GBS LiFePo4 20 and 40ah batts

PNjunction
PNjunction Solar Expert Posts: 762 ✭✭✭
To see what the marine guys have been going gaga over for the past few years, I decided to pick up some smaller capacity GBS LiFePo4 batteries for some test runs before going large scale. Competitors include the likes of CALB and Winston. Single-cell capacities range from 20ah to 200/400ah typically.

The basic qualities that make them attractive for solar is that there is no absorb. You charge with bulk until done, which is already efficient since the internal resistance is extremely low. There is no penalty for partial state of charge usage - in fact you are rewarded for it! Like my pure-lead Odysseys, they can handle extremely high charge/discharge currents compared to FLA/Conventional agm. Anything from 0.1C to 1C is fine, although I prefer to limit the max charge current to 0.5C. Simple charging, once balanced, is easily performed and there is no temp-comp. Cost has been coming down, and is about on par with high-quality agm. In fact, both my 20 and 40ah units were either cheaper or on par with the equivalent larger Optima battery.

The review here is for the 20ah GBS-LFP20AHA and the 40ah battery GBS-LFMP40AH. Both consist of 4 nominal 3.2v prismatic lifepo4 cells in a series configuration. They were pre-assembled and have either a plastic casing, or are strapped with aluminum side panels. Cell-links were already installed by the factory, so I didn't have to worry about over-torquing them. These cells are NOT constructed out of large quantities of smaller cylindrical cells. I opted out of getting the individual board balancers, because I can do it myself, and am not running a high-voltage, high-current EV application.

The box for the 20ah version comes in a plastic box and you can remove the top. Cutouts are provided on each side for you to attach your own lugs to the positive and negative terminal. Be gentle opening the top and the cutouts since the plastic is kind of soft, yet brittle. Full access to the individual cells is easy, but BE CAREFUL. There is a lot exposure there, and the slip of a wrench could be a major problem. Just use basic caution like you would with any battery.

The 40ah version has the strapping and aluminum sideplates instead of a box. On the top are individual snap-on covers for each cell. Keep them on when in use. However, I found that if individual cell maintenance was needed, only pulling the cover for the cell you need was a nice safety feature. This is the basic style that anything 40ah or larger uses. Cells are aquamarine, and the purple tops makes me think of high end graphic unix servers, or cool bicycle shops in the 90's. :)

The 40ah version has a different style of cell interlink, with 4 screws instead of just one big lug like the smaller version. These screws can be either phillips, hex or torx. Mine were hex. Unfortunately, one of my hex screws was already stripped/rounded by somebody down the line. Not a showstopper as I fixed it with pliers, but maybe the guy on the line should have slowed down a bit and abided by the manufacturer's torque specs.

To make life simple, and make sure I was reasonably balanced, I just used a 6A "single cell" 3.7v lifepo4 charger for my initial charge on each cell. (DO NOT CONFUSE with other lipo single cell chargers with a different voltage!). I charged to 3.55v, and removed it, and repeated that for each cell. Having it set for 3.7v, meant that if I got distracted, and let it finish, I wouldn't destroy the cell.

Of course, one can always use an adjustable bench charger. Or a hobby charger. Or whatever you come up with. Just don't let any cell get higher than 3.655v. In use, because we are not EV or RC modelers, a balance can consist of neither the highest nor lowest cell exceeding 0.050V (50 millivolts) apart from each other. With the single charger, you should be right on the money, but don't be surprised if the cells "walk around" just a little bit when new. After the initial balance charge, for normal use inside, I use a Samlex SEC 1215UL charger set for 14v. (reviewed elsewhere here) Outside I run a gamut of panels and CC's, also set to 14v, no float (or really low like 13.5v), and no temp-comp. Someday I'll consider a Genasun if they offer a user selectable voltage, but until then, my present CC's do fine.

The first time you do this, remember that these are wired in series, and you'll have to be CAREFUL about the polarity in which you attach the single cell charger. Almost zapped mine forgetting that the polarities alternate in a pack like this.

Speaking of which, in both the 20 and 40ah versions, yes there are LITTLE polarity marks on the edges of each cell. I made mine a little uglier by using a red and black sharpie and dotted the edge to keep me a little more alert.

I got both the preassembled batteries, and the single cell charger from a specialty battery company:

batteryspace

But, there are many other ways to get these, namely from EV parts companies. Just a FEW off the top of my head:

elitepowersolutions
evsource
ev-power (Australia)
electricautosports (Canada)
gwl/power (europe)

A search for EV parts should turn up dozens. But like any battery supplier, try to find one reputable. Get ALL the parts needed, like the cell-links and end-terminals, preferably already strapped. Don't kludge this stuff together. You don't want the major mechanical link holding 4 cells together being just the cell interlinks. Or go pre-assembled like I did. You may NOT want the so-called "drop in replacements" that are sealed and do not allow access to the individual cells. While they may be made of closely matched cells with an internal balancer or bms, if you don't want that, then no need.

Operate these things vertically with the terminals up. You may notice that they ship with stickers pointing "this end up". There is electrolyte in them, as small as it is, and you don't want it clogging the vents, or starving the plates on top if you were to mount them flat.

SOLAR:
I charge just as simply. 14 to about 14.2v. Disable temp-comp. In fact, Genasun uses 14.2v for their lifepo4 specific mppt controllers, unless you pay for reprogramming. No need to float at all. So you can get by with an existing CC, provided you can disable temp-comp, and select a voltage low enough. If your present low-end CC has a "gel" setting, see if that voltage is truly low, like 14 to 14.2v. Personally, I balance at 3.55v and charge at 14.0v. Essentially, a fully charged prismatic cell is one that has reached 3.40 - 3.45v. A little oversimplified, but fodder for some good forum reading!

Speaking of which, you may see online references to charging up to 4.2v for an initial "forming" charge - this is OLD OLD info, and not something any of us do today. The general trend for house banks anyway is not to live on the edge of specs like EV'ers do.

MISC: There is a LOT more not covered in this review of the battery. Some may like having "insurance" with some sort of automated balancer. Not me, but to each his own. Like any normal battery system, you'll want an LVC. That can be either a little planning, or monitoring and lots of hardware. Your choice. My choice is to set the LVC for no lower than 12v under load. Spec sheets from manufacturers may be tweaked by the marketing department. I like not to cut things so close. :)

Overall: These little versions gave me the confidence I needed in case I ever decide to build a huge bank. The largest cells are 200ah, but even larger capacities are seen with various manufacturers. Hint: if you do, parallel the batteries first to build up your capacity, and THEN series connect these groups to get your voltage. If you do it the other way, like we do with standard 12v batteries because we have no other choice, then balance will become problematic.

So far, I'm impressed. Once I got the distracting FUD out of the way demanding that I use a lot of external hardware beyond what any normal battery installation calls for, I felt relieved. Once again, others may not feel the way I do, but you can prove it to yourself with a small setup.
«13

Comments

  • john p
    john p Solar Expert Posts: 814 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Small cells of about 10 to 20 ahr are the way to go.. Extensive testing of hundreds of different sizes and makes over the last 4 yrs every day of the week. has proved this beyond doubt.... Why you think all major vehicle manufacturers use small cells.??
    Small cells can be balanced with nothing more than a zenner diode and a resistor on each cell
    Your cells will self destruct if taken below 2.2v. DONT DO IT EVEN ONCE .. or you will be buying many replacement cells.
    Its not an option its really mandatory to have an automatic permantly connected Low Voltage Cut off switch.
    There are many PWM chargers that can be easy adjusted to charge these cells just fine.. Look for some of the old Stecas or even copies of them. They are very reliable and cheap.
  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    Like any normal battery system, you'll want an LVC

    What's an LVC ? Being I've got an abnormal system and didn't get one with it.
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    john p wrote: »
    Why you think all major vehicle manufacturers use small cells.??

    Because they need to fit hundreds of them into small spaces. We don't fortunately. Using small cylindrical cells (remember, I am a fan of A123 cells for other purposes), means a lot of intercell connections that can go wrong. It also adds to the overall complexity, which is something I don't need in house-bank solar storage application. 4 Cells in series is all I need. For now. :)
    Small cells can be balanced with nothing more than a zenner diode and a resistor on each cell
    Perhaps in a flashlight application. Here, we are dealing with enough current and basically power, that it can overwhelm a low-quality blinky led board or shunt/diverter if care is not taken.
    Your cells will self destruct if taken below 2.2v. DONT DO IT EVEN ONCE .. or you will be buying many replacement cells.
    Actually, that is waaay too low for LiFePo4, 2.8 is a more real-world limit, but that is still extreme. Better to stop when you enter the discharge knee at about 3.2 to maybe 3v under load.

    What many don't realize is that when you take a cell too far, two things happen. 1) If left long enough, the cell degrades badly. 2) The subsequent recharge is what really kills it, and makes it swell ruining the cell. Far too many take their cells too low, but then apply the normal current back into it, causing major problems. You can slowly nurse a de-lithiated cell back to life, but we are talking using about .001/C currents. Even then, the cell may be compromised.

    So yes, using an LVC is very wise, just like you would with ANY battery chemistry. Proper planning, and not setting your lvc too low can go a long way.
    There are many PWM chargers that can be easy adjusted to charge these cells just fine.. Look for some of the old Stecas or even copies of them. They are very reliable and cheap.

    Yep - but be SURE to disable temp comp. Some morningstars you can't. Even so, an MPPT which has most of it's advantages in the bulk stage, is a nice fit for LiFePo4, which basically runs in nothing but a bulk stage until done.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    mike95490 wrote: »
    What's an LVC ? Being I've got an abnormal system and didn't get one with it.

    Low Voltage Cutoff.<grin>

    These are for those times when you have miscalcuated your needed capacity, or have left a device running too long without some other warning. Which is actually easy to do compared to lead. Since the charge/discharge curve is so flat, your 12v gear will just die immediately. With lead, you had the dimming of panels, bad audio, etc as a warning device.

    Most EV vendors can provide an LVC. Some have taken to using an LVC which monitors individual cells and will kick off on the one that gets there first. Very smart. Provided it actually shuts off the load. If it is a mere alarm, then there had better be someone around to hear it. :)

    For my use, I'm just using them like I do with my agm's, which don't have individual cell monitoring, nor have any way of balancing, unless you do an EQ, which most don't allow for. I just know my current consumption, adjust my time accordingly, and oversize a little bit. With these small packs, I'm taking a risk.

    For a serious installation, like with any chemistry, by all means incorporate an LVC. If you buy quality, plan capacity properly, and start out balanced, then you may get by with a pack-level voltage trigger. Some feel more comfortable with individual cell triggers, but they are usually EV'ers and others that run their capacity right to very edge or well into the discharge knee.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    mike95490 wrote: »
    What's an LVC ? Being I've got an abnormal system and didn't get one with it.
    That's because with your magical NiFe battery bank the LVC would only be necessary to protect the inverter, not the batteries from permanent damage. :)
    Actually, I imagine that your inverter really does have an LVC in it.
    SMA SB 3000, old BP panels.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Hah! I knew that was a trick question. :)

    What I really need is a 20ah NiFe starter's kit. Although I'm pretty sure I'll still need a dolly to move it around ...

    Seriously, Mike you've done great work and I thoroughly enjoy your hands-on experience and writing about it. Instead of mere theorizing, you put the rubber to the road by actually doing it and reporting real-world results and experiences. In fact, it was part of my inspiration for giving LiFepo4 a try in the first place, so thanks!

    I forgot to mention the lighter weight and size of LiFepo4. About 1/2 to 1/3 the weight and size, although those attributes were actually low on my list since I'm not moving my batteries around much.

    However, it does mean I can assemble and transport a much larger capacity system while portable, and even handle a much larger bank on my own as I get older. I think if I try to lift an 8D by myself, I'll be 6 feet under. :)
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Not my choice for standby / backup operations!

    While I like LiFePo4 for daily / regular use, I'm not so sure it fits very well into the standby or backup application, despite their very low self-discharge.

    For long-term storage, these should be discharged to about 40-50% SOC. In an emergency, that would mean that I'd only have about half the capacity available to start with. Not only that, but when stored like this, they should be recharged before using them again, limiting their usefullness in an emergency application. Unless of course I was ok with perhaps a very shortened cycle life, which in a backup application is not a major concern ... hmm. I'd have to think very carefully on this one to see if it is really justifiable.

    I still have my pure-lead agm's fully charged and ready to go for outages for this very reason.
  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    so LVC = LVD I'd always heard Lo Voltage DISCONNECT fair enough.
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    You are right, LVD is the better term.

    Speaking of which, I'm running the small battery down on a West Mountain Radio CBA-IV analyzer at 0.25C to see where it lands. It will be double-checked with the Fluke 87V, since there is a known difference in voltage on these units by about .2 volts or so, measured at the battery terminals, which can kind of skew the results if that is not taken into account. More to come!
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    This doesn't have to be rocket science!

    Well, the results are in and life is GOOD. Most importantly, it proves that with a storage application like ours, there is NO NEED to over-complicate the process! GBS gave me exactly what I paid for in capacity too without having to go to extremes.

    Top balancing the cells for a single initial balance with the individual-cell charger to 3.55v was enough. And that is just REACHING that point and stopping the charge. No absorbing down to zero current. Just pull the leads at 3.55v. Simple. Each time I checked after about 30 continuous cycles, on the Samlex charger charging them as a normal series connected pack to 14v, on both the 20ah and 40ah battery, my balance between highest and lowest voltage cell is no more than 0.025v. EVERY TIME.

    But will I get 16ah usable energy from the 20ah rated battery? (about 80%) The answer is yes. I am charging only to 14.0v (3.5v per cell) from the Samlex. I am using a CBA-IV to do the discharge at 0.25C (5A). Would you be discharging at this rate doing solar on a house bank? Probably not! The CBA-IV was also used to measure the Ah withdrawn over time. BUT, knowing the voltage is offset on the CBA-IV by about 0.23v, the Fluke 87V was relied upon to be the calibrated tool for determining voltage for both total pack and individual cells.

    Since I didn't obsess over top balance with microprocessor precision, shunt balance boards, ems systems and a rat's nest of wires all over the place, will I fall off the edge of the earth when I reach my goal? A resolute NO.

    Discharge down to 16ah depleted (still under load while testing)
    Total pack voltage: 12.76
    Cell1 3.187v
    Cell2 3.189v
    Cell3 3.200v
    Cell4 3.192v

    Doesn't look like any unbalanced sea-monsters are headed my way.

    Paused test and let it rest for 10 minutes:
    Total pack voltage (no load) 12.98v
    Cell1 3.248v
    Cell2 3.241v
    Cell3 3.245v
    Cell4 3.248v

    Even though I got enough out of the battery that I planned to, I took it further, this time to 12.6v under load. An additional 1.476ah was added to the 16 I already had.
    Total Pack voltage 12.6v (under load)
    Cell1 3.149
    Cell2 3.150
    Cell3 3.163
    Cell4 3.145

    After a 5 minute rest, the total pack voltage was 12.8v.

    That is as far as I need to go. This reiterates that there need be no fear of having to do nasa-like maintenance on these things. Unfortunately, the only major user reports we get are from people who use them in high-voltage, high-current motive applications, like EV's and Ebikes. RC modelers also do high-current motive applications, but they are using a different and unstable li-ion chemistry. We on the other hand, are low-current, low-voltage, and these prismatic lifepo4's will lead a sheltered life.

    If you have an EV, Ebike, or are a DIY using duct-taped reject/counterfeit lifepo4 TRASH, then yes, by all means overcomplicate it for your own safety. This is not meant to offend, as I have gotten great support from some of these users.

    Even if this weren't enough, for those that fear the bottom-balancing sea monsters, you'd be a fool to design your battery capacity down to 80% or more in solar anyway, leaving you no headroom for less than picture-perfect conditions. On the other hand, if you made a mistake and oversized your bank's capacity, and never reach a full charge, there is no penalty for not doing so as long as you have enough capacity not to run into the knees or the LVD. This PSOC type of operation may make the bank last far longer than expected, but I have not had enough time to witness that!

    Still, I would *occasionally* fully charge with an external source and see if the balance has gotten any worse than 0.050v (50 millivolt) from highest to lowest. If so, discharge the bank a bit, and put the individual cell-charger on them, stopping again when 3.550 volts is reached. Depending on how quick you can pull the leads off, you may be able to get better than the typical 0.025v spread than I did.

    It really is simple for folks like us using them for storage, and not motive power. The clue is to ALWAYS keep your application in mind, which may differ greatly from others using these batteries.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Ok, now that the good-news rant is over, there needs to be a caution in case you missed it.

    Like any good installation of any chemistry, an HVD and LVD are desirable. Since I'm using quality and not trash, I feel no need for a cell-level monitoring environment and take my chances by planning and generally not leaving the batteries unattended. I'm comfortable enough to leave that micro-precision to the EV and RC modelers. Pack-level monitoring is fine for me, but do what you want.

    For me, an early *warning* would be issued when the battery under load has reached 12.8 volts. THIS is where I plan to stop if I can.

    The actual LVD for me, would be 12.5 to 12.6v under load. Of course this could be fine-tuned depending on actual load. Still, this is a worst-case limit, and not one that I would use as a place to just hit the wall daily.

    IF for some reason, you dragged the cells down too low, say under 12.5v (or 12.8v rested), then do NOT apply a full current recharge. Apply no more than .01C current until the voltage under charge has reached 12.8v, and THEN you may continue charging at the normal rate. If you don't, your cells and your wallet may expand.

    Tip: If you read the forums, and run into horror-stories, what they usually don't tell you until about 300 messages down the line, is that they DID abuse their cells in the beginning and are now slaves to external balance systems. Sad that some don't mention that up front!

    While this review was mainly for GBS cells, I would have no problem using these very same charging and monitoring techniques on a CALB or Winston.
  • john p
    john p Solar Expert Posts: 814 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Pn junction.. Using small cells in vehicles the main reason isnt to fit them into small spaces.. Have a look at any manufacturers battery packs they are all on the lage size when assembled. In fact if you look at the battery packs carefully you can see it would be easier to use 200 to 400 ahr cells..Like I said after over 4 yrs doing every experiment just about possible. Ihave come to understand using cells about 10ahr makes life a lot easier, despite having many connections.. As they are low current connections very few problems arise.
    After building and installing 6 systems about 4 yrs ago for home security lighting for families that have husbands working absent and only housewife to look after system which they are not able to do as no electrical knowledge, Have not had one failure. All use 10 ahr cells. They are good quality units a bolt connection each end of cylinder.. I use a simple zenner 1w and 1w resistor across each cell. very simple.. No wiring at all running everywhere. The voltage differences are very lowso very little power is disipated in the diode and resistor.. The fact of never a failure proves my point.. And no circuit boards or blinking lights.
    The low voltage cut of is set to 2.4v and has no problem at that figure. Again it was determined to be ok after much extensive testing
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    john p wrote: »
    Pn junction.. Using small cells in vehicles the main reason isnt to fit them into small spaces.. Have a look at any manufacturers battery packs they are all on the lage size when assembled. In fact if you look at the battery packs carefully you can see it would be easier to use 200 to 400 ahr cells

    I'm glad you brought that up. The small cells used in those packs, even when assembled into larger units, are not necessary for house-bank duty. Those cells are designed for motive-power, and can withstand 10C or possibly more constant-currents for accelleration and regenerative feedback. We on the other hand, are a storage application, and don't need 10C current capability. You'll be paying a LOT more for a capability we don't use, (unless you get hoodwinked into buying hobbiest trash) and unnecessary complexity for the average guy.

    You'll even see these multi-multi small-cell units on the market as "drop in replacements" with their own internal bms - sometimes. The price of those at electronic specialty stores are nearly double-or more than these large single-cell prismatics! You are paying for a current capability feature you won't use in our application. The average consumer isn't even aware of that, and wrongly assumes that ALL lifepo4 batteries are just too dang expensive. If they are not using cells like A123 inside, then they may even be using TRASH, yet still charging you for high-quality prices!

    If you REALLY want to blow your money, get a powersports battery, like an Antigravity (which uses high-quality A123 cells inside), and use it to power trivial low-current stuff instead of using it for SLI duty.

    Case in point - when I was first starting out, I purchased the largest Shorai powersports lifepo4 battery I could find. Since I'm not using it for starting duty, that little 12ah battery (real amperage, not the "pbeq" they use for starting purposes) cost about $350. I could get a 60ah prismatic or more for that price! (Lurkers, that's 48ah usable with lifepo4 down to 80% DOD at 2000 cycles, vs needing a 96Ah lead-acid taken down to only 50% dod at possibly 1000 cycles with decent care. Don't compare prices directly across capacities without taking that into account.)

    Basically I paid through the nose for cells that were not truly designed for my low-current application. High-quality - yes! Waste of money on features that I didn't use? Yes because they are designed for sli / motive-power, and not storage-power.

    Aside from that, no hobbiest really has the RIGHT tools to properly and safely put all these little interconnects together. Witness the DIY duct-tape ebike batteries smoking, despite zener balance and other hobbiest measures. Your skills are probably very great, no doubt, but the point here is to move beyond the hobbiest environment, and into the off-the-shelf, take it home, balance it, and get on with your life for the average guy.

    This is kind of what I'm talking about - the guy who wants to put together a house bank of lifepo4 is first met by well intentioned others coming from a different application, be it commercial or hobbiest, and can be put off by the seeming complexity of it all.

    4 prismatic cells of the capacity you desire, wired in series for 12v nominal. Balance. Charge to 14v in normal cycle use. Don't drop below 12.5 under load, preferably stop at 12.8v. Done.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Use a high-quality voltmeter as your standard

    For those of you who wish to maintain a simple 4S housebank with nothing but a voltmeter, with lifepo4 you really should invest in high-quality. While we don't need lab-precision, it should be better than average because not only are we dealing very small percentages of voltage, but also something that is rarely mentioned - speed of the calculation/display! If you hit the knees, things start happening very fast, and you need a multimeter that can keep up.

    Voltmeters have been discussed and reviewed here before so I won't go deeply into the nitty gritty. Even though we have some leeway in our application as a house-bank, it pays to get as close as you can, and also as important a STANDARD you can trust and compare all the other readings from external gear against.

    My standard is a Fluke 87V. I also like the Fluke 179, but have no major experience with it. In addition to accuracy and speed of display, they have audible alerts every time a voltage changes by 1 millivolt if you use the min/max feature. At least my 87v does. This can be very handy if you are balancing or charging on the bench and get distracted. When you start entering either of the charge or discharge knees, the audible alerts will speed up catching your attention. Very valuable.

    Contrast this to the voltmeter on my Radio Shack ammeter. It is off by about .2v, and that accuracy is dependent on the voltage of the internal AA alkalines, which gets worse over time. Add to that it seems to take forever to settle down, and then bounce up and down by a 100mv or so. This is definitely not what you need when taking care of lifepo4, especially when under load.

    It really isn't tricky - just use a high-quality voltmeter that you can trust as your standard and is fast enough. Additional features like audible alerts are a plus.

    Tip: when reading about lifepo4 stories and experiences on the net, they rarely indicate how accurate their voltmeter is. "I charged to 3.6v and stopped". Did you really, or was your cheapo voltmeter lying, and were you really charging at 3.9v all the time well above the normal limit?
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: This doesn't have to be rocket science!
    PNjunction wrote: »
    Top balancing the cells for a single initial balance with the individual-cell charger to 3.55v was enough. And that is just REACHING that point and stopping the charge. No absorbing down to zero current. Just pull the leads at 3.55v. Simple. Each time I checked after about 30 continuous cycles, on the Samlex charger charging them as a normal series connected pack to 14v, on both the 20ah and 40ah battery, my balance between highest and lowest voltage cell is no more than 0.025v. EVERY TIME.

    Forgot to mention this, but this is important - you measure the balance after a period of rest, say 1 to 2 hours or so, not right after the charge. This should be familiar to us AGM users that are told by the manufacturer that taking relative soc measurements should only be done after many hours rest. Same here.

    For kicks, I tried balancing by letting the 6A single-cell charger reach 3.60v on each cell, and pulling it immediately not worrying about absorb down to zero amps, etc. When I finished the last cell, I grabbed some data from the Fluke.

    Here is the result. Read time interval / highest cell voltage / lowest cell voltage / balance (intervals are 15 mins)

    15: 3.439 - 3.396 / 0.043
    30: 3.420 - 3.393 / 0.027
    45: 3.409 - 3.390 / 0.019
    60: 3.403 - 3.387 / 0.016 <
    about 1 hour rest is all it really needed on the 20Ah GBS cell
    75: 3.401 - 3.386 / 0.015
    90: 3.399 - 3.384 / 0.015
    105: 3.398 - 3.384 /0.014
    120: 3.397 - 3.382 / 0.015

    From a full charge, cells should normally drop down to anywhere from about 3.3 - 3.4v after some rest.

    What I CANT say, is that my goofing around by rebalancing up to 3.60v is what caused my balance to improve from the normal 0.25v spread I was seeing, or if I am just seeing some settling-in improvements since I put about 10 more cycles on it before doing this. I was happy at 3.55v for balance, but I'm not going to make any proclamation here. I have no intention of writing things in stone, but 3.6v is NOT a cell voltage I'll be using under normal circumstances when I charge the battery like normal to 14.0v.

    With results like this, do I REALLY need a hobby-balance charger to obsess over getting it better to 0.007v? I don't think so.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    john p wrote: »
    Pn junction.. Using small cells in vehicles the main reason isnt to fit them into small spaces.. Have a look at any manufacturers battery packs they are all on the lage size when assembled. In fact if you look at the battery packs carefully you can see it would be easier to use 200 to 400 ahr cells..Like I said after over 4 yrs doing every experiment just about possible. Ihave come to understand using cells about 10ahr makes life a lot easier, despite having many connections.. As they are low current connections very few problems arise.
    After building and installing 6 systems about 4 yrs ago for home security lighting for families that have husbands working absent and only housewife to look after system which they are not able to do as no electrical knowledge, Have not had one failure. All use 10 ahr cells. They are good quality units a bolt connection each end of cylinder.. I use a simple zenner 1w and 1w resistor across each cell. very simple.. No wiring at all running everywhere.
    There is apparently a patent rights issue which limits many manufacturers to cells no larger than 10AH.
    Based on what I have read, Tesla uses 18650 (twice the length and slightly larger diameter than 123) cells for the best balance of cost and safety. They are easier to thermally isolate, monitor, etc. to limit the chances of a Boeing type incident as long as containment is not breached. :)
    SMA SB 3000, old BP panels.
  • john p
    john p Solar Expert Posts: 814 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    PN junction and some others most likely know this.. but to others. pulse charging the cells increases their life considerably..
    Extensive testing of all size cells over last 5 yrs ad some dcells are still under test and performing very well.
    Every cell over 100ahr that started life morte than 4 yrs ago has now either completely failed or very reduced capacity.. many of the below 20 ahr cells are still as good as new. cells are tested in group sizes . then some in each group get charged discharged at different rate.
    Many of the 10ahr cells have been charged to 3.9v and discharged to 2.4 v and still ok. same cant be said for cells over 200 ahr. even after one year. this test done automatically every day..
    Idont know about patents on sizes above 10 ahr and nobdy here at work seems to be sure of that.. Ido know a major problem with big cells is their expansion.even when stainless straps are used the still self destruct. or get so large they can be removed from their container.. not so good in a vehicle.. Even at low charge discharge rates this happens.. so be careful where and how you install in a house/building.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    inetdog wrote: »
    There is apparently a patent rights issue which limits many manufacturers to cells no larger than 10AH.
    You aren't thinking of the Ovonics patent debacle over NiMh patents are you? Runs out at the end of this year I believe, but that chemistry got strangled for so long that it is now unimportant, except for my beloved Fenix flashlights. :)
    Based on what I have read, Tesla uses 18650 (twice the length and slightly larger diameter than 123) cells for the best balance of cost and safety. They are easier to thermally isolate, monitor, etc. to limit the chances of a Boeing type incident as long as containment is not breached.

    Well, the problem with the Boeing incident that always gets raised is that it is not the same chemistry as Lifepo4, among other possible infrastructure issues. The average consumer doesn't know the difference among the li-ion chemistries, much like them not knowing the difference between an AGM and a Gel, and suffering for it. (loose analogy).

    Tesla, again a motive-power application and not a storage application, uses a special blend of li-ion chemistry of their own I believe. The important part here is that they are designed to handle super high-currents from charging and regenerative braking, a capability that we will never use, nor want to pay for!

    These things are good to know, but can be a major distraction if one is trying to decide on whether an off-the-shelf, safe, and uncomplicated string of nothing more than 4 simple cells in series will work for him in a low-current, low voltage house-bank.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    john p wrote: »
    PN junction and some others most likely know this.. but to others. pulse charging the cells increases their life considerably.

    Maybe so, but it is NOT NECESSARY to get the designed life out of these large prismatics. SIMPLE CC/CV charging is adequate. No need to get tricky. And the way we use them, well under their design limits, means good life right off the bat.

    We've gone into massive thread-drift, especially for a review thread. Why don't we take the debate over cylindrical vs prismatic to another thread of it's own. This will help those who are actually interested in using these things as a house bank to concentrate on the important factors, rather than debating the whole gamut of li-ion technology.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    40ah rebalancing for the fun of it...

    With the pack originally balanced once at 3.55v from the single cell charger, the balance as measured after a 1 hour wait was about 0.025v. Certainly liveable under the 0.050v (50mv) self-imposed limitation and I used it that way for many cycles.

    But could I somehow make it "better" by taking it to 3.600v per cell? Guess what, i did just that, and actually made things a little worse, resulting in a 0.047v balance. Still liveable, but I don't like making things worse than they were.

    My original 3.55v limit was actually ideal for the 40ah battery being charged by the 6A cell charger for the following reason - the normal rule of thumb when you are charging cells individually to establish an SOC, is to charge up to 3.6v, and when current drops to 0.05C (or C/20), you stop. This is how you do it with a bench supply. This is just an "mop" or procedure, and is NOT the voltage you want to take it to on a daily basis.

    When measured with my ammeter, that C/20 value was reached near 3.55v, and taking them up to 3.6v only served to overcharge each cell a little bit by taking it well below the C/20 current stop-point.

    Instead of going through the whole routine again by discharging it and starting over with the cell charger, I brought out the little RV incandescent bulb fixture that had a convenient switch and alligator clips put on it to manually discharge the high cells.

    So, after waiting an hour until after the LAST cell was charged, I brought out the lamp, and discharged each cell to get it very close to the lowest one which was sitting around 3.88v. This had to be repeated after waiting an additional 15 minutes since the cells tend to recover a little after discharging. No surprise as this is how my agm's act.

    Once that was done, I discharged the battery by about 20%, and recharged the whole thing with my Samlex charger, which stops at exactly 14.0v. Now for another 1 hour wait. Results:

    Cell 1 3.425v
    Cell 2 3.425
    Cell 3 3.420
    Cell 4 3.423

    So a balance of .005v ! Nice, but really, I could have been a bit more sloppy if I wanted. This is not a flight-critical RC model using another chemistry, or an EV taken to it's limits just to make sure I make it home. But I'm not complaining.

    The point here, is that you can use the KISS method of balancing and obtaining things like an RV overhead bulb around the corner at the auto parts store.

    Had this been a battery made of 200ah cells, then you'll need a headlamp, and not a little incandescent bulb unless you want to take a month away from work to accomplish this. Since these cells arrive somewhere around 50% discharged, you'd be wasting a lot of time trying to charge from a mere 6A cell charger. Either get a larger individual cell charger, like 10 or more amps, or do a bulk charge from another source, and stopping somewhere when the first cell reaches 3.3v, and letting the smaller cell charger finish the job. Do NOT do this with the batteries unattended!

    There are many ways to accomplish this, and may not even be necessary if the manufacturer did a good job for you, but it is wise to check at any rate. If they have not done a good job of matching, then typically what happens is that one cell will reach 3.6v before a total pack voltage of 13.8v is reached.

    Don't let all this verbage scare you off. Once you do it, you'll realize how much hot air I can put out. Still be SAFE and use the purple cell caps when you can.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Just in case anyone was desiring to do this with CALB cells instead of GBS, I thought I'd expand on my thought about using the battery to fit our application the best...
    PNjunction wrote: »
    Tesla, again a motive-power application and not a storage application, uses a special blend of li-ion chemistry of their own I believe. The important part here is that they are designed to handle super high-currents from charging and regenerative braking, a capability that we will never use, nor want to pay for!

    CALB currently produces three lines, the SE, the newer CA line and even more recently, the FI line. Each one is an improvement over the previous one in various areas, ie capacity, temperature, maximum current handling, etc, but for house-bank duty, none of these may be extremely important to you as you'll NEVER operate under those conditions like an EV'er would anyway. It's up to you, but I see the original SE line as the better match for our needs, unless perhaps you are sub-freezing.

    Even other brands of large prismatics that EV'ers may have found troublesome (although we don't know how well they are being treated in the beginning), may perform nicely under our rather tame requirements. But as always, try to go with the reputable dealers, and not necessarily the cheapest thing you find. You know the score - stay away from the used and abused repackaged as new.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Voltmeter update:

    ***
    (Update to the Update!) - I later decided to use a Fluke 116 designed for HVAC since it contains a built-in ambient temp sensor, and have no need for amperage measurements for simple battery voltage monitoring. I'm reaching 105F+ here in summer and needed a thermometer I could trust. I'll put the 115 to good use elsewhere.
    ***

    Added the inexpensive Fluke 115 to accompany my 87-V. The 115 is entirely suitable to the task of taking care of LiFePo4 properly, with enough precision, accuracy, and speed to do it right. There is a very very insignificant difference in the 12v range between the two, maybe 0.0005v if that.

    One difference is that the min/max function, which I use a lot when doing balancing or keeping an eye on charging to force a beep when there is a millivolt change, is that the 115 defaults to the "max" reading, whereas I prefer the way the 87-V does it by not putting it into either max/min/avg, and you choose to review those recordings manually. Not a big deal to just do another button push or two on the 115 so I can see if voltages drop OR fall in real time, yet still record them and review them later.

    From here on out, all my measurements will be coming from a Fluke 115. I'm very happy with it and can easily recommend it as a standard reference. Kind of surprising for a guy who used to hang his hat on an 8060A, and now the 87-V for many years. It may be the bottom of the Fluke line, but it is no joke for our electrical work. Put it this way - it comes down to trust. 10 years from now if someone said his readings were coming from a 115, that would be good enough for me.

    P.S. It DOES measure milliamps, that is it WILL measure below 1A but perhaps not with the accuracy you need for measuring tiny led currents precisely. In my line of work, a piece of gear pulling 652 milliamps on the 87V will read 0.651A on the 115. Close enough for electrical work.

    https://www.youtube.com/watch?v=3k9nVg03Wi4

    This guy's enthusiasm for multimeters kind of matches mine for LifePo4. :) This is for the 117, not the 115, but still very useful. Thanks Dave!

    Concerned about the protruding knob? Unless you are the type that when storing the meter, meticulously folds up the cables, twists a tie-wrap around them and put it into a sloppy toolbox (unlikely), in the real world you are just going to wrap the leads around the meter itself, providing a a small additional level of protection.

    UPDATE: Within hours I did the classic mistake, and blew the fuse. Do yourself a favor, and get a spare or two, Fluke part #803293
    Here's how it happens:

    You take a current measurement on a piece of gear, or are checking the ISC of a solar panel.
    You then go to take a voltage reading from the battery or panel, but forget to move the red test lead from the A(mp) jack to the V(olt) jack.
    Pffft. There goes the fuse. The 87V will sense the lead error and audibly warn me, but in the 115, even though it displays "LEAD" in the display when moving the rotary switch, I ignored it, and am now paying about $8 for moving too fast.

    Testing a BARE solar panel while not connected to anything else, presents an easy trap for blowing the fuse, because it is one of the few times that when taking measurements for voltage and current output, you keep the test leads across the panel wires. You take one measurement, and in your excitement dutifully swing the rotary switch without moving the test lead jacks from A to V for example, and pffft - there goes the fuse. Classic mistake. I'm getting tired of doing that. :)
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    PWM controller update:

    While not perfectly suited for the task, I did more testing on my Schneider/Xantrex C12 with the GBS cells in 97F shade conditions. At this stage of price/performance, these are what I'm going to use for the time being.

    Since I can't disable the ambient temp-comp in morningstars, they were out. And I'm not going to hang some cheap controller in front of my lifepo4's.

    Surprisingly when I set the Xantrex for 14.0v to mimick the setting I use on the Samlex ac charger, the battery actually never gets above 13.75 when nearing full charge, so it never actually hits the HVD, aka absorb set point. This threw me off at first, and of course being designed for lead, the led indicator is practically meaningless.

    What was happening is that the battery reaches a C/20 current state long before 13.8v is reached. What I believe to be happening is that even though I set the absorb point to 13.8v in the Xantrex, it may be following a lead based curve in regards to the duty cycle of the pwm, but I'm not really sure of that. So like the Samlex, it perhaps isn't taking advantage of the lifepo4 characteristics, but it does serve as some sort of HVC in any case. Either that, or the fact that a fully charged cell for rested prismatics is about 3.4v or so, although it takes just a little bit more to charge it to that level.

    What is blatantly clear is that the desire to drive these things with higher voltage, say 14.4v (3.6v per cell) is not necessary, and only done to satisfy the needs of external balancing circuits, or in an attempt to promote them as "drop in replacements". Gel's were once considered drop-in replacements and we know how well that went. Unless you absolutely need to push these things to the last drop, take it easy. 13.8 - 14.0v is just fine for normal use. If I could drop my Samlex down to 13.8v, I would, but I'll live with the fixed 14v setting for now.

    Long story short - Because the Xantrex pwm controllers allow for VERY precise adjustment of voltages (IF you use the setpoints and a voltmeter), and all undesired features are easily defeated (NO EQ, no temp-comp etc), until I make my mind up or not if I want to go Genasun MPPT, and pay for a lower voltage reprogramming, the ONLY thing allowed to touch my lifepo4's at this stage are the Xantrex pwm's. Thanks to NAWS for having a decent selection.

    Simply put, a generic recommendation is to set the Xantrex for 13.8v HVD (aka absorb for most of us). I put the float voltage at 13.8v too. Ideally it should STOP when C/20 current is reached, but they will soldier on as if it was lead. I may tweak these values later on a hair. But for now, at this price/quality range, the Xantrex pwm does the job.

    WARNING - You MUST be sure that these are NOT set for NICAD. Check the little resistor on the C12, or whatever they use on the others. That is a 2v increase over normal, and you don't want it going there by accident.

    So how will that solar charge to only 13.75 volts fare? Standby for the discharge test ...
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    20Ah GBS battery discharge test after being solar charged with Xantrex C12 set to 13.8v ..

    Temp: 90F indoors
    Voltmeter: Fluke 115
    Ah meter / discharger: West Mountain Radio CBA-IV. (volts measured at terminals by Fluke, not the CBA-IV) 0.25C discharge current (5A)

    After a 2 hour rest after charging (1 hour would have been sufficient), we started with the battery in this state unloaded prior to discharge:

    Overall voltage: 13.58v
    Cell1 3.403v
    Cell2 3.396v
    Cell3 3.399v
    Cell4 3.385v
    Balance: .004v

    Cells resting at about 3.35 or higher indicate a nearly fully charged cell for most prismatic lifepo4's.

    Target: to achieve at least 80% of the capacity I paid for with the 20Ah battery.

    At 17Ah withdrawn and still under load:
    Overall pack voltage: 12.7v
    Cell1 3.176
    Cell2 3.173
    Cell3 3.185
    Cell4 3.175
    Still well within the flat part of the discharge knee, balance here not relevant, but EXCEEDED 80% target.

    Just prior to pulling test discharge before hitting the 12.5v LVD:
    Ah withdrawn: 17.9
    Overall pack voltage: 12.6v
    Cell1 3.150
    Cell2 3.150
    Cell3 3.160
    Cell4 3.147

    So basically, even though the charge controller was set for 13.8v, which the battery never quite reached because it had already achieved a little more than C/20, I was able to withdraw about 90% of the capacity. I think I'll set my LVD to 12.6 since I never want to go beyond about 90% anyway.

    I'd say that's pretty good, and points out that you don't need to go bananas over trying to charge up to the very bitter end. There just isn't any need. 13.8v as your absorb point in a standard pwm controller seems to be fine as a stopgap until a dedicated lifepo4 controller is implemented.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Final conclusion, and an acknowledgement that the above could probably use a LOT of polishing and peer review for accuracy.

    It really is simple to implement, obtain, and maintain a LiFePo4 solar house bank with large prismatic cells. Complications arise from high-voltage, high-current applications, which a solar housebank is not a part of.

    If you have the motivation to take care of your flooded FLA's with a hydrometer once in awhile, then using a quality voltmeter to perform sanity checks should likewise be no problem. The alternative is to complicate things with external balancing equipment, which is not mandatory, but I understand the reasonings for those that do.

    Like any serious installation of any chemistry, an HVD and LVD would be nice additions, but again, not mandatory if you are willing to understand your system and maintain it.

    Since our needs differ from EV'ers, Ebike' and RC modelers, these batteries will live a pampered life. No need to push the limits in voltage, nor go beyond the knees, unless you need to for an initial balance.

    While we can take these down to 90% DOD without any harm (provided you are reasonably balanced), our need is still to provide a margin of comfort in autonomy, so I would NOT plan a solar system for daily cycles that go down this far. Realistically, instead of planning on a 50% DOD max like we do with lead, I might extend that to 60% with lifepo4. Not because it can't handle it, but because no matter the chemistry, we still want the power reserve to account for the unexpected.

    This has been an extremely fun trip! But I didn't really do this alone.

    I'd like to thank everyone who shared their experiences, suggestions, and advice, either directly or indirectly from the following:

    NAWS of course!
    SolarPanelTalk
    CruisersForum
    ADVrider
    CandlePowerForums (CPF)
    WheelchairDriver
    DIYElectricCar
    EVDL discussion list

    Any questions? I'll be happy to help you get started on your journey with prismatic Lifepo4.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    PNjunction wrote: »
    What was happening is that the battery reaches a C/20 current state long before 13.8v is reached. What I believe to be happening is that even though I set the absorb point to 13.8v in the Xantrex, it may be following a lead based curve in regards to the duty cycle of the pwm, but I'm not really sure of that. So like the Samlex, it perhaps isn't taking advantage of the lifepo4 characteristics, but it does serve as some sort of HVC in any case. Either that, or the fact that a fully charged cell for rested prismatics is about 3.4v or so, although it takes just a little bit more to charge it to that level.
    Does the Xantrex incorporate an End Amps setting for switching from Absorb to Float? Based on the bank size in AH that you entered?
    If so, you could get the full Absorb by entering a larger bank size. But it is probably better just to be happy with the behavior you have.
    SMA SB 3000, old BP panels.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    These little pwm's just let you reach your absorb voltage, hold it there for an hour, and then fall back to the float setting.

    Ideally, the Xantrex should wait to start doing pwm voltage limiting at exactly what I set it at, in this case 13.8v. But, I'm not sure if it is dipping it's toes into the water at 13.6v perhaps, and applying a sort of sliding scale to the duty cycle to the pwm waveform. Not sure.

    Add to the fact that I never really get there anyway. What I may be seeing is that when charging at just a bit over a 0.2C rate, (4.5A continuously measured from my 80w panel - at the start anyway with the 20ah battery), the battery never really needs to do an absorb of any consequence. That is one of the big draws of lifepo4, minimal Peukert. Since my charge current is so low to begin with, the battery is just accepting all it can, and when it finally reaches something like 98% soc, all I really need to do is just have it stop. I'm probably going to set the float in the Xantrex to something benign under 13.8v to emulate the off switch in case for some strange reason it flakes out and lets it absorb unnecessarily for an entire hour.

    3.4v for a prismatic needs to see at least 0.050v higher charge voltage to force any current to flow, so when charging at a very low rate, ideally I should just stop at 3.45v per cell (13.8 total) since there is practically no absorb. My GBS cells are either more efficient than that, or the Xantrex *may* be contributing by limiting current just a little bit - I haven't torn it apart with measurements to figure that out yet.

    I could go nuts by taking into account the rated internal resistance and doing it that way, say for a lifepo4 with 2.5m IR max (GBS spec for the 20ah):
    http://www.gbsystem.com/products-show-en.asp?id=58

    *** Lurkers! Spec sheets like these are typical across the various brands. There are the absolute immediate damage voltages, then there are "recommended" voltages intended to drive the SOC very high in order to make the external bleed/balancers work. If you don't take that into account, they also appear to be perfect drop-in-replacements! Just ask yourself why GEL isn't around much anymore. Even a commercial vendor like Genasun, with their own bms/balancers never take them this high. 14.2v is the max that Genasun takes these to. For those who don't need to drive an external balancing system, take it even lower!!. All I'm saying is that the manufacturer can't take into account all the possible uses for these batteries, and recommendations on spec sheets are not finely tuned for the application they end up in. ***

    0.0025 * 4.5A (my solar panel) = 0.012
    add the fully charged voltage: of 3.45
    3.45 + 0.012 = 3.462 per cell (13.84v pack total)

    If you charge at insane currents like 1C, then Peukert becomes more of an issue, and some may need to take it to a bit beyond 3.5v per cell, and allow for a much longer absorb. (But here, the penalty is like going from a 5 minute absorb to a 30 minute one - mind blowing.) But no solar house-bank with a 400ah battery is likely to achieve a 1C charge current - the largest you want to go with these, unless you are an EV'er needing 3C bursts for a few seconds.

    In the real world of my backyard, I'm seeing that I don't need to be dancing on the head of a pin like that. And what I'm experiencing like others, is that these batteries may be UNDER rated for capacity, even when you take them only to the start of the edge of the discharge knee.

    It really is hard for me to accept the fact that undercharging, as long as you have enough overall capacity to work with, means the loss of obsessing over achieving a 100% SOC each and every time. It has become a truly disruptive technology for me.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Quick note about why even pwm with lifepo4 is more efficient than lead when charging, and surprises you at first.

    With a simplistic charge scenario under pwm voltage regulation, consider that when you start out with a 12v lead battery driven down to 50% DOD like my Optima, the panel is meeting about 12.2v initially and has to work it's way up from there, very inefficiently due to the battery dragging down the panel voltage - hence the popularity of mppt to compensate.

    With lifepo4 at a nominal 3.2-3.4v per cell when just above the discharge knee for most of it's charge cycle, the voltage the panel meets right off the bat is 12.8v-13.6v, and that is more efficient since we aren't starting out at such a low voltage drop. Ok, yeah, it's not mppt, but an improvement since I'm starting out right off the bat in a more efficient part of the power point curve.

    This is why it is funny on the Xantrex to see that I'm immediately being told I'm 90% charged already! But that's for lead obviously, but funny to see how optimistic it is! :)

    So there is another tickbox for lifepo4 solar charging with simplistic pwm measures - you are already starting in a better power point than lead.

    I haven't yet tried to put numbers to an efficiency analysis between pwm and mppt with lifepo4, but I have the feeling that the margin narrows considerably compared to charging lead with pwm. Still, that's just a guess on my part.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    More tests with the Schneider / Xantrex C12 pwm ....

    It is obvious that one needs to set BOTH the absorb and float voltages to 13.8v due to the 1-hour timeout in absorb.

    I set the float to a benign value like 12.5v, and the controller just timed out in an hour, long before even reaching a decent charge or the upper voltage limit I set, but yes, current did fall to zero when it did switch to my low float voltage after an hour. So that's out.

    it proved to me that the timeout period begins LONG before the upper voltage setpoint is reached. I never noticed this before since I was constantly hooking, unhooking and reattaching things during my previous tests, resetting the timer.

    So, as expected, this pwm controller designed for lead will do the job, and it takes a trick to stop it from cutting off the charge too early. And I'm too chicken to attach the panel directly to the battery to see how much of a finger it may be putting on the pwm waveform. Perhaps this is justification for me to get an adjustable bench supply. :)

    At any rate the job will get done, perhaps not as fast as a dedicated lifepo4 CC, (or something like a totally controllable outback or midnite - but I'm not going to spend that much to maintain a 20/40ah battery) and as a stop-gap measure, is of high enough quality to serve as a glorified HVC.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    TEMPERATURE concerns ....

    Today this test was done in 105F in the shade (battery that is). That's making me a bit uncomfortable, even though I'm still a bit under the ambient specs, and the cell skin temperature is very cool considering I'm only charging at 0.25C. No bulging, no drama. Even the nickel plated cell-links were cool when checked with an IR thermometer.

    But getting close like this means I have NO TRUST in my radio-shack outdoor thermometer!

    That means another multimeter is coming to the bench - the Fluke 116 meter designed for HVAC electricians with a built in ambient temp sensor.

    Since I'm only using the meter for voltage checking and not amperage, I should have thought this through a little more when I picked up the 115 as my outdoor beater for battery monitoring. Oh well, I'll put it to use, but there goes my bench power supply budget. :)