voltage levels adjusted based on temp.

BB has answered this

My specific instance is for replacing a 250Ah 48V AGM battery bank which uses an Outback VFX3648, MX60, and Midnight Classic.

I am fairly sure that you can turn the Midnite Classic off via the AUX2 input and that you can program it for a bulk voltage of 55.2V and float voltage of 53.6V so the gear you have should work well with an LFP battery and hook directly up to the  BMS without any extra relays.

Simon

DaveB

Thanks for all the very informative information.  Two other questions:

- Can you expand your LiFePO4 battery bank with another string years later without adverse affects?  With lead acid you are told what a bad idea it is to add new batteries to an "old" bank.

- Is there much of an issue adding multiple strings to increase Ah with LiFeO4?  Again with lead acid the preferred setup is a single string.  


To reply to an above comment, I know flooded lead acid would be much cheaper per Ah but I do not want the associated maintenance, that is why I currently have AGMs.  I'd also have to worry about extended periods of low state of charge, making the necessity of generator usage more likely.

karrak

As far as I am aware I can't see any adverse effects to adding extra strings to an LFP battery at a later date. My understanding is that LFP batteries loose capacity at a reasonably fixed rate and the battery impedance/resistance only goes up slowly with time. So far I think the batteries that I have installed are loosing between 1%-2% per annum and I haven't notice any change in battery impedance. If this continues I intend to add more capacity when the capacity loss of my battery becomes too much to extend the life of the battery.

There is no problem with running LFP cells in parallel as long as they are wired up correctly. This post shows how to wire a 2p16s LFP battery. I think allot of the problems with wiring LA batteries in parallel is that the wiring is not done correctly so the batteries do not share the current evenly.

On batteries that are cycled everyday the cost per Ah or kWh cycled trough the battery is less for lithium ion batteries that LA batteries.

Simon

Raj174

@DaveB

- Can you expand your LiFePO4 battery bank with another string years later without adverse affects?  With lead acid you are told what a bad idea it is to add new batteries to an "old" bank.

I think the way to look at this is to consider that capacity is reduced as the number of cycles increases.
If the new string is added within the 500 to 600 cycles and the bank is rebalanced, I don't think there would be a problem. At higher cycles, I believe the difference in cell capacity could show up as longer charge and discharge times for the newer cells. Just my opinion.

This is a sample LFP life cycle vs capacity graph:


   
- Is there much of an issue adding multiple strings to increase Ah with LiFeO4?  Again with lead acid the preferred setup is a single string.

I don't know of a problem using multiple strings, however I do try to follow the KISS principle, so I will stick with a single string.

Rick 

nickdearing88

I just upgraded to BattleBorn batteries. I know you're looking for more of a DIY solution but my primary reasons for choosing LiFePO4 were: Not worrying about partial-SOC, VERY low internal resistance, great surge capacity, higher charge rates, lower self-discharge rates, very flat discharge voltage curves, and lower overall cost (provided they hold up to the expected cycles). Then there are added benefits like major weight reduction: 155lbs of AGM vs 58lbs of LiFePO4 providing me more Ah capacity (due to cycling down to 20-30%) My application is not mobile but if I were planning a system for a camper or boat, this weight difference could be a huge consideration.

I'm very practical and price was surely a factor but Lead Acid and Lithium are very different technologies and don't think they can be equally compared (apples to oranges).

Sure LiFePO4 has its limitations, cold-weather charging being the biggest issue for many, but for my money, I appreciate the above-mentioned benefits.

I also checked with Electric Car Parts Company, as other users have mentioned, and their customers seem to have good results. They have a variety of cell brands and BMS components.

A custom built system with prismatic cells and BMS is in my future for sure. I just wish I had @karrak 's software development skills!

karrak

Raj174 said:

This is a sample LFP life cycle vs capacity graph

Rick, thanks for posting the graph, very interesting. Do you have any more information on it like battery manufacturer, charge and discharge rates and depth of charge and discharge.

Simon

Raj174

@karrak
The graph posted was from a few if found in a search. I can't find any info on it. It is possible that it may be the graph of a cylindrical LFP cell, although I did not think they would be different than the prismatic cells. As I now recall your 4 year old set has lost very little capacity. Is that correct? I have not been able to find a graph for prismatic cells yet. Anyway, this is a study that indicates that the graph posted may be for a cylindrical cell.

Rick

karrak

The problem with allot of these graphs and studies is how do we relate them to our systems. What are the test conditions for the graph from a few posts ago. If it is cycles at 100%DOD cycling and a charge/discharge rate of say 0.5C how does this relate to my system where the average daily DOD cycling is ~30% and average charge/discharge rate is ~0.025C or my friend's system where the average daily DOD cycling is ~40% and average charge/discharge rate is ~0.033C? With the study in the last post we have no idea what is defined as a cycle or what the charge/discharge parameters were. Were the cells being charged and floated at 3.65V/cell?, what was the cycled DOD?

I estimate the loss in capacity of my battery to be between 1% and 2% per annum. Unfortunately I didn't have the equipment to do a capacity test on my battery when it was new so can't say with any accuracy how much capacity has been lost. Only clues I have as to how much loss there has been is that earlier this year my battery got down to an SOC of ~9% with an average cell voltage of ~3.04V which from my charge discharge curves corresponds to an calculated SOC of ~5%. The other clue is that I did a capacity test on a cell out of my friends battery and it came back to my surprise at ~90Ah for a 90Ah cell.

Simon

nickdearing88

I'm not sure about the testing conditions on the graph Rick posted but from my Battle Born research, their stated tests are as follows:

12v nominal 100Ah pack (4 cells in series)
     - Charged at 1C to 3.6v/cell (14.4v), terminating charge at C/100 end amps
     - Discharged at 1C to 2.75v/cell (11v)
     - ~80% capacity remaining at 3000 cycles

I didn't ask about temperature but I would assume STC (25 degrees C).

I assume that by charging and discharging at specs like mine (and @karrak), we should see much greater cycle ability.

westbranch

and if you plan for 2 days of autonomy, 1 extra day before it needs to be charged, it will still last 3000 cycles and have a lifetime that is 2 x longer than you currently expect as a cycle is every 2 days , not 1 day...

DaveB

The biggest issue I see right now with building your own LiFePO4 bank is dealing with a battery management system.  There does not appear to be a simple solution like is provided built-in with the Battle Born or SimplePhi.

The Orion Jr. for instance which is from the USA is going to be at least $600 and on top of that you are highly recommended to rent a device to verify everything is connected properly before you even connect the BMS to the batteries, otherwise you risk destroying it and that isn't covered by warranty. It also requires you to program the BMS with a windows or linux computer.  If I only have a 180Ah bank initially which I know will be too small for going off-grid I could add another 180Ah in parallel but with the Orion BMS I'd have to spend another $600+ to cover that set of cells as well.

So now I am looking at getting battery balancers such as http://www.electriccarpartscompany.com/3V-1S-Lithium-Lighted-Battery-Balancers, a means to monitor the voltage of each of the 16 cells, and latching relay if voltage it too high or low to disconnect the bank.

mike95490

So now I am looking at getting battery balancers such as http://www.electriccarpartscompany.com/3V-1S-Lithium-Lighted-Battery-Balancers, a means to monitor the voltage of each of the 16 cells, and latching relay if voltage it too high or low to disconnect the bank.

The sad thing about the balancer boards, is failures.  Any way to check on their track record? 

And they can only bypass 6A.  If you are charging with 20A, you still have 14 going into the battery and you have to rely on the balancer to trip the Charger Stop Command and stop the charging.  And you have one cell at Full charge voltage, and the rest of the pack is still behind. I suppose you could rig up another 5 A  charger and get all the cells topped off,

Not saying you can't use them but that they are not magic and can fail, just like the rooftop microinverters are doing

DaveB

I'd really prefer a battery bank with little maintenance which is why the LiFePO4 are so attractive.   I'm worried about longevity of larger AGM bank which would be traditional option for little maintenance.

jonr

There is some advantage to charge shuffling vs top balancing.  The former is always working to balance vs only kicking in after voltage reaches some limit.  So it's more likely to keep up.

Another option is a charger that can throttle amps depending on battery state.   Ie, if a 6A bypass is on, limit charging to 6A.  Otherwise use 20A.  Yet another is matching the bypass capability to the charger (ie, 20A for both).

nickdearing88

mike95490 said:

So now I am looking at getting battery balancers such as http://www.electriccarpartscompany.com/3V-1S-Lithium-Lighted-Battery-Balancers, a means to monitor the voltage of each of the 16 cells, and latching relay if voltage it too high or low to disconnect the bank.

The sad thing about the balancer boards, is failures.  Any way to check on their track record? 

And they can only bypass 6A.  If you are charging with 20A, you still have 14 going into the battery and you have to rely on the balancer to trip the Charger Stop Command and stop the charging.  And you have one cell at Full charge voltage, and the rest of the pack is still behind. I suppose you could rig up another 5 A  charger and get all the cells topped off,

Not saying you can't use them but that they are not magic and can fail, just like the rooftop microinverters are doing

I agree but most users have experienced very little imbalance in cells, even after heavy cycling. This means that by the time ANY of the cells are near their high voltage limits, the charge controller should be in Absorb (constant voltage) and the current being "taken" by the batteries should be very low.

It would seem to me that if the cells are so imbalanced that one or more is near the top-end voltages while others are so low that the CC is still in Bulk mode (higher currents), the cells were not well-matched to begin with.

That's my experience so far, but only a couple weeks (and seems to correspond to the longer-term users here). @Raj174 and @karrak , wouldn't you agree?

PNjunction

DaveB said:

I'm one of those people who use and wrote up exactly what you are talking about here.

- Where are people getting new 3.2V LiFePO4 cells from?  My understanding is I should be looking for 1C or less cells.  Are some brands preferable then others for a renewable energy environment?

You can use any "C" you like, but for our application, solar storage, properly sized in capacity means that you don't need high-C cells if your system is designed for 0.1C discharge or less so you have adequate runtime and for many that means extra days of no sun!  So the "high-C" rate cells are basically a waste of money if you are only discharging at say .1C or less right?

For a DIY'er, I've always said to keep it simple with 4 prismatics for a typical experimental 12v battery.  I find that the 40ah cells are a good budgetary limit in case you make a total mistake!  Ie, don't bite off more than you can chew, but hey, if you wanted to start out with 200ah cells, and have the budget and risk factor taken into account - go for it.

The cells we deal with appropriate for low-rate discharge compared to ah sizing, are not high-rate cells.  1C continuous, and 3C for short <30 second bursts.  Still, that is waaaay overkill.  I mean, who designs a solar system with lead-acid for 1C, or even .5C discharge?  Nobody.  So the so-called lower-performance cells like GBS, Winston, CALB etc suit our purposes just fine.  We're not starting a diesel locomotive, but running lights off inverters, etc.

GBS (aqua cells with purple cell covers) works for me, and the reason I recommend it for DIY'ers starting out is *safety*.  Only GBS provides the cell covers so you don't have to also DIY your own terminal cover system.  In your haste to get things running, the cell covers with other brands become a "get a round to it" kind of thing, and then you drop a wrench and ruins your day.  ElitePowerSolutions is just one vendor.

- Does it matter if the bank has aluminum strapping around it that I've seen on some banks that is supposed to prevent deformation if the battery gets hot?  I assumed in a RE situation this would not be an issue unless something is seriously wrong.

The strapping is to help force the vents to open, rather than let the cases expand if you system goes haywire.  Often overlooked, the need for strapping is to prevent terminal-torque twisting (side by side or up and down), which will kill the battery.  Somebody backed into your battery box?  Vibration from a neaby air-compressor hitting the box/cells?  etc etc.  Because the cells are so light, accidental movement which twists the terminal connectors the wrong way is a bad thing. 

- There have been numerous mentions of different Battery Management Systems.  Is there currently a reliable BMS that is simple, and has all the safety features? (balancing, low voltage disconnect, high voltage disconnect) The "simplest" BMS I've seen and has been mentioned here is the one from evparts.com.au but requires importing from Australia.  Does anyone have experience using that one?

Talked to death on most forums.  Something I touted early on was the KISS principle, which we can get away with *due to our low-current discharge application*.  Even the hard core marine guys have gotten in on the act, many running lower voltages than they used to, disabling the parasitic balancing (mosfets failing closed, and UNbalancing a perfectly balanced bank), etc.

Some BMS companies have gone out of business, or to another product, and replacement parts are hard to find.  Once again, the KISS principle works here (nothing more than top balancing initially by charging each cell individually first), and setting a conservative low-voltage disconnect.  It is a catch-22 situation - the reason I need individual cell monitoring is because the bleeder/balancer mosfet failed, draining one of my cells!  And sadly, for those balancing systems that do work correctly, underneath there may be bad quality cells that DO need this kind of babysitting - and the time needed to do the balance hurts the cells in the long term.

- There was concern about disconnecting the charge controller from the battery bank while PV is still connected to charge controller.  Is this an issue I need to be concerned about?  You already have a circuit breaker between the CC and battery bank that can be turned off manually even when PV is active.

Just treat your LFP bank with the same concerns externally as a PB bank.  *HOWEVER*, because lithium cells will give you everything they have until they are nearly depleted without much voltage drop, FAST fuses, and quality dc disconnects cannot be skimped upon.  Ie, no car-stereo stuff.

- The evparts.com.au has a 240A latching relay to disconnect loads.  Would this not work for disconnecting an inverter?  I realize this is a fail-safe and the Outback inverter should be programmed to disconnect below a certain voltage already.

Most inverters are set for Pb, not LFP low voltage disconnects, such as 10.7 or 11.2v or thereabouts.  For LFP, you would want that to be higher, about 12.7v.  *At our discharge rate*, that is conservative enough.  Still, 10.7v is better than going down to zero volts, provided you bring the bank back up SLOWLY (no more than .05C) until each cell reaches 3.2v, whereupon you can apply the normal charge rate, typically 0.1 to 0.3C or so.

- With a LiFePO4 battery bank I should remove any temp sensors used by the system to control the voltage charge levels since the LiFePO4 is not need voltage levels adjusted based on temp.

YES, *disable* temperature compensation.  One inexpensive controller I had that didn't disable compensation was tricked into doing so by setting the temperature variable to 0 (zero) mv per c.

My specific instance is for replacing a 250Ah 48V AGM battery bank which uses an Outback VFX3648, MX60, and Midnight Classic.

Don't bite off more than you can chew at first.  Ie, would you recommend to a newcomer to assemble a 250ah / 48v lead-acid bank right off the bat, or would you recommend they start out with just one or two batteries and a simpler system to get their feet wet with experience?

Hence, my recommendation for the DIY'er not to spend too much money up front, and prove to themselves that they can take care of a larger bank so they don't blow a hole in their wallet.  It's called the "learner battery", whether Pb or LFP for a reason.

Dave Angelini

That was excellent, as usual Mr. Diode 

karrak

nickdearing88 said:

mike95490 said:

So now I am looking at getting battery balancers such as http://www.electriccarpartscompany.com/3V-1S-Lithium-Lighted-Battery-Balancers, a means to monitor the voltage of each of the 16 cells, and latching relay if voltage it too high or low to disconnect the bank.

The sad thing about the balancer boards, is failures.  Any way to check on their track record? 

And they can only bypass 6A.  If you are charging with 20A, you still have 14 going into the battery and you have to rely on the balancer to trip the Charger Stop Command and stop the charging.  And you have one cell at Full charge voltage, and the rest of the pack is still behind. I suppose you could rig up another 5 A  charger and get all the cells topped off,

Not saying you can't use them but that they are not magic and can fail, just like the rooftop microinverters are doing

I agree but most users have experienced very little imbalance in cells, even after heavy cycling. This means that by the time ANY of the cells are near their high voltage limits, the charge controller should be in Absorb (constant voltage) and the current being "taken" by the batteries should be very low.

It would seem to me that if the cells are so imbalanced that one or more is near the top-end voltages while others are so low that the CC is still in Bulk mode (higher currents), the cells were not well-matched to begin with.

That's my experience so far, but only a couple weeks (and seems to correspond to the longer-term users here). @Raj174 and @karrak , wouldn't you agree?

I totally agree, from my own experience and the experience of the vast majority of people who post on line  who have set their LFP batteries up properly in the first place is that balancing is just not the issue that those who don't have LFP battery based systems make it out to be.

If an LFP battery is properly balanced as part of the commissioning process to start with unless there is a fault within one of the cells or in any equipment connected to the individual cells they will only go out of balance very slowly over a period of months or years.

If an LFP battery is top balanced, when charging the individual cell voltages will all stay very close together until the SOC percentage has reached the high nineties . By this stage the overall battery voltage will have reached the Absorb/CV voltage and the current will start tapering off. The longer the absorb/CV period the closer to 100% charged the battery becomes, the more the individual cell voltages will diverge and the lower the overall charge current. Usually the charge is terminated after a set period of absorb/CV time or when the charge current has decreased to between C/20-C/50.

Under these conditions the cell voltages may have diverged by say 0.050V-0.100V but  the highest cell voltage will still be well under the voltage where damage is being done to the cell. Over time unless there is balancing circuitry to keep the overall cell balance within say 0.050V or there is someone monitoring the imbalance and manually re-balancing the battery the voltage difference will slowly get larger until a point is reached where the cells at the higher voltages are being damaged and the cells at the lower voltages will not be being fully charged.

The graph below shows the cell voltages diverging and the charge current (the bottom line) during the absorb/CV phase of the charging of my friend's system. This battery is slightly out of balance and has since been manually balanced.

Simon

karrak

jonr said:

There is some advantage to charge shuffling vs top balancing.  The former is always working to balance vs only kicking in after voltage reaches some limit.  So it's more likely to keep up.

I also see problems with charge shuffling with LFP batteries because of how the difference in individual cell voltages can vary depending on the impedance/internal resistance characteristics of the individual cells. As an example lets say we have two cells both of which have exactly the same capacity and are at the same SOC i.e. they are balanced and Cell A has a lower internal resistance than Cell B. The difference in OCV (open Circuit voltage) of the two cells will most likely be within 1mV (0.001V) of each other over the SOC range 75-25% which will not trigger the balancing function but if we draw a current of say C/2 from both cells it is quite possible to get a voltage difference of between 20-40mV between the cells due to the different cell impedance/resistance. This may trigger the balancing circuitry to shift energy from Cell A to Cell B and thus make the battery out of balance. When both cells are nearly full Cell B's voltage will rise above Cell A's due to the imbalance caused earlier so energy will be shifted from Cell B back to Cell A. This unnecessary shunting of current backwards and forwards will decrease the life of the battery.

I think the best way to balance either manually or automatically is to note the voltage difference and charge current at the end of the absorb/CV charging phase. By looking at the charge curves you can calculate how much charge has to be taken from or added to each cell. You then either put a load or charger on each of the cells that is out of balance for the time that it will take to add or subtract enough charge to balance the cells.

For example assume there are two cells in series charged to 6.90V (3.45V/cell) with cell voltages of 3.50 volts and 3.40 volts at the end of charge with the end charge current being 0.05C. From the charge/discharge curve below we can see that at 3.50 volts the SOC is ~99%, at 3.45V it is ~98% and at 3.4V it is ~95%. So to get the battery back in balance we need to subtract 1%SOC from the high cell and add 3% to the low cell. We could of course just subtract 4% from the high cell to balance the cells. So as to not overshoot rather than the 3% and 1% I would divide this figure by 2 and only subtract 0.5% from the high cell and add 2% from the low cell.


Simon

karrak

PNjunction said:

The cells we deal with appropriate for low-rate discharge compared to ah sizing, are not high-rate cells.  1C continuous, and 3C for short <30 second bursts.  Still, that is waaaay overkill.  I mean, who designs a solar system with lead-acid for 1C, or even .5C discharge?  Nobody.  So the so-called lower-performance cells like GBS, Winston, CALB etc suit our purposes just fine.  We're not starting a diesel locomotive, but running lights off inverters, etc.

That is one of the big advantages of LFP batteries over LA batteries. I can easily run a 4kW inverter off a 10kWh battery. It doesn't happen very often but is very handy to run my welder or be able to have the microwave, toaster and electric oven on at the same time.

- There have been numerous mentions of different Battery Management Systems.  Is there currently a reliable BMS that is simple, and has all the safety features? (balancing, low voltage disconnect, high voltage disconnect) The "simplest" BMS I've seen and has been mentioned here is the one from evparts.com.au but requires importing from Australia.  Does anyone have experience using that one?

Talked to death on most forums.  Something I touted early on was the KISS principle, which we can get away with *due to our low-current discharge application*.  Even the hard core marine guys have gotten in on the act, many running lower voltages than they used to, disabling the parasitic balancing (mosfets failing closed, and UNbalancing a perfectly balanced bank), etc.

Some BMS companies have gone out of business, or to another product, and replacement parts are hard to find.  Once again, the KISS principle works here (nothing more than top balancing initially by charging each cell individually first), and setting a conservative low-voltage disconnect.  It is a catch-22 situation - the reason I need individual cell monitoring is because the bleeder/balancer mosfet failed, draining one of my cells!  And sadly, for those balancing systems that do work correctly, underneath there may be bad quality cells that DO need this kind of babysitting - and the time needed to do the balance hurts the cells in the long term.

I agree that having a fault in automatic battery balancing circuitry can and has led to LFP batteries going out of balance. Maine Sail on the Cruiser forum comes to mind. He has cell level voltage monitoring so was able to detect the fault before it damaged his battery. Gordon on the Alternate Energy website had a fault with one of his CALB cells which led to it loosing around 25% of its capacity. He hasn't got individual cell monitoring but monitors it in 4 cell blocks which allowed him to detect the fault. These sort of faults are fairly rare but they do happen.

Most inverters are set for Pb, not LFP low voltage disconnects, such as 10.7 or 11.2v or thereabouts.  For LFP, you would want that to be higher, about 12.7v.  *At our discharge rate*, that is conservative enough.  Still, 10.7v is better than going down to zero volts, provided you bring the bank back up SLOWLY (no more than .05C) until each cell reaches 3.2v, whereupon you can apply the normal charge rate, typically 0.1 to 0.3C or so.

IMO 12.7V (3.175V/cell) is too high. Just looking at the logged data from my friend's system shows that the lowest voltage under load that his battery got down to this month was 50.37V (~3.15V/cell) and yet the lowest SOC was ~33%

Simon

Dave Angelini

karrak said:

That is one of the big advantages of LFP batteries over LA batteries. I can easily run a 4kW inverter off a 10kWh battery. It doesn't happen very often but is very handy to run my welder or be able to have the microwave, toaster and electric oven on at the same time.

Right but to be fair, I have not seen much out there that can run more than 5KW continuous and 7KW peak on Li-ion. Don't want to mislead anyone that for large systems LA is still the way to go. Outback and Schneider are throttling down their inverters to protect the Li-ion banks from large surges. If they were not, the BMS would shut down the battery next.