Motorcycle Batteries .. AGM, GEL, Wet, Lithium Iron Phosphate (LiFePO4)

Discussion in 'The Garage' started by _cy_, Jan 20, 2012.

  1. Anorak

    Anorak Woolf Barnato

    Joined:
    Jul 23, 2004
    Oddometer:
    40,905
    Location:
    Belvedere, Marin County
    It looks that way to me too. Not the best battery because there are so many variables.It's a very good battery as a starting battery. Much better in my opinion than the generic Chinese lithium iron phosphate batteries.
  2. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    Okay, I am stumped. Consulting text books, online battery sites, and fired off an email of data to a crazy professor sort that teaches battery technology and may eventually respond.

    This is what I got so far on a Shorai LFX18 used roughly for 4 months.

    Noteworthy that the BMW F800GS has a wimpy pimply 13.8 charging voltage with substantial AC ripple and lots of HF switching noise.

    Also worth noting that the F8 has a hot battery compartment because the basic controller (giant SCR switch bank with some logic) bakes the battery with it's heat sync pointed at one face of the battery, about .5" of air gap, asymmetrically with heat concentrated towards cells 3 and 4 IF MEMORY SERVES, it may be cells 1 and 2 that are getting the hottest, or none of the above depending on how the cells are stacked in the battery, but I think the cells towards the positive terminal are getting hottest.

    In any case, here are my raw notes so far, very raw. Nothing seems to match up. AH and SOC voltage aren't linearly connected. Its late so maybe I just need sleep but right now I'm about as clued in as an ameba.

    I have a high degree of confidence in the measured numerical values, but nothing makes sense of what they are saying.
    Any thoughts on what on earth can be determined from this shit salad?


    Shorai LFX18, 4 months of heavy service on F800GS

    As arrived, at least 4 days of resting13.149
    cell1 3.2965 70%
    cell2 3.2962 70%
    cell3 3.2938 70%
    cell4 3.2633 35%
    35% imbalance

    Capo w F8 regulator 4.5 hours 3 starts 125a withdrawn for 10 seconds, rest 1 hour
    cell1 3.2998 70%
    cell2 3.2995 70%
    cell3 3.2970 70%
    cell4 3.2657 35%

    Withdrew current from cells individually at approx 0.2 amps to 3.2165 volts (would be 12.886 volts for battery or 20% SOC) This is a 30 hour rate. Shorai says some BS about “PBEQ amp hour” as do all the LiFePo4 power sports battery manufactures I can find. Dig and you find out they rate the battery at 6 actual amp hour, but state that using the bottom 20% will damage the cells and void your warranty. This means you effectively have a battery that is 4.8 amp hour
    cell1 2.56 amps 53%
    cell2 2.99 amps 62%
    cell3 2.87 amps 60%
    cell4 0.81 amps 17%
    charged cells individually same amount as withdrawn + 10% for a guestimate at charge eff

    After 26 hour rest
    cell1 3.2991 70%
    cell2 3.2997 70%
    cell3 3.2968 70%
    cell4 3.2646 35%
    close enough for government work

    Capo 6.5 hours 7 starts, over 2 days
    1 hour after 125 amps for 10 seconds
    13.708 volts
    cell1 3.5820 100%
    cell2 3.4448 90%
    cell3 3.3447 90%
    cell4 3.3142 75%
    25% imbalance

    3 more hours of Capo riding, all freeway
    cell1 3.5819 100% (rounded up ever so slightly)
    cell2 3.4494 90%
    cell3 3.3492 90%
    cell4 3.3181 80%
    20% imbalance

    Withdrew current from cells individually at approx 0.2 amps to 3.2327 volts (would be 12.93 volts for battery or 20% SOC to bottom balance all cells

    cell1 3.99 amps 83%
    cell2 4.33 amps 90%
    cell3 4.74 amps 99%
    cell4 3.04 amps 63%
    Charging it at 14.5 volts till acceptance down to 200mA
  3. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    @joel... excellent raw data! collect enough and some pattern will show itself. sure is nice to see data generated that's lab grade.

    there's sooo many different sources for Lithium battery information it's hard to ferret out which one is worth spending your brain cycles on. here's one worth a cruise..

    Charge LiFePO4 to 3.65v per cell then terminate charge if using a regulated power supply. if your power supply has constant current triggered by volt termination feature. set to 14.4v then allow time to top off. Your battery will absorb decreasing amount of current, closer to fully charged condition. ramping up voltage up to 3.65v allows faster charge for last phase. higher the volts delivered by charger, higher the amp rate pushed into battery. this is especially true during final phase before LiFePO4 reaches 14.4v fully charged state.

    intelligent chargers that support LiFePO4 will have custom algorithms to support that chemistry. best chargers will have balance options that charge different cells at different rates. others will charge all cells at same rate, then shunt off excess current.

    when you pull a fully charged LiFePO4 battery off charger. voltage should read just under 14.4v depending on individual battery. then surface charge will quickly bleed off to 13.9v range. after sitting overnight with nothing hooked to battery. resting voltage should be about 13.7v range. each battery will be slightly different. then as battery ages resting charge retained after a full charge will go down. when battery reaches 80% DOD it's considered spent.


    Charging
    [​IMG] Because lithium-ion batteries can have a variety of vocathode and anode materials, the energy density and voltage vary accordingly. Lithium-ion batteries with a lithium iron phosphate cathode and graphite anode have a nominal open circuit voltage of 3.2 volts and a typical charging voltage of 3.65 volts. The charging procedure is performed at constant voltage with current-limiting circuitry (i.e., charging with constant current which should be 30% of the battery ampere hour rating until a voltage of 3.65 volts is reached in the cell and continuing with a constant voltage applied until the current drops close to zero). Typically, the charge is terminated at 7% of the initial charge current. In the past, lithium-ion batteries could not be fast-charged and typically needed at least two hours to fully charge. Current-generation cells can be fully charged in 45 minutes or less; some lithium-ion varieties can reach 90% in as little as 10 minutes.

    http://lithiumstorage.com/index.php?main_page=page&id=3&zenid=774df594ed1201f57f11233f8fcc52f3
  4. Rajin Cajun

    Rajin Cajun Been here awhile

    Joined:
    Jan 30, 2007
    Oddometer:
    724
    Location:
    The DEEP South
    Finally got home and was able to check out the new pack for the SE. It arrived a little over a week ago and as it sits now it's at 13.44v. Hopefully I will be able to get the bike put pack together this week and see how she cranks over with this one vs the stock battery.

    [​IMG]
  5. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    What I know....

    When the battery arrived from being used hard by Lost Rider for 4 or so months, it was horribly imbalanced. Lost Rider does not have any charger so like many, only the bike charges the battery.

    His mount is an F800GS I know well from being a BMW dealership shop foreman and being front line on many many BMW issues.

    The F800GS has a 13.8 volt charging voltage. I don't know why, it almost as if they were planning on putting a GEL VRLA battery in the F8's and changed their minds and dumped AGM's in them at the last moment without changing the regulation voltage. It is not as if BMW is stupid, the R1200GS/GSA, S1000RR, K1300S/GT, K1600GT/GTL all have charge voltages around 14.4

    I would have to say it was well on it's way to complete failure though likely without symptoms since LiFePo4 batteries have a low enough internal resistance that they will put out, till one day all at once they won't.

    Best I can dig up is that a Shorai LFX18 is a 6 amp hour battery, and that includes the last 20% of charge which they don't want you to use, so effectively a 4.8 amp hour battery.

    This being the case, I am seeing amp hour capacity per cell that is all over the frisking board!

    Cell 1 appears damaged, and this seems logical. It was the cell with the highest state of charge and in series with 3 other depleted cells, likely saw much higher voltage then designed for.

    Cell 2 is sorta kinda reasonable.

    Cell 3 is actually showing more capacity then spec'd

    Cell 4 is also showing a reduced capacity compared to it's state of charge. Perhaps damaged by Anode damage from too low of SOC or even cell reversal?

    I don't know. I don't have any special charge equipment and won't be buying any as I am not at all convinced that lithium chemistries are the way to go for adventure riding which is all I do. This may change, but not with this sickly battery lol.

    What I do have is a charger that will charge across the main terminals at 800mA pairing back to 200mA at 14.4 volts, then ceasing all together when current declines further.

    It will not exceed 14.4 volts under any circumstance confirmed by plotting it's algorithm on a paperless recorder.

    I built a bottom equalizer with some spare parts and a bread board with a pot I can use to adjust disconnect voltage which I had set for 3.2165 volts and hit its mark on every cell within 1% It only applies a 200mA load, so waiting now for it to finish the second draw down so I can charge the whole thing equalized.

    The F8 also has massive AC ripple and a lot of HF switching noise and has a battery compartment that is hot but asymmetrically hot.


    Shorai keeps ramping up the size of battery they recommend for the BMW F800. The started with an LFX14 recommended and after a high failure rate with both that and the LFX18, currently recommend the LFX21.

    This is odd, because compared to the R1200GS and many other bikes, the starting demands cold or hot of the F800 is a walk in the park.

    As shown in tests, the Shorai will balance in series slowly when seeing the charging voltage my Capo puts out which varies greatly by load and engine speed but gets into the 14.55 volt range for brief periods and averages about 14.34 on the freeway.



    Working hypothesis: Shorai does not balance either cell resistance or capacity well or has self discharge that varies widely. 13.8 volt charging of the F800 is not high enough to permit the very limited balancing LiFePo4 packs will do when voltage is continued after fully charged. The cell with the lowest capacity, highest resistance, or lowest leakage current gets damaged by high voltage as the cell the most opposite gets damaged by under voltage until the battery rapidly fails.

    Just a theory, but all evidence points to high incidence of failure of Shorai in F800 application and Shorais every increasing recommendations of battery size for the F800 bares this out.

    If true, this argues strongly that 13.8 volt charging systems are not adequate for the Shorai battery.

    Big hedges: As stated, the F800 also has unusually large AC ripple, loads of HF switching noise from PWM pump control, and a hot asymmetrically heated battery compartment.

    I have spent half my life reading up on new technical things though only 2 or 3% on batteries. I will continue reading up on LiFePo4 but any first impressions or suggestions for experiments to confirm? :)
  6. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    question... at what rpm is the 13.8v charging voltage measured?

    if F8 is putting out to 14.55v even for brief periods (what rpm?).. then voltage regulator is showing it's limits. F8 is so darn computer controlled, it's flat amazing... how is the charging volts controlled? via conventional voltage regulator? or via CPU?

    it's OK to use any conventional lead acid battery charger with Gel/AGM settings. if you keep track of volts LiFePO4 battery is getting charged to. setting any AGM/Gel charger at 2amp setting is the safest. for termination voltage is limited to about 14.4v.

    "The peak charging voltage for Gel batteries is 14.1 or 14.4 volts, which is lower than a wet or AGM type battery needs for a full charge. Exceeding this voltage in a Gel battery can cause bubbles in the electrolyte gel, and permanent damage."

    LiFePO4 batteries like getting charged at .5C rates. Shorai spec's 3C max charge rate, but can easily take 5C rates. note battery will only swallow large amps until it reaches about 80% full, then battery will slow down accepting current. unless volts is stepped up to push higher current towards battery.

    normally I'll charge at 1-2C rates ... for Shorai LFX21 AH (pb eq) that's 7AH actual rating... so 2C = 12amp charge rate. when charged at 2C to 3C rates it's amazing how fast these LiFePO4 batteries charge up to 80% of full charge. then rates slow down for final phase.
  7. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    Two different bikes...

    The F800GS, the bike this battery was used in for so long, puts out 13.8 volts from just above idle all the way to red line.
    At idle, the F800GS will put out as much as 13.95 volts, but only if you aren't running any accessories. I know this is backwards of the norm, it's due to something called refraction that was not intended for the model, but trust me, the F8 spends 100% of it's time at 13.8 volts except for when idling.

    The F800 also typically has a 120mV AC ripple at a frequency of 3x engine RPM.

    The F800 is the bike the terrible imbalance in the Shorai occurred in.


    The Aprilia Caponord is the bike I have been using to see if higher voltage will balance Shoria LiFePo4 cells that are in series, which surprisingly, it will though slowly and likely with significant stress to the fullest cell.

    At a continues RPM anywhere between 2000 and 10,000, the Caponord hunts and pecks between 14.1 and 14.55 volts. 95% of the time it is between 14.2 and 14.4 volts. This behavior does not change at any RPM once the RPM is high enough to meet demand.

    Recorded in averaging mode, it comes out to 14.34 volts with no idling.
  8. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    Both the F8 and Caponord have SCR shunt regulators on PM alternators.
  9. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    give me a bit... will yank same Shorai LFX21 AH off G/S.... charge to 100% full, perform a measured discharge test. then charge back up to exactly 13.8v using HP regulated power supply at 1C.

    then will measure resting voltage after several hours, to find out how much charge was achieved by a charging system limited to 13.8v.

    by the way it doesn't hurt Shorai to fully discharge (9.2v) a few times. but for for all practical purposes battery is dead below 12.7v. ... consistently draining LiFePO4 down to 9.2 will kill the battery and only squeeze out a tiny bit more energy.

    how far one drains down any Lithium ion battery effect number of cycles battery will achieve to 80% DOD. battery life can be dramatically different. battery could deliver low as 500 cycles vs 2,000 cycles to 80% DOD.

    my guess is this the reason Shorai tell folks to never discharge below 80% DOD.

    [​IMG]
  10. Rajin Cajun

    Rajin Cajun Been here awhile

    Joined:
    Jan 30, 2007
    Oddometer:
    724
    Location:
    The DEEP South
    Cy, have you had a chance to try out the Powerlab balance charger?
  11. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    unfortunately no ... still trying to get the correct 4S adapter needed for Shorai. the assortment of balance boards and cables are designed for RC battery packs. which all have a male lead that plugs into balance board.

    Shorai uses a female 4S cable. so been having fun coming up with a proper cable to fit to Shorai's balance port.

    minor issues with getting correct cabling to access the power of Cellpro Powerlab 8 is well worth it. No other charger allows me to discharge at 1350 watts. or charge at 40 amps. Powerlab 8 is the most powerful RC LiFePO4 charger available.
  12. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    Little more info on the sickly Shorai LFX18 Lost Rider sent me.

    I got it fairly balanced, after 12 hours of standing with no charge or load, all cells are within 0.5 mV.

    Interesting thing, If I look at each cell out to the micro volt, you can watch each of the first 3 cells hold steady and the last cell slowly count down. I can't measure that many digits accurately on a 4 volt range, but i can measure repeatably.

    Cell number 4 has a self discharge rate much higher then the previous 3, and cell 4 was the one that arrived way discharged compared to the others.

    After charging the now balanced Shorai, there is still an 11% disparity in cell capacity measured by running each cell down to 3.2165 volts at a 1 amp rate.

    I'd love to see performance on a newish LFX18 to get an idea if this battery was damaged in service, defective, or if Shorai just has shit QC in cell matching.
  13. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    starting charge/discharge/charge cycle

    Shorai LFX21 AH (pb eq) is charged to 100% full
    [​IMG]

    starting regenerative discharge at 99% full
    [​IMG]

    1,700 mAH discharged at aprox 7.5amp
    [​IMG]

    4,071 mAH discharged
    [​IMG]

    discharged to 5,386 mAH final volt 12.64v indicating discharge stopped at 90%
    [​IMG]

    battery is quickly charged to above 12.86v then stopped to transfer to HP regulated power supply
    [​IMG]

    battery is charged at 1C at exactly 13.8v ... HP regulated power supply will terminate charge at 13.8v
    [​IMG]

    Shorai LFX21 AH at 1C is quickly coming back up
    [​IMG]

    13.76v .. charge rate is down to 500 milliamps
    [​IMG]
    [click to show pic]
    [click to show pic]
    [​IMG]
  14. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    never mind, thought u were balancing, checking balance is next step i'm assuming
  15. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    what I'm doing is duplicating F8 charge rate of 13.8v with HP power supply set to terminate current at 13.8v. after battery reaches 13.8v ... will allow battery to rest to determine charge level achieved.

    yes I can check balance, but that's not the purpose of this particular test.

    Shorai LFX21 AH (pb eq) is now 13.78v charge rate is down to under 50 milliamps. at this rate will take another few hours to reach 13.8v. for all practical purposes we are there.

    [​IMG]
  16. JoelWisman

    JoelWisman Long timer

    Joined:
    Apr 22, 2009
    Oddometer:
    1,392
    Location:
    STL, MO, USA
    Thats interesting, the down to 50mA part. That means there is high resistance to any more current going through the battery at that voltage, yet those cells are no doubt still charging, very slowly.

    If current continues to fall off, it's clear why they can get out of balance at that voltage!

    With a lead/acid battery of motorcycle battery size, around 100mA+ will continue to flow at 13.8 forever. Full cells will continue to pass appreciable current and expend it by gassing electrolyte which in a VRLA will recombine without venting.

    It doesn't harm the full cells but continues flowing current so any lower cells will all come up to full, hence, the battery is self balancing.


    If the current acceptance of the Shorai continues to fall off at 13.8...... Of course it would get out of balance!

    Minute differences in capacity, internal resistance, self discharge...... one cell ends up lower. BUT the fullest cell stops flowing current and over time the imbalance compounds till you have a light paper weight.

    Can you continue for some time and see how low the current gets?
  17. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    current acceptance of Shorai battery at 13.78v is higher than the 50 milliamp rate measured. meaning reduction in current as battery approaches set 13.8v set point is caused by HP regulated power supply not battery. closer to set voltage of 13.8v charge rate will decrease until it terminates entirely at 13.8v. note I set the 13.8v limits on HP power supply before applying load.

    if one delivers say 18amps or 3C to Shorai battery would swallow amps until battery reaches aprox. 90% full, then slowly taper downwards as battery approaches fully charged state. this is providing voltage delivered to battery remains constant.

    at 13.8v LiFePO4 chemistry is also approaching fully charged state and is slowly tapering downwards in terms of further current acceptance. to find out difference one would increase charge rate to 14.4v after 13.8v is reached. then measure current battery accepts at higher volt.

    higher the volts, higher the current rate pushed into battery. it's acceptable charge at 3.65v per cell rate to reach fully charged state of 3.6v per cell.
    _______________

    Shorai battery has been resting for 4 hours since it was charged to 13.78v.
    resting voltage is 13.74v ... this means F8's charging system terminating at 13.8v, still results in a fully charge LiFePO4 battery. provided long enough runtime is allowed to for battery to reach 13.8v state. end of charge would be reached faster at 14.4v vs 13.8v. but the differences would be small.

    if Shorai battery had been charged to fully charged state of 14.4v. after resting overnight. voltage would still be about 13.7v.

    -------------

    10 hours later... 13.73v resting ... proves charging LiFePO4 battery to 13.78v still achieves a fully charged battery. difference is how long it takes to top off the final 10% or so of charge cycle. noticed after R80G/S voltage regulator was changed from 14.2v to 14.5v. how long it took to fully charge LiFePO4 battery went down. at 3,000 rpm, amps delivered to battery went from 19.5amp to 22.5am with same load of lights and heated grips.

    when dealing with heated gear in winter conditions. at slower discharge rates actual AH ratings comes into play. then if riding in slow traffic conditions before putting bike away for the night. 15-20 minutes riding in those conditions with heated gear on could easily deplete a 6 Amp hour battery to under 70% condition.

    so yes contrary to what some folks say about amp hour ratings doesn't really matter. yes it does under cold conditions with heated gear on. then factor extra amps needed to wake up LiFePO4 battery the next morning before starting bike in cold conditions.

    [​IMG]

    10 hours later resting 13.73v
    [​IMG]
  18. Speedo66

    Speedo66 Transient

    Joined:
    Jan 14, 2010
    Oddometer:
    7,248
    Location:
    Cin City, OH
    Anybody have any experience with the BikeMaster TruGel AGM batteries?
  19. _cy_

    _cy_ Long timer

    Joined:
    Sep 25, 2011
    Oddometer:
    6,731
    Location:
    Tulsa, Oklahoma
    finally got correct lead with double male 4S to make final connection from balance board to Shorai battery.

    Shorai battery was charged to 13.78v by HP regulated power supply. then allowed to rest for a few days. question was raised if an F8 charging system regulated to 13.8v would fully charge a LiFePO4 battery?

    answer was documented above ... yes a charging system limited to 13.8v will charge to full. but where difference is how fast the last 10% or so of charge is stuffed in.

    Powerlab 8 has an A123 extra accurate charge mode with balancing. that's what I used on Shorai LFX21 AH after sitting a few days. last charge was to 13.78v .... Powerlab 8 registered Shorai has being 92% full starting point. all cells all balanced to zero within a few milliamp charged.

    very low total milliamps delivered to Shorai before terminating at 3.65v per cell. in accurate charge mode with balance. with a resting 14.30v four hours later. Powerlab 8 really stuffed Shorai to completely full.

    Shorai accepted 190 milliamp @ 3.65v per cell during last phase of charge.
    [​IMG]

    all cells show 3.65v delivered
    [​IMG]

    volt drop to 3.62v after full charge reached
    [​IMG]

    resting after 4 hours .. an amazing 14.3v after getting stuffed full by Powerlab 8 in accuate charge mode.
    [​IMG]

    resting voltage 14.24v after sitting overnight
    [​IMG]
  20. write2dgray

    write2dgray n00b

    Joined:
    Mar 15, 2012
    Oddometer:
    2
    Excellent testing here - nice work! Also, one heck of a vetting process to get registered here :).

    I hope you are enjoying the PL8 and beginning to flex it's capabilities. Lithium's the way forward for starter batteries in motorcycles (and cars). Next is replacement of the IC engine!

    Hope the graphs are helpful and please let me know if there are ever any questions I can assist with.

    Cheers,
    David
    ProgressiveRC