Simple LiFePO4 battery bank for DIY?

2

Comments

  • jonr
    jonr Solar Expert Posts: 1,386 ✭✭✭✭
    "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"

    At C/2, a charge shuffler moving less than C/2 is actually moving current from high cells to the load, not cell to cell.  Ie, it's protecting the weak cells from wear, increasing life of the string.

    Charge shuffling results in top and bottom balance - which is going to prevent wear-inducing extremes in some cells.

    I am available for custom hardware/firmware development

  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    Dave,
    While this is true for currently manufactured packs like the Resu 10 and  DIY LFP 10 kWh banks, it's not a limitation for larger 20 kWh (400AH@51.2V) DIY LFP banks capable of 6700 watts continuous and much higher peak. And 400AH is by no means the largest prismatic LFP cell. But I do believe that the availability of much larger prefabricated banks is not far away.
    By the way, my 10 kWh LFP bank will start our 2 ton 2 stage heat pump. The 4000 watt Radian inverter isn't crazy about it since it's an 8000 watt surge, but it does manage it.

    Rick
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.
  • solar_dave
    solar_dave Solar Expert Posts: 2,397 ✭✭✭✭
    Anyone thought about using Chevy Volt or Nissan Leaf batteries?  Seems to be fair number of them showing up as pulls from the junk yards at a pretty reasonable price.  The Volt uses liquid cooling so simple overheat protect while charging seems to make them last a long time. I have a Volt with over 50K miles and 6 years of use. The battery looks to be performing just below new specs.  Lots of them up on eBay. 
  • Dave Angelini
    Dave Angelini Solar Expert Posts: 6,730 ✭✭✭✭✭✭
    Rick, this may be true, but none of my clients would ever do what you guys are doing. Over 100 and there are none who would do this with my name on it.

    I also have been told that surging over the spec is like taking 5 cycles away, or worse. This is why it is measured and a really good BMS will shut down if the inverter has not already. A really good BMS can also have firmware changed to allow certain users to exceed the limits but at reduction in warranty. When they are in the store here and at my supplier, is when "not far away" is here to me. Wiring together a bunch of parallel banks is not what I want to install either.

    I do like watching the homebrew side!  Keep it up! 
    "we go where power lines don't" Sierra Nevada mountain area
       htps://offgridsolar1.com/
    E-mail offgridsolar@sti.net

  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    I agree, it's DIY only. I do plan to go to a 20 kWh 51.2V bank when I have scraped up enough shekels.
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.
  • Dave Angelini
    Dave Angelini Solar Expert Posts: 6,730 ✭✭✭✭✭✭
    It does always come down to the shekels, and using them easily. I do fine on 9 kWh, but not many can in a home with people offgrid.
    20 kWh is very nice
    "we go where power lines don't" Sierra Nevada mountain area
       htps://offgridsolar1.com/
    E-mail offgridsolar@sti.net

  • bill von novak
    bill von novak Solar Expert Posts: 891 ✭✭✭✭
    Anyone thought about using Chevy Volt or Nissan Leaf batteries?  
    A company out here called EV West has Tesla/Smart batteries for $230/kwhr.  57 volt, which is a voltage that works well with most 48V inverters.  They expect liquid cooling, but in a temperature controlled indoor application at low C rates that shouldn't be needed.
  • karrak
    karrak Solar Expert Posts: 326 ✭✭✭✭
    jonr said:
    At C/2, a charge shuffler moving less than C/2 is actually moving current from high cells to the load, not cell to cell.  Ie, it's protecting the weak cells from wear, increasing life of the string.
    The active balancers/charge shufflers that I have seen transfer the current to the the adjacent cells. I agree that if this raised the voltage of the adjacent cell by enough to activate that cell's balancer and so on that the current would be transfered directly to the load. As the balancing current is fairly low I doubt it will raise the cell voltage enough so the current will end up going to the next cell until it gets too far out of balance. This will mean the current will end up "hopping" its way through each of the cells or even worse hopping backwards and forwards
    Charge shuffling results in top and bottom balance - which is going to prevent wear-inducing extremes in some cells.
    I can maybe see this being beneficial in high power applications like a bus or truck. Tesla don't bother with active balancing on their Model S, not sure what they do on their other cars.

    My off-grid battery spends most of its time in the top 50% SOC,  the lowest recorded SOC was 9% and is regularly at an SOC in the high 90s. Under these conditions I don't see any advantage trying to keep the battery balanced at both the top and bottom, in fact I only see that the extra shunting of charge between the cells that would be needed to keep it balanced at both the top and bottom would cause extra wear on the battery.

    Simon 
    Off-Grid with LFP (LiFePO4) battery, battery Installed April 2013
    32x90Ah Winston cells 2p16s (48V), MPP Solar PIP5048MS 5kW Inverter/80A MPPT controller/60A charger, 1900W of Solar Panels
    modified BMS based on TI bq769x0 cell monitors.
    Homemade overall system monitoring and power management  https://github.com/simat/BatteryMonitor
     

  • jonr
    jonr Solar Expert Posts: 1,386 ✭✭✭✭
    edited August 2017 #40
    > The active balancers/charge shufflers that I have seen transfer the current to the the adjacent cells.

    Sounds like this is the source of confusion.   An active balancer/charge shuffler doesn't transfer charge to adjacent cells.  It transfers charge to the *terminals* of the adjacent cell.   If 10A is flowing out of a cell and you apply 5A to it's terminals, no current will flow into the cell.    This is always true, there are no thresholds involved.

    Possibly the pulsed nature of charge shuffling could cause an issue.  Ie, a 20A 50% duty cycle square pulse, averaging out to 10A, would cause some flow into a cell that had a continuous 10A outflow.  It's unclear that such pulses would cause any significant wear.

    I expect that the best argument against charge shuffling may be that it's expensive and isn't required.

    I am available for custom hardware/firmware development

  • karrak
    karrak Solar Expert Posts: 326 ✭✭✭✭
    jonr said:

    I expect that the best argument against charge shuffling may be that it's expensive and isn't required.
    I agree, one positive with the charge shuffling balancers is that because they can shunt a larger balancing current than the passive resistive balancers and do it in both directions is that there is less likelihood of any damage in the event of someone buying a number of cells and whacking together a battery without balancing it to start with. Not recommended, but I am sure it does happen!

    Simon
    Off-Grid with LFP (LiFePO4) battery, battery Installed April 2013
    32x90Ah Winston cells 2p16s (48V), MPP Solar PIP5048MS 5kW Inverter/80A MPPT controller/60A charger, 1900W of Solar Panels
    modified BMS based on TI bq769x0 cell monitors.
    Homemade overall system monitoring and power management  https://github.com/simat/BatteryMonitor
     

  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Back to the DIY aspects - although too many replies to replies gets confusing. :)

    Leaf batteries / etc - NOT LFP chemistry.  Different voltages.  Not KISS.  Buying crash victim rejects.  'Whole different DIY story.

    Re Maine Sail and other boaters - actually removed the balancer / shufflers as one failed and unbalanced a good bank.  Most are running very conservative like I proposed years ago.  Initial charge/balance, followed by sane low-voltage disconnects, low-ish upper CV voltages.

    The simplistic DIY aspect follows these rules, but nobody believes them.  Either that, or they get waaaay too complicated by those addicted to gadgetry.

    It's not that hard to go KISS for a DIY'er, as long as you don't start out with used / reject trash.  KISS is secret that bms manufacturers and other scare-mongers don't want to be widely talked about.

    Kind of like buying quality 2v lead-acid cells, building a bank, and doing your own maintenance.  It's not that hard.  Neither is LFP, but outside interests make it so.  Sad.


  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Here are the KISS rules again for diy'ers:

    1) Buy reputable quality cells from a reputable dealer.  That means each cell will be matched reasonably well for both capacity and internal resistance.  An *exact* match is not strictly necessary.   DO NOT be a cheapskate, which is tempting for diy, and that means used / old stock, or deals too good to be true.  THIS is the cause for most of the failures starting out on the wrong foot.  In the lead-acid world, I don't buy "Brojans" that are "slightly used" for the same reason. :)

    2) Charge each with a CC/CV algo, set to 3.6v absorb, and when the current naturally drops to .025C, stop.  Use a charge source preferably between 0.1 to 0.33C.  If you don't know how to *properly* use a programmable bench-supply, "single cell" 3.2v nominal chargers are available for this purpose.  Shoot for 0.2 - 0.3C current capability.  Don't use the wrong voltage charger designed for non-lfp lithiums, ie 3.7v nominal (4.2v cv) chargers !!!

    3) Clean and assemble your "links".  Be SAFE and TAPE your tools from the handle to the head in case you accidentally drop it.  A LIGHT coating of aluminox, penetrox or the like is beneficial.  Dont use wire-brushes for this task.

    4) Set your low-voltage pack-level disconnect to 3.1v per cell as a conservative point to trip at.  Because you are top-balancing, you should expect a small amount of "ragged" levels towards the end.  Because you sized your bank properly (0.1 to maybe 0.2C maximum draw), you have the luxury of time to go pack-level.  Not so with EV'ers or RC'ers.  (Beware of helpful hints from those outside our low-current application, as they are not always correct)

    5) Your CC/CV charger should have a CV from anywhere between 13.8v to 14.4v.  Lower voltages will take longer for the CV absorb to finish.  There is no "holy grail" voltage - it is up to YOU when to stop.  Stop at a specific voltage?  Stop at a specific absorb charge current?  That's up to you.

    6) When your cells reach the upper end of charging, the cells should be between 3.45 to 3.6v.  Ideally, they should be no more than 0.1V apart from each other.  Usually it is much less.

    7) Obsessing over getting the cells to match *exactly* in voltage during charge is a fool's game.  It's nice if they do, but less than 0.1v difference is acceptable.  Playing this game with flaky bleeder / shunt boards that are an additional point of failure, introduce the TIME DEGRADATION aspect of holding cells at full charge while others catch up.  Cycle over cycle, this time factor adds up (oxidation of the cells).   TIME spent doing this needless obsessive exact matching is something overlooked by those who want to sell you all sorts of rat's nest gadgetry.  (or needed to babysit TRASH cells - but we warned about that from the start!)

    8) Did I say don't buy TRASH?  I'll say it again.  When doing a lead-acid bank made of 2v cells, I don't buy "Brojans" from the back of some guy's truck, or lightly-used "Bolls" batteries that sat in a warehouse for 5 years.  Common battery-sense should apply, from purchasing to safety concerns.


  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭
    Hi all,
    I've got an existing Outback Radian 8KW power center with 1150AH Gel batteries.  I'm looking to replace these batteries with 700AH LiFePo.  I have a line on some Chinese made batteries from Eastar battery company.  The batteries are rated 1C discharge and 0.5C Charge, which is plenty sufficient for my application.  The batteries come with a BMS.  The picture of the battery case (each case contains 200AH of batteries) shows two sets of terminals; 1 pair for charge and 1 pair for discharge.  My Outback power center only has a single pair of wires to the existing Gel battery bank.
    Question: Will this pose a problem connecting the charge and discharge to the Outback?  Will i need to bypass the MFG BMS and connect directly to the batteries?
    BTW: the price that I've been quoted is tough to resist.  700AH 48V delivered for $10K.
    Thanks for the help.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭

    Picture of the Eastar battery enclosure. FYI
  • bill von novak
    bill von novak Solar Expert Posts: 891 ✭✭✭✭
    The picture of the battery case (each case contains 200AH of batteries) shows two sets of terminals; 1 pair for charge and 1 pair for discharge.  My Outback power center only has a single pair of wires to the existing Gel battery bank.
    Question: Will this pose a problem connecting the charge and discharge to the Outback?  Will i need to bypass the MFG BMS and connect directly to the batteries?
    If you are going to bypass the BMS you might as well get raw cells for even cheaper.

    Most batteries that have separate discharge/charge terminals do so for a few reasons:

    1) Easier/cheaper/lower loss to design a single unidirectional switch per terminal pair than a bidirectional one
    2) That way if the discharge terminals get disconnected, charging is unaffected (and thus the battery can still be charged.)  Single-terminal systems can do this too but it takes more control logic.

    So if that is truly a battery with separate terminals for both it will be hard to use on your system.

    (BTW there have been several accounts of people buying LiFePO4 battery systems rated for hundreds of amp-hours, only to find it containing a string of much smaller cells.  Caveat emptor.)
  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    @bear_essentials
    I agree with @bill von novak . It is a great price, but I would want to see inside the box first. If it contains larger LFP prismatic cells, then they can be used to set up a battery bank. If it has lots of small cells then I would pass.

    Rick 
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭

    Battery case contains a number of 3.2V 100AH cells, see pic above.
    Thanks for the suggestions/comments.  I'll see if this MFG has option for single set of terminals.  They very well might.  I know that they offer a variety of BMS' for different 'C' capacities.
    I'm leary of bypassing the BMS and using the raw cells.  I know that Outback has settings for Li Ion, but my experience with their charger/charge controllers leaves me concerned that they won't provide all the necessary settings/protections needed for these batteries and I'll wind up with a fire, or at least a mess.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭
    I've also attached the Battery cell specification sheet
  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    Looks like 84 batteries, how are they configured. These batteries have pretty good specs.
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭
    I am guessing that the box is configured with 32 batteries per 200AH 48V box.  3.2V * 16 = 51.2V 100AH * 2, etc.
    I have seen pictures of the box and test specs, but have not seen inside the box.
    Note: the picture that I showed above of the pallet of individual cells did not mean to imply that all of them are used in a single box.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭

    Here's the pack specification.  I was correct, it is configured as 2 parallel sets of 16 cells in series to achieve 200AH 48V. 
  • karrak
    karrak Solar Expert Posts: 326 ✭✭✭✭
    edited September 2017 #53
    Hi all,
    I've got an existing Outback Radian 8KW power center with 1150AH Gel batteries.  I'm looking to replace these batteries with 700AH LiFePo.  I have a line on some Chinese made batteries from Eastar battery company.  The batteries are rated 1C discharge and 0.5C Charge, which is plenty sufficient for my application.  The batteries come with a BMS.  The picture of the battery case (each case contains 200AH of batteries) shows two sets of terminals; 1 pair for charge and 1 pair for discharge.  My Outback power center only has a single pair of wires to the existing Gel battery bank.
    Question: Will this pose a problem connecting the charge and discharge to the Outback?  Will i need to bypass the MFG BMS and connect directly to the batteries?
    BTW: the price that I've been quoted is tough to resist.  700AH 48V delivered for $10K.
    Thanks for the help.
    I agree with @bill von novak that if you can get the individual cells for less do that and organise your own BMS.

    Ideally your Outback Radian can be configured so that external inputs to the Radian can be used to stop charging or turn the inverter off in the event of any of the cells being overcharged or over discharged. The overcharge or over discharge outputs from the BMS you choose would be connected directly to the Radian. What is the model number of the Radian you have?

    Simon
    Off-Grid with LFP (LiFePO4) battery, battery Installed April 2013
    32x90Ah Winston cells 2p16s (48V), MPP Solar PIP5048MS 5kW Inverter/80A MPPT controller/60A charger, 1900W of Solar Panels
    modified BMS based on TI bq769x0 cell monitors.
    Homemade overall system monitoring and power management  https://github.com/simat/BatteryMonitor
     

  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭
    The Radian model is 8048A.
    The problem has been resolved.  The MFG has agreed to provide me with a BMS that has 2 terminals supporting 0.6C charge/discharge rate for the same price.  This will meet my system needs exactly.
    According to the salesperson, the reason for the 4 terminals is cost savings.  When dealing with different rates for charge and discharge (in my case it was 1C discharge and 0.5C charge), it is cheaper to separate the terminals.
  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    That's good news. Let us know how it goes with the new bank.
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.
  • bear_essentials
    bear_essentials Registered Users Posts: 8 ✭✭
    Will do.  I just placed the order yesterday and they will be on a slow boat from China, so it may be several months before I check in.
    I imported the lead acid batteries directly from China and it is never a simple process, but if you have the stomach for it, it can definitely save you a ton of money.  There's a lot of risk, so I don't advise anyone to try this at home.
  • kellylipp
    kellylipp Registered Users Posts: 21 ✭✭
    I just upgraded to BattleBorn batteries...

    Nick,

    How are your BattleBorn batteries working out for you?

    Thanks,
    Kelly

  • nickdearing88
    nickdearing88 Registered Users Posts: 100 ✭✭✭
    kellylipp said:
    I just upgraded to BattleBorn batteries...

    Nick,

    How are your BattleBorn batteries working out for you?

    Thanks,
    Kelly

    Great, I'm very happy with them.

    The system was down for about a month (Sept 15 - Oct 14) because I moved to the new property. During that time, I disconnected the solar and turned off the charge controller and inverter breakers. The batteries were in the garage, in mild temperatures, about 60F average.

    I tested capacity to check self-discharge. By my rough estimate, there was very little to no capacity loss --- I'd say less than 1 percent.
    Current system: 8-100w Renogy panels mono/poly, 2 strings of 4 panels in series - 24v 100Ah AGM Battleborn LiFePO4 batteries - Morningstar MPPT40 CC - 1500W Samlex PSW inverter
  • Jacob_U
    Jacob_U Registered Users Posts: 5 ✭✭
    Hi all,
    I've got an existing Outback Radian 8KW power center with 1150AH Gel batteries.  I'm looking to replace these batteries with 700AH LiFePo.  I have a line on some Chinese made batteries from Eastar battery company.  The batteries are rated 1C discharge and 0.5C Charge, which is plenty sufficient for my application.  The batteries come with a BMS.  The picture of the battery case (each case contains 200AH of batteries) shows two sets of terminals; 1 pair for charge and 1 pair for discharge.  My Outback power center only has a single pair of wires to the existing Gel battery bank.
    Question: Will this pose a problem connecting the charge and discharge to the Outback?  Will i need to bypass the MFG BMS and connect directly to the batteries?
    BTW: the price that I've been quoted is tough to resist.  700AH 48V delivered for $10K.
    Thanks for the help.
    Hi,

    Did your 48v 700ah work with Outback already?
  • Ampster
    Ampster Registered Users Posts: 174 ✭✭✭
    @bear_essentials, I am interested how they worked out as well. $336 /kWhr is a good value for Lithium.
    8 kW Enphase micros AC coupled to a SolArk 12K
  • Raj174
    Raj174 Solar Expert Posts: 795 ✭✭✭✭
    @Ampster,
    200 and 400 amp hour CALB LFP cells are available for 119 to 126 and 259 to 249 dollars per cell respectfully, depending on quantity. I know 2 people who have bought and installed them. In fact, I installed one of those, a set of 16 400AH, for a friend. Here's the link if interested.

    https://www.alibaba.com/product-detail/3-2V-200Ah-Prismatic-LiFePO4-Battery_60790265531.html?spm=a2700.wholesale.pronpeci14.6.55e33d15pxql5M

    Rick
    4480W PV, MNE175DR-TR, MN Classic 150, Outback Radian GS4048A, Mate3, 51.2V 360AH nominal LiFePO4, Kohler Pro 5.2E genset.