GBS LiFePo4 20 and 40ah batts

13»

Comments

  • zoneblue
    zoneblue Solar Expert Posts: 1,220 ✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    PN, to me theres a bit of difference between cell level balancing and cell level monitoring, at least in terms of reliability. Surely a monitor which has a high input inpedence, is going to be more reliable than something with mosfets in active use? Still a nest of wires i guess.

    I bought a few panasonic 18650s and a 8 cell monitor alarm to fit to an old nicad cordless drill, to take your advice, to gets some air time. Ill let you know how it goes.
    1.8kWp CSUN, 10kWh AGM, Midnite Classic 150, Outback VFX3024E,
    http://zoneblue.org/cms/page.php?view=off-grid-solar


  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    zoneblue wrote: »
    .. I bought a few panasonic 18650s and a 8 cell monitor alarm to fit to an old nicad cordless drill, to take your advice, to gets some air time. Ill let you know how it goes.

    Ok, be SURE they are 3.2v nominal Lifepo4's, not some other li-ion chemistry! Note that these are high-rate cells, not the lower-rate prismatic cells, so characteristics will be a little different. And of course I don't know the QC of these Panasonics. I'd still recommend using actual prismatics made by GBS or CALB like the 20 or 40ah 4S packs, but maybe you can get an idea with these.

    Scale your charge / discharge rates to those of a house bank. If those Panasonics are say 3.5A, then think of charging and discharging at 0.2C max perhaps to simulate what you would still do with say a 350ah prismatic. :) At this point I guess we are talking in terms of only a few hundred milliamps at most.

    A lot of guys use the Junsi cell-logger, but you'll want to power that seperately and not drive it off the Panasonics themselves. Keep an eye out for idle / sleep currents waking up and draining cells overnight.

    I guess the point here is to see if you can get AWAY from all the RC hobbiest stuff hanging off the battery, and live comfortably at no more than 14v total with a reasonable balance under 3.6v.

    WARNING! You don't often see this mentioned, especially in the RC world, is that once you start to get into the steep discharge knee of about 80% DOD and beyond, you MUST come back out of it slowly with about 0.001 to maybe 0.005C current until you get back to 12.8v (3.2v per cell), and THEN you can apply full current. This is a secondary reason NOT to go past 80% DOD. Real world, 3.2v or so is where I would set my voltage alarm to under load, and a total LVD at 3.15 under load. Ideally simulate a solar installation and not go past 50% DOD. I guess you'll want to calculate / measure the current draw instead of relying on voltages, which in the middle of the flat discharge curve can be misleading.

    Some EV'ers and RC modelers drain well past to 90% or more, and although they do the right thing by getting a charge back on it quickly, many do so with too much initial current. PUFF. That's just as bad as too much voltage at the top.

    To do this right, I suppose a small variable bench supply would be the ticket, but if it goes awry, try again with the real prismatics. :)
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Optimate TM-291 lithium charger and Samlex charger are a great combination!

    After purposely unbalancing my battery to play around with manual charge / discharge balancing, I could get no closer than 0.020mv delta between the cells on my 20ah unit and lost of lot of hair in the process - or at least what's left. :) I got tired of playing around like that, yet I'm still not going to put external balancers on the battery. No way.

    To make a long story short, the Optimate charger rebalanced the battery, even though at first glance, the voltages are not exactly the same after rest. HOWEVER, when I follow up with a discharge and a recharge with my Samlex charger set to only 14.0v, they are about perfect, no more than 0.005mv difference. This only implies a balance of voltages, and not of true capacity, but since I'm not an EV'er taking this down to the danger zone, I'm not too worried, and have proven to myself that I can get at least 80% DOD out of the cells easily, without any one cell taking a nose-dive, even when not perfectly balanced. Being conservative is the key.

    And yes, I feel it coming from those saying that there is no way you can balance a lifepo4 battery in series without using individual charging / discharging of cells. Yes, I know. What I have proven to myself through real-world experience, is that the technique the Optimate uses actually does work. I was prompted by the sales of these things to powersports users of high quality lifepo4 batteries that don't use internal balancers, like AntiGravity. While that is not my application, nor the cell types I use (same basic chemistry though), the balancing act appeared before my very own eyes. Trust, then verify. I did, and it passed.

    Now that the battery is back in balance, I'm going to continue to use the higher amperage Samlex charger, or my own solar setup with the simple Schneider pwm controller set no higher than 14v. Every once in awhile, I'll run the small Tecmate-Optimate lithium charger on it, and let it do it's own thing, but like before, these cells tend to stay in the state of balance you leave them in under a "Sub-C" application like ours, unless you do something to change that.

    I have only tried this on the 20 and 40ah GBS batteries, but the Optimate says it will handle up to 100ah. I'll have to save up some more bucks to prove that to myself. :)

    My suggestion is that before using the Optimate, the cells need to be in *reasonably* close state to each other, as they are supposed to be from quality dealers. Check them at say 13.9v or so, just before the Optimate goes into the cell-balancing routine, to make sure you don't have anything way out of the ordinary, like 3.7v on one cell, and 3.2v on another. A single cell charger can get these reasonably close, and then you can try again with the Optimate.

    (Just in case you thought I have dropped lead altogether, I haven't. I also have a lead-based Optimate 6, and am super pleased with that for equalizing my agm's, when doing an EQ the normal way with lead is prohibited by most agm manufacturers. I don't have access to the cells of course for immediate proof, but I have seen slight improvements in my discharge testing with the West-Mountain-Radio analyzer.)

    A conservative upper limit voltage of 14.0 max, and with the Optimate-Lithium to initially balance and later use as a PM measure is the way I'm going to keep it all simple. And of course an LVD of about 12.5v max under load. It really is this simple.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Another test, another charger!

    Things are really settling down. Tried charging my 40ah GBS batteries with an automotive-style Vector 25A VEC1095A model. Set for GEL (which is the default), it doesn't take the battery beyond 14.2v. Used 25A for charge (about 0.65C, something I'll never do in real life with a large housebank), and it didn't bat an eye - totally cool and never rose above ambient.

    The charger remained at 25A charge until about 14.1v was reached, and a 15-minute absorb took it down to 3.5A, at which point I just stopped it.

    The nice thing is that the Optimate Lithium charger did it's job balancing the cells on a previous charge after I was done monkeying around with purposely unbalancing it.

    Results at end of charge with the Vector:
    Voltage 14.21 (limited by the gel setting)
    Current-end during absorb 3.5A (manually pulled it - no need for more)
    Balance prior to withdrawl from charger:

    3.559
    3.539
    3.565
    3.549

    0.026 delta. Maybe I'll play with more runs of the Optimate if I want to become obsessive. At any rate, no cell is exceeding 3.6v, and is safely under that.

    NOTE: The Vector charger was solely for a test. Because it has other functions, like desulphation, engine-start, EQ, etc, I'm not comfortable trusting it in case the microprocessor flakes out, and applies one of these other modes. The Samlex will remain my go-to ac charger.
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Compliments to your detailed approach to tests that can significantly improve on what seems to me the next generation of solutions to energy storage for us on/off gridders!

    I'd appreciate your opinion on some of my long term goals;
    1. Use my on-grid 4000 watt SMA Sunny Boy on-grid system to the maximum for off-grid use (now simply using the Secure Power Supply to charge some Club Car Golf cart batteries (48 V)).
    2. Use my old Prius as the backup "generator" in the system (following the great work by techntrek on this forum)
    3. Rather than a traditional 48V lead acid system, use a ~200 volt DC LiFePO4 (LiFeMnPO4?) that is compatible with the Prius (and possibly it's charging) system
    4. Use an APC Smart UPS (the DC to AC inverter section) - these have capabilities to produce 240 V (single phase) AC with DC feeds between 170 and 240 volts. Adding a center tapped transformer gets the 240 split phase needed for my house.
    5. Last but certainly not least is to use the 4000 watts of solar panels with a VERY high voltage charge controller (or some type of DC/DC converter?) to recharge the Li batts to limit the use of the Prius. Don't believe this piece of equipment actually exists, but the DC voltages from the panels could be set up to be a fairly good match. My current config has ranged from 180 volts up to 220 volts, but of course current/voltage combination is "all over the map" with ambient temps and sunlight conditions.

    Some multi use possibilities of the PV/UPS/Prius/Li batt system
    1. Use the Li batt pack IN THE PRIUS to improve gas mileage and simply hook it into the solar system when in the off-grid mode - there are sellers of these type add-on battery packs that exist now. Your designs seem far better - BUT can they stand the rigor of EV use?
    2. Add a crappy 240 V split phase generator (propane) for long/long term backup and input this to the AC to DC inverter of the UPS , which can then output very clean AC power - and could possibly be set to auto switch if needed to recharge batteries and supply backup generation. Lose some efficiency here, but may be worth the simplistic approach?

    Why on earth would you do this?
    These UPS's seem to be readily available as the big networkers apparently "throw them out" when they either upgrade or have to change the batteries!
    The Prius seems to be close to if not the most efficient "generator" - including the very expensive Honda inverters
    Wiring becomes very manageable with 8 awg probably the max in the 4000 -5000 watt range.
    The PV panels voltages seem to match well
    The li batteries seem to be a great fit from almost every perspective, except maybe cost and level of experience in the off-grid sector.

    Likely, new safety practices will need to be thoroughly vetted with such high DC voltages - but nothing outside current knowledge!

    Any comments are appreciated
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    MarkC wrote: »
    I'd appreciate your opinion on some of my long term goals;

    All of those applications would be fine for those cells technically, but don't jump the gun. For example, what exactly are the GC batteries not doing for you now that you think lifepo4 will resolve? I'd hate to see anyone just do a chemistry-swap without putting a lot of thought into it.

    Your Prius and high-voltage solar application would be best served by consulting an engineering firm, unless you really are on the ball DIY. And that is what I'd recommend here - if you want to get your feet wet before blowing a huge hole in your wallet, grab a simple 20 or 40ah GBS (or CALB or Winston, or whatever you like) to get some real "hands on" before you spend the megabucks on a system. With some hands-on behind you, you will be less likely to be taken by crooks just wanting to dump some poor drop-in type of engineering on you, now that you have some knowledge to back up your request.

    A simpler application would be to replace your GC's with lifepo4 - but if that bank is not really doing much cyclic service, then you are wasting the capabilities of lifepo4. In other words, would improving your current setup with lead be a better option?

    Lifepo4 is great, but it is not the "power pill" for the masses just yet, despite what saleseman tell you. :)
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    After "experimenting" with my APC SURT 6000 and two sets of older AGM battery cartridges (these operate at a nameplate 192 volts), my appreciation of battery characteristics of voltages during charging, discharging at various loads, resting, has certainly increased (again thanks to techntrek of this forum).

    My ultimate would be to have a single set of batteries that could be used by the GC normally (with maybe a spare) and possibly be rewired to use in an off-grid mode for night time storage AND work with the Prius as the (likely substitute) traction battery as my backup generator. Realizing that this will take some time as the charge controller for this does not appear to exist???, I agree starting simple is best. So - what seems to be the best alternate now is to buy the Li batteries that can be used in the GC (wired 48V of course) and "learn" about them. I can later add/reconfigure to the nominal 192 volt configuration that will hopefully, eventually be useable with the PV panels, APC SURT and the Prius. Appears we are not going to get the big hurricane (knock on wood) this year, so I have time to get the "hands on" before worrying too much about off-grid considerations. The GC will give plenty of real world experience, especially with my Grandkids using it continuously!

    BTW, I am going to spend significant money for a backup set of GC batteries and keep them charged as mine are getting quite old and is located in a rural area far from any deep cycle battery suppliers. A huge advantage SEEMS to be that the weight savings will allow me to carry an additional adult up some steep hills in the GC. Deciding on the useable energy needed for the GC is another challenge.
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Mark - Because your needs straddle BOTH EV and much simpler solar energy storage, you really need to consult a system expert who may be able to bridge both of them for you:

    http://www.elitepowersolutions.com/

    Your "learner" 48v bank replacing the GC's may be inadequate for use in the Prius. Those are more complicated as seen here:
    http://gwl-power.tumblr.com/post/361692495/toyota-prius-plug-in-conversion-using-thundersky

    Bridging the gap between the two is best left to an engineering firm from which you may also need support despite our best efforts here.

    What is presented in this review, is more of a simpler DIY effort that one can use to educate themselves, and not necessarily be the ultimate lifepo4 solution.
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Thanks for all the advice. After much searching, emailing, phone calls, it seems any "off the self" system components are a long way off to integrate solar/PV and "traction" type battery systems (and of course back up generation from EV's) - and maybe never felt to be worth the effort. One EV battery supplier is making a convertible battery pack for the Prius that can be easily re-wired for nominal 48 volt service in the traditional off-grid solar environment. Their take is that going much higher in DC voltages for renewable energy storage will be met with great resistance due to safety considerations. Doesn't actually make that much sense to me as many PV panels are now operating at 400 plus volts!!

    Not giving up easily on this as it seems too natural of a "fit", but realize that it will take much time and experimentation.

    I'll certainly post any results.
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Another approach could be to replace the SLA batteries (16/12V 5ah) in the UPS with an assembly of Li batteries - sort of a custom assembly using the "best fit" in the existing battery case of the UPS. The UPS has a semi-adjustable charger that can be set to various final charge voltages of 215-225 volts. I'm wondering if designing the size/number of Li batteries to match the cases and charger is a possibility? These units have management (network) cards that can record battery voltages, wattage usage, % remaining load, and the charge voltages and % charge when back on charging circuit - nice test setup - but not necessarily correct for the Li batts. Again, it would likely take a systems designer to verify all the affects of the Li batteries on the charging ciruits. The DC/AC inverter has already been utilized with the Prius traction battery (techntrek of this forum) and appears to be just fine! More musing!
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    I personally wouldn't use lifepo4 in a UPS application. This is mainly due to wasting their various capabilities of just sitting there. You'll get the most "value" out of them if you actually use them in a cyclic application, where their capabilities are put to full use. For UPS use, I'd simply stick to a high-rate discharge AGM. Perhaps a commercial customer would use lifepo4 for a UPS, but I wouldn't.

    Many questions before purchasing a much bigger system can be answered by picking up one of these smaller 20 or 40ah GBS cells (although they have much bigger versions) and experimenting with it.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Remarks on a Junsi iCharger 306B and the smaller GBS batteries:

    Once I proved that I could properly charge and maintain these cells manually with conservative charging techniques (like not letting any cell go higher than 3.6v), and running at 3.5V maximum instead (14v max on my nominal 12v battery), I thought I'd see what the hobby-charger environment would be like. The findings are interesting as I really didn't want to go that route initially.

    Naturally, it made very short work of balancing each cell. I just cut the 6S balancing cable in half, and attached my own short jumpers to the GBS battery cells and let it do it's work (two wires went unused). The most amazing thing is that when all was said and done, I got VERY close to the same balance manually as it did - but it got much closer and MUCH quicker. :) At first I was concerned when the voltage in the display didn't match up with my Fluke multimeter until near the end of charge. I found out that the iCharger's compensate for the voltage drop in the cabling.

    BUT, my intentions here are NOT to balance on each and every cycle. Fortunately this charger will also do just regular two-terminal charging. So guess what? After an initial balance, I've been doing simple two-terminal charging and no cell drift is taking place. Which it shouldn't since as explained before, this doesn't happen in a "sub-C" application. Because my solar charge controllers don't come with balance leads, it was very important to me to make sure that simple two-terminal charging and maintaining balance would work without them.

    Because I can do two-terminal charging (without balance leads) in addition to balance charging, I'm going to retire my Samlex lead-based charger since this iCharger has some features above and beyond the Samlex.

    I won't get into this like an RC-modeling review, however I do feel that it is a GREAT charger for bench-testing a setup prior to making a commitment to a much larger system. You can prove to yourself what is real, and what is FUD from those who don't have any hands-on. You have to provide your own power supply, or large battery to run the charger. Yes, there are plenty of menus to wade through.

    A 6S charger like mine is fine for a nominal 4S / 12v system, but if you are playing around with 24v, then an 8S or 10S would do. While I like the Junsi iCharger line, there are others like Revolectrix which are highly regarded. At some point however, say with an 800ah bank, balance charging might take a bit too long, and manual techniques like using an automotive bulb to bring down cells that go high are in order. Like anything, I wouldn't pick a bottom-of-the-line product, especially with our larger cells.

    While I hesitated at even looking at a "hobby-charger", it has proven to be a very good tool for this experimenter.

    The best thing about it was that it proved to me that with some common-sense and conservatism, it is NOT MANDATORY to use a complex charger in our application - especially when that is primarily a two-terminal charge from a solar-controller!
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    PNjunction wrote: »
    I personally wouldn't use lifepo4 in a UPS application. This is mainly due to wasting their various capabilities of just sitting there.
    It strikes me that the biggest mismatch is that you want the batteries in a UPS to float at 100% charge to maximize the available capacity during an outage, but with these batteries that will result in a notably shorter battery life, even if they are not cycled.
    SMA SB 3000, old BP panels.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    That's a good point, and another reason I don't see the practicality of using them for UPS duty for the average consumer.

    However, for a commercial customer, things may be different. While it is a best practice not to store li-ion fully charged, most of the warnings come from the RC modeling application, where they are using a different higher energy-dense chemistry, such as LiNmc, or LiCo02, and not lifepo4.

    Yes lifepo4 also is shipped and stored at 50% or less SOC, but in a real-world ups application utilizing them where cycle-life is not of the upmost priority, then it may be possible to accept the compromise of storing them at a high soc for an infrequent backup. I have no hard-data for this, only anecdotal, and if I find it, I'll let you know. I know that if I were forced to store them at a high SOC, I wouldn't take them to 100%, but something conservative yet again, like maybe 90%. Or if your wallet is golden, just double the capacity you need and store at 50% DOD. Ouch!

    Still, without any hard data on how much *lifepo4* REALLY degrades under these conditions, this is just guessing. I'd like to dig around and see if yet again, all the major concerns come from OTHER chemistries, and perhaps lifepo4 is liveable under certain conditions. When I find it, I'll get back.

    In the meantime, for an average consumer UPS application, I'd still stick to using high-rate quality agm's like those from Enersys.
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    PNjunction wrote: »
    That's a good point, and another reason I don't see the practicality of using them for UPS duty for the average consumer.

    I have crossed two post threads without the proper explanation! The UPS service that I have referred to is the APC SURT models that have been used as the DC to AC inverter with a Prius to provide a backup generator for off-grid use (techntrek thread on this forum). The AC to DC power unit and charger are not used. However, these UPSs have much more potential possibly? The battery packs are modularized, and up 4 complete 200 volt systems (8 - 100 volt packs) can be controlled and monitored by one UPS - simply hooked up by 200 volt Anderson connectors. Could this system provide the energy storage from limited PV panels for an off-grid design? The "missing link" seems to be the 200 volt (nominal - output) charge controller. So the Li batts would be in a more traditional off-grid energy storage mode - BUT would be operating at EV voltages (190 - 240 volts). These UPS have network cards so that operating characteristics are monitored and available remotely. Again, thanks for feedback and advice.
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Since the thread here is more about the generic operations of GBS lifepo4 cells, I think that you may be better served by getting more specific answers from a more focused group on your Prius:

    http://www.diyelectriccar.com/forums/
    and
    http://www.evdl.org/
    and
    http://priuschat.com/forum/

    Note that the first two forums / lists pertain mostly to those who build/convert the whole thing, and not necessarily concentrate on commercially produced EV vehicles.

    For DIY'ers, even though we aren't using these cells in the same application as EV'ers do, there is a wealth of good information in the first two references in the batteries / charging subforum that is generic to us. Just keep in mind our relatively low-voltage / low-current housebank application which will differ from some of the results from an EV application.
  • MarkC
    MarkC Solar Expert Posts: 212 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    For DIY'ers, even though we aren't using these cells in the same application as EV'ers do, there is a wealth of good information in the first two references in the batteries / charging subforum that is generic to us. Just keep in mind our relatively low-voltage / low-current housebank application which will differ from some of the results from an EV application.

    PNjunction;
    I'll check out the subforums - thanks.

    Note that this thread is very helpful to me in that obtaining a better understanding of use of LiFePO4 cells for off-grid use - the proper set up, handling, sizing, and specifically balancing/charging. The "Prius as a generator" thread by techntrek is the most helpful to me of any of the "EV" forums around for that purpose. It may be that a new "thread" about the use of "UPSs", specifically the APC SURTS 5000 and 6000 Units, combined with the use of LI battery technology might have merit for off-gridders. The much higher voltages do seem to present a new set of challenges. In your opinion, do you believe that 65 Li cells in series will be the most challenging issue (or show stopper?) for long term off-grid use? These SURT battery packs (16/12V AGM batteries - 96 cells) typically last 3-5 years, but as you (and others) noted- are kept at a relatively high state of charge continuously (217 volts typically) - which does seem challenging for a "sealed" type battery with no realistic method of adding water.

    I'll keep up with your progress - very helpful and thanks
    3850 watts - 14 - 275SW SolarWorld Panels, 4000 TL-US SMA Sunny Boy Grid tied inverter.  2760 Watts - 8 - 345XL Solar World Panels, 3000 TL-US SMA Sunny Boy GT inverter.   3000 watts SMA/SPS power.  PV "switchable" to MidNite Classic 250ks based charging of Golf cart + spare battery array of 8 - 155 AH 12V Trojans with an  APC SMT3000 - 48 volt DC=>120 Volt AC inverter for emergency off-grid.   Also, "PriUPS" backup generator with APC SURT6000/SURT003  => 192 volt DC/240 volt split phase AC inverter.  
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts
    MarkC wrote: »
    In your opinion, do you believe that 65 Li cells in series will be the most challenging issue (or show stopper?) for long term off-grid use?

    Oh man, certainly. There is NO NEED for us to use very high-voltage, and thus our batteries are vastly simpler in construction and management.

    Since we aren't designing for EV, all we have to do is string either 4, 8, or 16 cells together of the proper capacity for the typical 12/24/48v bank. Management and proper wiring infrastructure is vastly simplified, especially when operated conservatively.

    Part of the reason for my review of these GBS cells is to help prevent some from making a knee-jerk purchase like a high-voltage ups, which may be entirely inappropriate for our non-EV application, and get some hands-on first with a small setup.

    In fact, I'd even go so far as to say that just building a simple 4S / 12v battery with the 40ah cells would be the most appropriate, as the 40ah cells and higher capacity use basically the same cell interconnect setups, and the latest in their chemical advances. The 20ah cells work, but they differ just slightly in operation from the 40ah and higher, which I suspect is still the "gen 1" stuff. Not bad, just a little different.
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Modification to Morningstar Prostar pwm controller for lifepo4:

    One of the problems that stopped me from using the Prostar with my lifepo4 batteries was that there seemed to be no way to disable temperature compensation, which we don't typically use with lifepo4.

    Fortunately, if the little 2N3904 transistor fails, the unit still operates, but without compensation. The correct way to do this would be to open the case, and unsolder / clip leads off the transistor. Since mine is dedicated to these lifepo4 batteries, I just wiggled it like a loose tooth, and it came away cleanly from the board.

    NOTE ** This applies only to the Prostar as they say in the manual. Other Morningstar products with this temp-comp transistor removed may fail to operate at all!

    Now I just set it to "GEL", which has an absorb of 14.0v. Perfect! Or, if I really want to, set it to "sealed", which is actually 14.15v absorb. Still good.
    In fact, the "sealed" position has a mild boost/eq of 14.35v for an hour every 28 days which is just under the normal lifepo4 limits. Still, this doesn't give you much headroom if your batteries are not really well balanced. I'll stick to GEL (14.0v), since that gives me some nice headroom, and there is no monthly boost/eq. My batteries are not permanently connected, so when using the "sealed" setting (14.15v), I may be resetting the 28-day boost timer anyway. Still, the simplest solution is to just use GEL (14.0v).

    Do NOT set it to "flooded", which has a very absorb and high EQ voltage every 28 days. Nicely engineered, if the battery-selector switch fails in the field, the unit will default to GEL. Nice - that's what I use anyway.

    Of course the SOC led's mean nothing as they are calibrated for lead - but the unit still functions and does the job.

    But, I would not use this on a permanent-fixed basis where there is not much cycling. This is because the controller is designed for lead-acid, and as such the controller will happily try to go below C/20 in absorb, and there is no need to do that with lifepo4. In essence, you may be faux-floating at 14.0v while the controller tries to go below C/20. The REAL float is set at 13.7v, which is good and benign, but I'll have to see if the Prostar has a timeout when absorb current time is lengthened by such minimal current.

    I use this thing for portable / emergency charging and not as a fixed controller for daily use unattended. My Prostar has metering and I just pull it when I see the current is near C/20, and the voltage is at/near 14.0v, but I don't obsess over not having a super-programmable controller. At least not yet. :)
  • PNjunction
    PNjunction Solar Expert Posts: 762 ✭✭✭
    Re: GBS LiFePo4 20 and 40ah batts

    Update on the modified Morningstar Prostar 15M:

    Upon reaching my "gel" preset of 14.0v, it maintained that voltage at the terminals (made my Fluke 87v happy!), and absorbed all the way down to Zero amps. This is not necessary with lifepo4, as C/20 will do.

    After an hour of 14.0v and 0 (zero) amps, it fell back to 13.7v. That is, the controller displayed 13.8, but lifepo4's low self-discharge meant that I'd have to wait for 12 hours or so to actually see the battery voltage drop to that level. Catch-22.

    So while it isn't the most ideal thing, a 1-hour extended absorb down to 0 amps at 14v, and falling to 13.7v isn't really a deal-breaker for my non-daily usage. It isn't enough to freak out about if I forget it. I do NOT leave any batteries charging unattended, and pull it when convenient - preferably before going into the unnecessary extended absorb.

    Float - the ONLY time you should ever float a lifepo4 is at well under the fully-charged voltage, and is only really done to catch any small parasitic loads that might occur should it be enough to drain the bank that far in the first place. Ideally, if you HAD to float, it would be under 3.45v per cell (13.8v for a nominal 12v battery). The Morningstar is rated at 13.7v float so this is just a smidgen underneath that. If I can find a convenient way to disable float altogether in the Prostar, I'll let you know.

    So in this non-critical application, the Morningstar Prostar, even though designed for lead-acid, can be set to do lifepo4 with a bit of a compromise in the absorb mode. Just remove the temp-comp transistor, use the GEL setting, and you are good to go. Again, for non-critical setups. I would use a much more programmable controller on a larger or more critical bank.

    What really makes me happy is seeing the Prostar matching up with my Fluke 87V for the most part when measuring voltages. The display voltage will actually be a bit off during charge, since it takes into account the voltage drop of the cabling. Only at the very end will it become extremely accurate. OR, if you like you can use the "sense" wires - I didn't need to as the Prostar stopped very accurately with just a 2-foot jumper from it to the battery.