Announcement

Collapse
No announcement yet.

Simplified lifepo4 charging and care

Collapse
This is a sticky topic.
X
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Simplified lifepo4 charging and care

    Lifepo4 charging and care is actually very simple. In this example we are going to use a 12v battery made up of 4 cells, aka a "4S" configuration.

    We'll assume that you have purchased a pre-built battery from a reputable manufacturer a 4S configuration of quality CALB, GBS, Winston, Sinopoly, Hi-Power large prismatic cells, and NOT some random collection of RC modeling cells. Pre-built means that they have connected up all the cells with links, and have strapped the battery together for physical stability. Each cell is typically pre-charged to about 50% SOC before shipping.

    You'll need to babysit the very first charge, and accordingly have a voltmeter of decent quality.

    1) Upon receipt, measure the voltage of each cell. They should be relatively close to each other, somewhere around 3.2v. What we are looking for here are wide swings in voltage, like one cell reading 2v, and another reading 3.4v. Either that low cell is bad, or it just didn't get charged properly before shipping. Anything below about 2.6v upon receipt is a red flag. Having cells that read close in voltage upon receipt does NOT mean you have a good balance since during the flat part of the discharge curve they mean almost nothing. What we are looking for are wide variations for obvious signs of dead cells, or an indication that you may have received just a random collection that was put together.

    2) Set your AC charger or solar charge controller (disable any temp comp!) anywhere from 13.8 to 14.0v. There is NO NEED to set the charge voltage any higher. Those that do are trying to drive external balancing circuits, which can be additional points of failure. If you follow these precepts here, and of course are not using used/abused trash, but quality cells, there is no need for constant balancing. But don't tell them that.

    3) Apply the charge, and just make sure that no individual cell exceeds 3.6v during charge. If the cells are high quality like those above, and have been charged to about the same amount prior to shipping, ideally each will read very near 3.5v at the end of charge. If they are slightly out of balance, they might read anywhere from about 3.45 to 3.55v.

    So what is "end of charge"? Technically, that would be when the current to each cell drops to about 0.05C. But because we don't have to worry about sulfation, there is NO NEED to ever complete a charge to full! YOU decide where end of charge is, as long as it provides the capacity you need to get the job done without going beyond the low voltage limit upon discharge. This makes lifepo4 ideal for solar where obtaining a TRUE full charge is often not met. It has no problem with partial state of charge operations, and in fact prefers it.

    There is no need to go nuts over a small SOC voltage difference like this. No cell is above 3.6v, and due to minor differences in a cell's actual manufactured capacity and internal resistance, it is not uncommon for the SOC voltages to differ a little bit. Assuming you have quality cells, you'll still be able to achieve 80% DOD without suffering any major imbalance at the bottom end.

    4) Don't over-discharge. Set your alarm for a pack voltage of 12.8v under load. Disconnect them if they reach 12.6 to 12.7v under load. (3.15 to 3.175v per cell). This will be approximately 80% DOD, and you won't be travelling too far down the steep discharge knee. Because we are not going well beyond 80% DOD, small imbalances just before the knee drop are not that big of a deal. If you want to be conservative, just use 12.8v under load as a disconnect. Remember this: as solar users, we still should only design our systems not to use more than 50% DOD even with lifepo4. So go ahead and be conservative here.

    Optional: if you feel you must, or if your cells WERE somewhat charged hastily prior to shipment, and one cell wants to rise to 3.6v or more before the charge is finished, you can easily discharge that individual cell a little bit and check it again on the next cycle.

    In the case of my 20 and 40ah GBS batteries, this was nothing more than an RV incandescent brakelight bulb fixture. I "top balanced" them just for fun, even though it wasn't strictly necessary. When one cell was higher than the others, and when OFF CHARGE, I applied about 30 to 60 seconds of discharge to that cell. The battery quickly came into line after about 2 or 3 individual cell cycle/discharge compensation attempts. BUT note that it didn't change or improve things operationally! I was still stopping at a max of about 80% DOD, so nothing was really gained.

    What I am saying here is that upon receipt, with good quality cells and a decent charge from the factory, you may never have to "balance" at all!

    Is it that simple? Yes. Of course, most will want to use automation / relays for a low-voltage disconnect, like any other serious battery installation, but here I'm just showing how easy the majority of charging and care really is. In our application of being "fractional C", that is, low voltage and low current, there is no need to go nuts over balance as long as each cell rises no higher than 3.6v at the top, and just don't let any cell go below about 3.15v during discharge. That's IT!

    Individual cell monitoring? Sure, good stuff. But I consider that excessive since I don't (actually can't!) do that with my lead-acid batteries. If I suffer a dropped cell or other signs of poor performance, I'll be taking that up with the dealer / manufacturer much as I would with lead-acid, rather than hanging a rat's nest of wiring all over the top of my batteries - or worse yet letting an external balancing system fail or mask an issue that needs to be dealt with sooner than later.

    The overarching moral to the story is that we are NOT EV'er or RC modelers. OUR needs allow us the headroom to use lower charge voltages and conservative discharge depths. If you don't keep this in mind, you can easily be led astray by by those who operate in entirely different application scenarios.

  • #2
    I nominate this thread as a sticky at some point.
    SunnyBoy 3000 US, 18 BP Solar 175B panels.

    Comment


    • #3
      Thanks!

      I'd like to add that the reason I mention the use of GBS, CALB, Winston, Sinoply, etc large lifepo4 prismatics, is that those are the types of cells we would use in a typical house-bank, and NOT smaller cylindrical cells like Headway or the usual 18650 types.

      The reason is that Headways and 18650's are HIGH-RATE cells, typically rated up to about 5-10C or more of charge discharge current. You will never do this with a solar housebank sized appropriately to get you through the night. Thus, you'd be paying more for a capability you'll never use, nevermind adding instability with large amounts of wiring infrastructure to build up capacity, each of which is an additional point of failure. Stick to the large prismatics, which themselves being rated commonly at 1C continuous with 3C or more bursts, and are still almost over the top in capability as far as our relatively low-current needs are.

      Our application could be titled "Fractional-C" usage, where one doesn't under-size the bank on purpose. This would make it easier to differentiate and identify the actual application we are in as compared to EV or RC modeling.

      Comment


      • #4
        Simple low-voltage disconnect

        I mentioned the use of 12.8v under load as roughly an 80% DOD point which you don't want to exceed. For sticklers, we know that SOC voltages are a poor indicator, but this is close enough for our usage.

        The problem is that finding a quality LVD that operates under these conditions can be tricky, or incorporated into a larger part of a bms system that you may not desire.

        I've found and tested one. Although designed for lead-acid in emergency vehicles, the Powerwerx ITS-12 (35A continuous), or the APS-12 (25A continuous) has a convenient timer that starts at 12.7v, and the timer can be adjusted within minutes to hours before shutoff. In the case of lifepo4, I'd recommend using either the 2-minute or perhaps the 15 minute timeout max. I've tested my own ITS-12 and it works fine.

        Note that the both the ITS-12 and APS-12 incorporate a hard high-voltage-disconnect at 16v. This is too high for lifepo4, and your batteries will suffer, but will prevent a catastrophe if your charge source goes berserk. The ITS-12 has a hard low-voltage-disconnect of either 11.0 to 11.4v, and again, these are too low for lifepo4, but will prevent total discharge catastrophe. The APS-12 doesn't have the hard LVD, so the timer is it for the low voltage disconnect.

        It draws about 150ma as measured on my Fluke. Shouldn't be a problem if you are using large prismatics.

        What we rely on is the timer which conveniently starts at 12.7v. Wiring is a simple 2-terminal in/out, and no need for additonal sense wires. Originally, these units started the timer because at 12.7v, it assumes the ignition is off with the typical voltage drop on a lead acid starter battery. Fortunately for us, this voltage is just about right for an 80% DOD on lifepo4.

        While not perfect for lifepo4, I give them a thumbs-up for a simple LVD based on the timer. Just don't set the timer too long, as you are already at 80% DOD at this point. Maybe in the future I could get them to readjust their hard-limit voltages to those more appropriate to lifepo4, but for now, the 12.7v timer will do.

        This is right up my alley - keeping it simple yet effective.

        Comment


        • #5
          What happens if you go below 80% DOD?

          If you do find yourself going well below 80% DOD a few things happen.

          If not taken too far, say down to 90-95%, you'll be cutting cycle life.

          If you leave it in a heavily discharged state long enough, the battery starts to eat itself, and this may or may not be accompanied by gas formation swelling. So you need to get to a recharge asap.

          Note that a swollen or bulged cell is already compromised. While you may be able to physically squeeze them back together, the reaction that caused this condition in the first place cannot be reversed. Note that the strapping and banding that is done on pre-built packs is done primarily for terminal connection stability to avoid unintentional terminal torque stress, and is not designed to prevent charge / discharge abuse.

          HOWEVER, if you value your cells, in a heavily discharged state, you'll want to limit the charge current to no more than 0.01C until the cell terminal voltages reach 3.2v. THEN you may apply the normal charge current, which is usually limited to about 1C max on most large prismatics.

          So while some may do the right thing by getting a charge back on it asap, they can do additional damage by applying too much current in a heavily discharged state. You may not immediately see the effects, but it is accumulative.

          Moral - don't go much below 80% DOD if you don't have to.

          Reminder - as solar users, we don't normally design our banks to go down to 80% DOD anyway, with only 50% DOD typical, to give us some headroom during bad weather or unexpected loads.

          Comment


          • #6
            Originally posted by PNjunction View Post
            So what is "end of charge"? Technically, that would be when the current to each cell drops to about 0.05C. But because we don't have to worry about sulfation, there is NO NEED to ever complete a charge to full! YOU decide where end of charge is, as long as it provides the capacity you need to get the job done without going beyond the low voltage limit upon discharge. This makes lifepo4 ideal for solar where obtaining a TRUE full charge is often not met. It has no problem with partial state of charge operations, and in fact prefers it.
            OK I have two concerns on a technical and economical issues. First one I quoted because it is missing vital information IMO and vague. There are 2 basic ways a LFP battery can be charged. I call them Good and Better. Technically the first one which you use in your example is FORCED charging. Forced charging is really nothing more than a Constant Current mode in which you charge at a fixed current until you reach a set point voltage and terminate. At best it can only take you up to 90 to 95% SOC, and only if the set point is set for 3.6 volts or 14.4 volts. What is good about this I just basically described the BULK phase of about every solar charge controller out there. While I think the lower voltages are fine that you mentioned will be a significantly lower SOC. Not a problem on battery cycle life, but might leave you short on power when there is no harm going higher.

            The second way which is a bit better and how you get from 95 to 100% is using a 2-Stage Algorithm. First is exactly the same Constant Current (CC) or Bulk Mode to 14.4 volts, then you switch to Constant Voltage - Current Taper aka Absorb Mode of about every solar controller out there. In this mode, initially it is a CC up until the voltage set point of 3.65 volts. Once the battery reaches the voltage set point the charger holds 3.65 volts until the current tapers down to 3 to 5% of C of the battery. So for a 100 AH battery 3 to 5 amps @ 3.65 volts.

            The second issue I have is Battery Management System or BMS being discounted. For a really small set of batteries under $1000, no problem not using a BMS. Just check balance from time to time and Bottom Balance when needed. $1000 may sound like a lot of battery, but not for LFP. $1000 worth is about a 2Kwh battery vz a $1000 6 Kwh FLA battery. LFP is about 300% higher in cost. Really talking about a small 12 volt system for an RV or campsite. For a home system operating on much larger batteries and the investment really jumps quick. A BMS is well worth protecting that kind of investment.
            MSEE, PE

            Comment


            • #7
              Originally posted by Sunking View Post
              OK I have two concerns on a technical and economical issues. First one I quoted because it is missing vital information IMO and vague. There are 2 basic ways a LFP battery can be charged. I call them Good and Better. Technically the first one which you use in your example is FORCED charging. Forced charging is really nothing more than a Constant Current mode in which you charge at a fixed current until you reach a set point voltage and terminate. At best it can only take you up to 90 to 95% SOC, and only if the set point is set for 3.6 volts or 14.4 volts. What is good about this I just basically described the BULK phase of about every solar charge controller out there. While I think the lower voltages are fine that you mentioned will be a significantly lower SOC. Not a problem on battery cycle life, but might leave you short on power when there is no harm going higher.

              The second way which is a bit better and how you get from 95 to 100% is using a 2-Stage Algorithm. First is exactly the same Constant Current (CC) or Bulk Mode to 14.4 volts, then you switch to Constant Voltage - Current Taper aka Absorb Mode of about every solar controller out there. In this mode, initially it is a CC up until the voltage set point of 3.65 volts. Once the battery reaches the voltage set point the charger holds 3.65 volts until the current tapers down to 3 to 5% of C of the battery. So for a 100 AH battery 3 to 5 amps @ 3.65 volts.
              And then there is the group of Li battery enthusiasts who feel that because of the harmful effect on overall life (as opposed to cycle life) of sitting at or near 100% charge, it is not a good idea to routinely take your LiFePO4 battery to 100% charge. Instead they take something between 80% and 90% as the new "full charge" point and operate their battery bank that way. It means that you will be using less than the full energy capacity of the batteries, but perhaps end up with them lasting much longer. Unlike Lead Acid where never getting back to 100% will lead inexorably to sulfation and loss of capacity, Li chemistry batteries do not have this mechanism for deterioration.
              I think that PNjunction may be basing his design on this strategy.
              SunnyBoy 3000 US, 18 BP Solar 175B panels.

              Comment


              • #8
                I'm actually very glad you brought up these issues!

                What is good about this I just basically described the BULK phase of about every solar charge controller out there. While I think the lower voltages are fine that you mentioned will be a significantly lower SOC. Not a problem on battery cycle life, but might leave you short on power when there is no harm going higher.
                We won't be short on power if we design our banks just like we do with lead-acid - that is to about 50% DOD. Plenty of room to spare on the top. Thus there is no need to reach high voltages (above 14v, or more than 3.5v per cell). The other issue is time. How much time you spend at a higher than necessary voltage is what contributes to electrolyte oxidation and other parasitic reactions. If one tries to finish absorb, you are spending more time at an elevated voltage - even if it is within spec - than necessary. We can afford to be conservative be it CC or CC/CV.

                It's not so much the actual act of charge / discharge (de-intercolation) that kills cells, but the parasitic reactions, along with heat that nail us. While we can't do anything about the SEI layer and the manufacturer's choice of electrolyte additives, we can contribute to longevity by limiting voltage exposure creating parasitic reactions at the very top and very bottom of charge. Since there is no need to go there, why? This is what is lost on the "drop in replacement" crowd.

                The second issue I have is Battery Management System or BMS being discounted.
                I think the problem here is that most consider a bms as only a balancer, and not one that includes an LVD or HVD, which would be appropriate for any battery chemistry.

                My contention is that in a fractional-C application like ours, with conservative charge/discharge, you'll always be for the most part in the flat part of the power curve, so there is no need to obsess over minor balance differences. Of course cells needs to be *reasonably* close to equal SOC's, and of good quality to begin with. So yes, an LVD and an HVD are proper. One may put their faith in their CC or AC charger as the HVD instead of relying on an external unit but bms manufacturer's basically saying that lifepo4 won't properly charge without one are wrong.

                What I'm concentrating on is the hype surrounding "balancing" in a fractional-C application like ours. Sure, go ahead and invest in a bms, but just know that with most of them, you don't HAVE to activate the balancing function, which tends to keep cells artificially high in elevated voltages for periods of time longer than necessary. And of course, are a possible additional point of failure in and of themselves.

                What I present here is not touted as the ONLY way to deal with lifepo4 - just one of the alternatives that works. It may not be for everyone, especially an unknowledgeable consumer market, but those who come here aren't that sales demographic.

                Comment


                • #9
                  LiFePo4 economics

                  NOTHING ruins a good thread more than the discussion of lifepo4 economics!

                  What I will say is that lifepo4, like going solar off grid itself, is about 10 times more costly than what you would be charged from your local power company. You go off-grid solar for reasons other than saving money, and at some point, the whole issue of battery chemistry becomes moot!

                  Likewise, the use of lifepo4 vs other battery chemistry is done for reasons other than pure economics, like the ability to thrive in a partial-state-of-charge environment. Although costs have been coming down, they haven't come down as far as we would like, and it becomes a favorite target for the lead vs lifepo4 crowd. Don't go there - and no need to pick sides. I have a very heavy chunk of lead right next to my lifepo4's for different reasons.

                  We can argue all day about "long term" investments in cycle life, DOD's, the lack of 100 years experience, no major lead-acid players in the game so far, dramatic newsbytes about batteries that are li-ion but NOT lifepo4 after all, and entirely different user applications than what we do, the guy next door with a bad power drill battery, ebikes that won't run, nissan, tesla, boeing - man I've seen it all. All that does is raise the signal to noise ratio so high that nothing is of value.

                  I'm just pleading that we don't go there, but stick to the technical side of things, and let the user make up his/her own mind if lifepo4 is a good fit or not for their needs. In many cases, it doesn't! I really dislike unjustified lifepo4 cheerleading as much as I do nay-saying especially when either party has no first-hand experience with these large prismatics.

                  Ok, rant over.

                  Comment


                  • #10
                    Well here is my first Lithium pack I had in my racing golf cart V1.

                    lithium bat.jpg



                    Here is V2

                    lith #2.jpg
                    MSEE, PE

                    Comment


                    • #11
                      Originally posted by Sunking View Post
                      Well here is my first Lithium pack I had in my racing golf cart V1.
                      So that PCB across each cell in the first picture is some sort of balancing/OV protective circuit?
                      Not quite the same as a full-fledged battery management system, but still more than you later felt you needed?
                      SunnyBoy 3000 US, 18 BP Solar 175B panels.

                      Comment


                      • #12
                        I *knew* Sunking was holding back on us.

                        That is a very nice collection of Sinopoly and GBS cells. Sunking - make me feel better by putting the purple covers back on those blue GBS cells after the shoot! They look like a 48 volt battery, but wonder if you have more on the other side of the vehicle...

                        This is an excellent example of what is not our application. High voltage - and high current. Notice too all the additional interconnects to obtain this. The mantra for solar house bank users is to *reduce* possible points of failure.

                        In a simple 4S (12v) application of only 4 cells, (or perhaps 8S for you 24v house-bank users) and by not taking the voltages higher than 13.8 or 14.0v, you can rid yourself of any additional points of failure strapped across the terminals. Because there is no need to make it home on the last ounce of power, or accelerate up a freeway ramp pulling hundreds of amps, we don't need to go to these extremes.

                        I'll let Sunking speak for himself, but I personally do not want any of those balancers to go bad and actually unbalance a good pack. However, those that feel they need this, are under no obligation to leave them connected once they have finished balancing. Ie, if you feel better about it, one option would be to drive these thing up to 14.5 - 14.6v, (about 3.6v or so per cell) which is necessary for them to start doing a discharge, and once all the blinky lights go out, you *could* remove them, and merely operate at a lower overall voltage afterwards.

                        I'm not comfortable with a $5 to $8 dollar board, and would have to verify each and every voltage trigger for proper operation, and then hope that they stay within specifications (typically burning .5 to 1A each depending on cell size, about 1ma when off, voltage is *supposed* to be 3.55v ), during the lifetime of the bank. If one burns up and latches into a constant discharge, that is not a risk I'm going to take when all I have to do is run a lower overall bank voltage. I'm also not sure of RFI confusion. While I think that the battery would absorb most of it, it does make me wonder if they have tested them near hi power radios / broadcast towers and the like across the full spectrum from powerhouse FM broadcasters, NOAA weather radio stations, cell towers, heck even nearby cellphones. The boards on their own may pass on the bench, but has anyone tested them with all the "antenna" wiring (dc infrastructure) attached?

                        The point to remember here is that we are not operating under EV conditions. And some EV'ers won't put these on their banks either. But here too, I hope we keep the EV application issues to the EV forums.

                        To each his own, but once something like this is introduced, we get sidetracked into EV and RC modeling needs and that's the death-knell for house-bank users.

                        Comment


                        • #13
                          Mounting cells upright

                          The BEST position to mount these cells in are with the terminals facing up.

                          Facing down is prohibited. This is because the electrolyte (what little there is) can seep down and clog the safety vents.

                          Sideways is also not recommended - however some seem to get away with it, but only with the manufacturer's approval. Here, sideways or flat mounting may starve the topmost anode or cathodes of electrolyte given enough time, and may also interfere with the safety vent.

                          You may notice that upon receipt, the box may indicate "this direction up" or something like it. Those are not only shipping instructions.

                          I don't have any figures on how long it takes for the electrolyte to equalize again in a cell not stored upright. It could be seconds, minutes, or hours. I don't know. But what I wouldn't do is cut open your prize, and immediately put it into use, since you have no idea of how it is stored in the delivery truck on the way to you.

                          Comment


                          • #14
                            Don't cheap-out!

                            Yeah, I too would like to see the prices of lifepo4 prismatics come down. Just don't be tempted to go with fly-by-night or no-name manufacturers just to save a buck here and there. Personally, I'd stick to GBS, CALB, Winston, and perhaps Sinopoly or Hi-Power.

                            Why? Aside from possibly of using poor quality chemicals, and maybe shoddy exterior construction, little concern may be placed upon the precision of how it's made INSIDE, and that's important.

                            What I'm talking about is precision alignment and size of the materials themselves. Essentially the cheap ones may have little knowledge of what is known as "overhang", or purposely making one set of plates if you will, larger than the other. This is to help prevent dendrite formation. They also need to be aligned inside. Some of the counterfeiters of the small cylindrical 18650's may cut corners in just this fashion - leaving you with a dendrite short later.

                            This is also the reason that once you have a bulged / swollen cell, and somebody thinks that all you have to do is squeeze it back together again is wrong. Sure, you may be able to do so physically, and the cell may even appear to operate ok - for the time being. BUT, you may have lost the internal alignment, and now you face the great possibility of future dendrite growth, and an eventual internal short.

                            Comment


                            • #15
                              China Aviation Lithium Batteries

                              Just checked the fabricator of our batteries, Manzanita Micro. They primarily produce battery packs/banks for EVs. The cells are China Aviation Lithium Batteries (CALB). It turns out that CALB makes Manzanita small.jpg lithium battery banks of up to 200 kW-hours. Attached is photo of a battery (4 cells) similar to ours. "...Available in 180Ah packs. Includes: 4 Lithium Ion Cells, Aluminum Battery Box, Lexan Top, Fully Assembled Reg Deck, All necessary hardware, and Regulator..The battery box is 8" wide, 12.5" long, 13.5" tall (without lugs). Total weight with 180Ah cells is 63 lbs/ 28.5 kg."

                              Mazanita Micro designs for the high power demands of EV but work well with the lower requirements of RV living. We have four of these in series for a 48 V (nominal) system. We do have room for another two batteries and 700 W more on roof, but this would cost quite a bit more, require redesign and configuration, and we can see no reason to do such.

                              Currently in an RV park (oh the shame and humiliation) and we are in total shade (get back about 600 kW-hours a day). Have paid attention to PNJunction and only hook up the 15 amp cord through the battery chargers when we get down -3500 W-HOURS

                              Reed and Elaine

                              Comment

                              Working...
                              X