Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Solar series battery charging problem

  1. Jun 21, 2017 #1
    I have a large array of solar panels, which are 12v panels. I have wired them two each in series to produce 24v at the array and carry that to a 24v charge controller which then connects to two 12v batteries wired in series for 24v, to power a 24 volt inverter.

    But I have observed a problem with this configuration. The problem is that while the system is still charging up the battery bank, if I separate two batteries and measure the voltage on them, the first in series (from the + connection end) will be fully charged at 12.6v and the second will be less. For example, when my voltage on the array shows 28v during charge, two batteries separated and tested for voltage may show one to be 12.6v and the second one at 10.8v. This means to me that one of the batteries is being overcharged while the other battery is not yet charged. It usually results in the first battery of the pair failing. Never the 2nd one.

    I am not certain but I get the feeling that the cause of this is that the power coming in to charge the batteries must pass through each cell to get to the next? Whatever the cause of this is, it causes battery failure as the first battery in line in the series configuration gets overcharged while the 2nd battery is still charging to get up to full potential.

    After consideration of this problem recently, I wondered if it might not be better to carry the common between two solar panels on the roof and tie it to the common (+/-) post that tie the two batteries together. In this way both batteries would see 12v from the array and charge simultaneously, rather in a cascade fashion. The solar controller would still control the charging of the array as both the + and - of the array go through it, so when it turns off, there would only be the common leg connected directly to the solar array, hence no path for power to bypass the charge controller. Likewise the inverter would not change it's connection and would still see 24 volts.

    I would like to understand the cause of the difference in voltage (is it really due to charging in a cascade fashion?) I know that it is not a bad battery issue, as I have twenty 265Ah batteries in the bank (10 banks of two batteries each) and all batteries show the same voltage when tested. So it's not just a battery going bad.
  2. jcsd
  3. Jun 21, 2017 #2


    Staff: Mentor

  4. Jun 21, 2017 #3


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2017 Award

    So you are saying that the one reading the higher voltage tends to fail first?
    What happens if you switch positions of those two batteries?
  5. Jun 21, 2017 #4


    Staff: Mentor

    Nope, the undercharged battery will fail first due to "sulphation" of the plates.

    (BTW we are talking about lead-acid batteries, correct? And you have checked the water level in all cells, correct?)

    Swapping position won't help, because both get the same current in series.

    The better charge controllers have a built-in "equalization" cycle for exactly this reason. Check yours, you may have it but not know about it.

    If you don't have it, then you periodically (30-60 days) need to separate the batteries and put them through an equalization charge at 12v, not 24v as a maintenance procedure. My charger's equalization cycle brings each battery to 15.2 volts for 60 minutes, once every 30 days. You need to check the water levels frequently.
  6. Jun 21, 2017 #5
    Yes. The higher voltage battery is the one that always fails. Never the lower voltage one. I never let my array fall below 24 volts (1/2 charged.) So you see one battery is at full charge (12.6v) the other is at less than 1/2 charge (10.8v) while system voltage shows 1/2 charge or greater (24.5v). Originally I believed it was too much charging current so added more battery banks to cut down the amount of current per bank. (Please note that the voltages given are examples, and not actual measurements, they are just to illustrate the problem.)

    It seems that the first cell in the first battery (the battery that has it's + terminal connected to the array) fails always. Swapping batteries around gets rid of the problem, but handling 200 pound batteries isn't any fun ( I have to physically swap the batteries position due to the wiring. All wires are equal length to a central point where all connect to the power so that there is no difference in potential that any bank sees from the rest.)

    I equalize the batteries regularly (once or twice a month at 32v for 2-4 hours.) These are 7-10 year batteries that are failing in under 2 years. Not good. Very expensive problem to have. I have had 4 die now. I saw this same problem occur with golf cart batteries on a friends' array that was set up for 12v. So he had two 6v golf cart batteries in series. Same exact issue.

    I am not a novice when it comes to solar, as I live off grid and have lived off grid now for 2 years with solar being my sole source of power. I do regular maintenance.

    The array is about 5KW of panels and there is about 100 amps @ 24v coming in from the panels at max.
  7. Jun 21, 2017 #6

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Do i understand you have ten 24 volt 'banks' wired in in parallel ? Do you equalize them one at a time or all in parallel?

    Do you have any 12 volt loads connected across just one of the two batteries that's in series comprising a 'bank' ??
    The only other wire on "the common (+/-) post that tie the two batteries together" should be a high impedance voltmeter..
  8. Jun 21, 2017 #7


    Staff: Mentor

    What is the failure mechanism? Cracked plates are most often caused by low water levels.

    That is almost an excessive amount of equalizing. I would expect that to evaporate away lots of water.

    Are the batteries flooded or AGM or gel cell?

    If flooded:
    How often do you pull the caps to check water level? I recommend once per equalizing cycle. That is quite a chore if you have 120 cells..
    Do you have a watering system that distributes water to all cells without the need to pull the caps to check?

    If gel cell:
    The max charging voltage for a flooded battery is 14.2, for a gel cell 13.8, for an AGM (don't know).
    The max voltage for gel cells is very sensitive to temperature. If the gel cells are at 120F or more when charging, failure is likely.
  9. Jun 21, 2017 #8
    20 Flooded Cell batteries in 10 banks of two each. In other words, pairs of batteries wired together for 24 volts, the pairs wired in parallel.
    The only load is the inverter itself.

    Batteries are 8D's
    Due to the low amount of current, I have very little water loss. At peak each bank gets at max 12 amps or so. That is why I equalize them a bit longer and more often.
    I have to manually add water, but as there is never much water loss, it hasn't been a problem.
    I use Trojan's guide on voltage for charging, which is 28.2 - 29.4 volts, I use 28.8v as the high voltage setting on my charge controller.

    When a battery fails, I can charge it to full voltage but soon after disconnecting from the charging source, after the surface charge is gone, the voltage will go to 11.8v where it then stays. It will not stay at 12.6v. To me this means a half cell problem (full cell failure would be 1.6v) which I believe may be a short internally, but I can not understand by what mechanism this would happen on all batteries regardless of type. As I said this same thing happens on 6 volt golf cart batteries as well. I came to the belief that too much current was passing through the first cell, but at 10-12 amps, with this large of batteries I couldn't see that being enough to do this kind of damage.
  10. Jun 22, 2017 #9

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Four failures out of twenty batteries, i think you said.

    Are they close together ? Batteries in parallel need to be about the same temperature. Are they all in the shade?
    All about same age? We're not discussing recycled batteries, are we ?
  11. Jun 22, 2017 #10
    Your frustration is understandable - it is quite an expensive investment.

    If possible, could you provide a couple more bits of information:

    1) You say the 12.6v and 10.8v are illustrative voltages. Are the "actual" measurements in the same ball park? If not, any chance you could relate figures from actual voltages you measured?

    2) Have you ever equalized a pair of batteries exhibiting the problem, individually - i.e. not as a pair. So each battery equalized at around 15v - 16v?
  12. Jun 22, 2017 #11


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    Using a mix of batteries (age / model / temperature) is best avoided. Not a pleasant prospect to change them all at one time if that's much sooner than they're guaranteed for.
    I wonder if it could be worth while checking the wiring between the batteries in detail? Sounds a bit daft but it wouldn't cost any money to measure voltage drops between adjacent + and - terminals etc..
  13. Jun 22, 2017 #12


    Staff: Mentor

    We're the failures sudden or gradual?
    Do you keep records that can be used to plot trends?

    The Trojan Battery Company promotes their reputation for making superior batteries. Contact them and ask if they would be willing to do a complete post mortem on one of the failed batteries. They may even be willing to send someone to look over your entire installati M.

    If the cause is sulphation, then it might even be recoverable.

    Here is a paper that claims to present a fault tree causal analysis of lead acid battery failure.
  14. Jun 22, 2017 #13

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Like that idea .

    I've been wracking my brain too. Old troubleshooter instinct , plus i have OCD tendencies.

    When you're baffled it's helpful to try baseless random measurements and see if anything peculiar shows up. Ever watch a veterinarian poke gently at a puppy until it yips?

    Bernie separates his cells to check for failure.
    Pick a pair that's been having trouble. Don't disconnect them, just measure... What is voltage across each when being charged? And when loaded? Check the AC component too by switching the dmm to AC. Feel one with each hand - are they same temperature?

    Since i knoiw nothing about your installation.
    Is that inverter's input filtered ? Reason i ask is
    Old fashioned SCR inverters take huge gulps of current at twice line frequency.
    Large capacitors adjacent inverter are necessary to keep them out of the battery bank.
    I once had to add capacitors to old fashioned SCR inverters to tame our battery voltage. They modulated it at twice line frequency , imposing peak to peak ripple about 75% of battery voltage . We placed 1 microfarad per milliamp of inverter current, about 60,000 μf, at input terminals of each inverter .

    PWM inverters draw high frequency current pulses. Are his filtered ?
    What about that charge controller output? Is it pwm too?
    Reason i mention that is , knowing nothing about his physical battery wiring layout i wonder if it has substantial inductance?
    If so it's possible some battery pairs carry the lion's share of the high frequency current. High frequency pulses heat the electrolyte lowering cell voltage, perhaps starving other battery pairs on float..


    While he's checking DC, measurements of the AC ripple across several batteries might give a clue.
    Draw a physical map and write them on it looking for a pattern.
    But the meter will have to have frequency response out to inverter's and charger's pwm frequencies. Old fashioned Simpson 260 using "output" jack is good to about 40 khz. Check your dmm book...

    Yes it's a long shot . But looking into it might lead to some other shorter shot.

    Necessity may be the mother of invention but desperation is at least its good uncle.

    old jim
    Last edited: Jun 22, 2017
  15. Jun 22, 2017 #14


    User Avatar
    Gold Member

    Very weird.
    @BernieM , can you tell us what the average daily depth of discharge is?
    At 50%, a 2 year lifespan is standard for flooded lead acid.
    I'm still scratching my head as to why the order of the batteries makes a difference. It makes absolutely no sense.
    And your one paragraph is rather incomprehensible:

    "carry the common"?
    "cascade fashion"?

    Can you doodle a schematic for me. I've had a lifelong difficulty with language, but, I've found that a picture is sometime worth more than a thousand words.

  16. Jun 22, 2017 #15


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    The appropriate 'language' for describing this sort of problem is Pictures. I am confused by written descriptions, too.
  17. Jun 22, 2017 #16


    User Avatar
    Gold Member

    I saw someone post one day; "I don't understand pictures, nor equations. Just tell me what the problem is!"
    I'm quite the opposite, and the lady who posted that has a PhD, so I just assumed I was stupid.
  18. Jun 22, 2017 #17
    All batteries are the same age and two different brands, but same specs and type. Four failures out of 24 batteries originally, I now have 20 left. They are about 2 years old at this point.

    They are not recycled batteries. They came out of a 12v solar power system that someone was upgrading and were 1 year old when I got them. I tested them all and they all tested about the same on a load tester. They are two different brands but have the same specifications. Both brands have failed, so failure is not due in my opinion to a difference between the brands.

    There is about an inch of space between batteries. They are housed in a small room on the porch which is covered and at lower temperature than outside currently (it's hot here but these failures have happened when it's cold too, so it isn't heat related.) I also have temperature compensation on the charge controller and the sensor is near the batteries. Difference in temp from top batteries to lower batteries is about 3c max when it's hot outside. That's the best I can do with where I live and the conditions I have to deal with here.

    1) --- 12.6v and 11.8v. I put one of the failed batteries on a stand alone solar charging station with 250 watts available to charging. After a month of charging cycles with no usage on it, it still would drop to 11.8v and stay there after fully charging up.

    2)--- I also ran some equalizing cycles on it during that time. I was baffled by the problem and was trying to figure it out myself. I even went so far at the end to put a bit of hydrochloric acid in it to dissolve the sulfate (hydrochloric dissolves lead sulfate but does not dissolve pure lead and after a while will dissipate, leaving the cell as chlorine gas.) This changed nothing either. So I was relatively certain at this point that it was not a sulfation issue, though obviously I could be wrong.

    These are considered 7 year batteries as I recall. If one keeps depth of discharge less than 50%, one should be able to nearly double the life span. The life span of a battery is considered to be the number of full discharge cycles the battery can tolerate without failing. So a 50% discharge would take 2 discharges to that depth to equal one complete discharge. This is how I run my setup, I do not allow it ever to get below 50% (24v.)

    When I said 'carry the common' I meant to connect the common connection between two solar panels on the roof, where they are joined in series, and connect that directly to the battery's common connection, where they join in series. In this way each battery would have a separate solar panel circuit charging it. Simplest model would be two batteries, each with their own solar panel to charge them, then joining the two batteries in series. In this way there would be 24v across the batteries as well as 24 volts across the solar panels, yet each battery would only see 12v passing through it's terminals.

    What I mean by cascade fashion (and I will make no illusions that I understand battery physics here) is that I visualize all the power to charge both batteries having to pass through the first cell, the current moving to the next cell, will be less than the first cell sees as the first cell has used up some power in charging, and on and on through to the last cell of the second battery. Perhpas this is an incorrect picture of how it works (and I am sure someone will correct me if it is) but I see it that the + terminal sees all the incoming power to charge both batteries, where the + terminal at the second battery sees only half of that due to the power dropped in the first battery to charge it. That's what I meant by cascade fashion. Though the voltage potential is the same across both batteries, the amount of current would not be as I see it.
    I have another small array that I put in place that is 12v and I don't see this problem there.
    Last edited: Jun 22, 2017
  19. Jun 22, 2017 #18
    Voltage is equal across the batteries while not disconnected from each other. After disconnecting, one will show a lower voltage, the other a higher voltage, but the voltage across both will be whatever they were charged to as a pair.

    There is no source of a/c here as I am off grid 20 miles from a city, nearest power lines are 5 miles from here, no generators on the system, etc. Sources of high freq a/c would be a microwave oven, a computer (which isn't always on) and a tv perhaps.

    I have been using a 8KW/32KW true sine wave inverter. It has been a great inverter and gone through a lot of tough tests over the last couple years and has performed excellently. It's made by Powerjack. It's split phase 220/110.

    The old charge controller I used was PWM, recently I have put on a relay type that can handle 300 amps. Kind of dumb but it doesn't shut down when it gets too hot. Before I had issues when the amperage going through the controller got high and temps were high. I have almost finished an underground building to put the batteries, inverter and charge controller in to guarantee that the batteries don't see any high temps. Currently they can get up to about 95F for short periods (a few hours at the end of a day.)
  20. Jun 22, 2017 #19


    User Avatar
    Gold Member

    Everything has an "advertised" "rated" lifespan if kept within its normal operating parameters.
    I think the first battery in my first car lasted 20 years, even though it was probably rated for about 5. Not sure. That was nearly a half century ago.
    Not sure where you heard that. It's not true. You should stop listening to whomever told you that.

    ps. It's partially true, but in a hand-wavy, really bad maths kind of way.

    pps. Are you still living in the Mojave desert in Arizona? Sorry to be so nosy, but I checked out your profile, in an effort to figure out your battery's environment, saw that it was blank, and checked out some of your posts.:redface:
  21. Jun 22, 2017 #20

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Meh. Take heart , my friend.
    We picture thinkers are in good company.

    They do have a following. It's a pretty good bet they do draw some sort of pulsing current
    this guy adds a 18 microhenry inductor to reduce idle current draw.

    That is very confusing. They're not all joined in the middle, are they? All those common +/- junctions tied together? Something's gotta be holding voltage equal.
    Sounds to me like internal battery failure but i'm at a loss as to what causes it. You've described what should be a quite good system.

    divided by 12 cells = 2.4 volts per cell
    Sounds like you are doing everything right.

    Gotta be something very simple or something very esoteric.

    That huge inverter has to draw pulsating current. Were this mine I'd look at the AC ripple across every one of my batteries with inverter running and with it off, objective being to see if they're sharing it equally. If your dmm has a frequency button, check that, too.
    Fortunately it's high frequency so if it's there it will be easier to filter than was my 120 hz. A few good high current oil filled 'SCR commutation' capacitors right across its input terminals might do the trick. This CDM Cornell-Dubilier one's good for 60 amps rms.
    Wish i knew how to measure the ripple current your inverter draws. Most clamp arounds i've seen are only good to 2 khz or less.


    you might get away with less expensive aluminum electrolytics, depending on what frequency you find.

    Wishing you Good Luck
    and when you solve this one i hope you'll enlighten us.

    old jim
  22. Jun 22, 2017 #21
    Yes I live in the Mojave desert west of Phoenix, where yesterday it was 121F here. So my battery bank and solar are hugely important and the difference between life and death literally, as I live remote, 1/2 hour drive from nearest town. If power goes out here it could get really serious (though I do have a generator for backup, which I nearly never run.)

    If you just go to a good search engine and search for 'depth of discharge vs battery life' you will find plenty of corresponding articles and webpages that confirm the same thing, including some that are reputable.

    The post about that PowerJack inverter is old. I discussed this with the engineer/inventor/owner of Powerjack in China, and he says that all of the issues in that post have been dealt with several years back. I find very little not to like about it, especially considering it's ability to output 4x it's rated output for up to 2 minutes (instead of 15 seconds) without shutting down to protect itself or frying itself. Many times while my system was pretty much under full load I have fired up a big power hungry table saw and ran it for a minute or so while I did something I needed to do in a hurry, and the inverter didn't whine about it. I have two of them, one as a backup, and the backup is a year older. It has a few idiosynchrasies that the newer one doesn't, and a more recent model I bought for a friend and his house, is even better. So they are continuously improving their inverter.

    I have an o-scope and have checked the output of the inverter out and it's really clean and a very true sine wave, that's really stable. I have never checked the input of the inverter for a/c on the scope, though I did check it with my meter with no a/c shown on the batteries. I put the meter on Freq and did a check and it shows .001 which is the default display so it's not detecting any a/c up to 3khz (range of this meter.)

    I am suspecting that it might be a combination of not a high enough amount of current during equalization and battery electrolyte stratification. It's the only thing that makes sense to me. I guess I can disconnect 1/2 of the array during equalization and see if that increase of current per battery changes anything. I should be able to tell right away by disconnecting two of the batteries and measure the voltages on them after equalization. If they don't read the same voltage then I still have the problem, if they do read the same then it's fixed. So I guess that's my next step.
  23. Jun 22, 2017 #22
    I need to make a clarification here. I see that part of the confusion is what is happening before a battery failure and then after the failure. So let me put it like this:
    Ignore the charge controller and just imagine I have two 12 volt solar panels connected in series to produce 24v at the array. This in turn connects directly to two 12v batteries, likewise connected in series. As it should be.

    At any time, if I measure the battery voltage across the two batteries it will show, of course, the system voltage, which will reflect the solar input and load. If I disconnect the two batteries from each other so they are no longer in series, and measure the voltage of each, they will both have the same voltage.
    But when a problem occurs, if I disconnect the two batteries, one will show a lower voltage than the other. The lower voltage battery will always be the battery that connects into the system at the negative terminal (battery #2 in the series.) But at this point neither battery is failed. Both will charge fully and keep the charge if charged separately. I have found that I can delay this problem if I swap them out physically so that the battery that had it's negative terminal connected to the array is now the one that connects to the positive of the array and vice versa.

    When a failure finally occurs however, the failure is the battery that had it's positive terminal connected to the array. At this point when I read the voltage of it, it will be .8v lower than fully charged, regardless of how I try equalizing or charging it thereafter.

    So it seems to me that if sulfation was the cause, it should be the battery that spends more time in a discharged state before the failure, (battery #2,) not the one that is constantly at a higher voltage (battery #1.) That's why I rule out sulfation. The only difference between why one has a higher charge state than the other in my mind is current. That all the current to charge both the batteries must pass through the first battery (the one with the positive connection to the array) where the positive terminal on the second battery is carrying only half the current of the first one, because of the power used in charging the first battery on its way to the second. Although both batteries see the same voltage, the first battery sees double the current than the second one. If I have that wrong let me know.

    So it is the current that I feel that has a role in this but I don't know how it plays a role.
  24. Jun 22, 2017 #23

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Yes i think you have it wrong, your words do not describe behavior of electricity.
    Two 'things' that are in series have the same exact current through them. They can have different voltages, though.

    Only way for them to have different current is for a leakage path around one of them to appear.
    That might be spilt electrolyte on top of the battery that bridges its terminals
    or something connected to that +/- junction between a pair which you've already said isn't there
    or a hole in the case of a battery through which current gets from the electrolyte to a conductive battery rack or something.
    One's imagination can run away.

    This is a decent little writeup
    http://batteryuniversity.com/learn/article/serial_and_parallel_battery_configurations Batt_U_Series.jpg

    Thought experiment:
    Imagine yourself very small and seated at the observation point with a "charge counter", a hand held gizmo that increments by one count every time a single quantum of positive electric charge moves past your observation point in the direction of the arrow.
    When it increments 6.24 X 1018 times a second that's one amp of current flowing.
    Look up "Coulomb".
    An amp is one Coulomb per second flowing past a point. It's that simple.

    Observe those charges can't get out the top or bottom of the battery string because of the insulation..
    All the charges that go in the left end must come out the right end, traversing all four cells.. You could observe anyplace in the series string and get the same current.
    That's why things in series have same current. It's their voltage that can differ.

    not talking down here, just it's important to get the basic definitions nailed down straight early on. Saves a lot of unlearning later.

    Rethink your sulfation conclusion?

    old jim
  25. Jun 23, 2017 #24

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Of course , some folks consider current to be movement of negative charges because usually it's electrons that carry them.
    Hence the old conundrum of "Conventional Current"(positive charges in motion) versus "Electron Current"(negative charges in motion).
    They work identically and give the numerical same answer, just on your drawings your current arrows point in opposite directions .

    old jim
  26. Jun 23, 2017 #25
    http://journal.esrgroups.org/jes/papers/4_2_2.pdf was a link provided in an earlier post in this thread. In it it shows an electronic model of the cells in a battery. If you will notice the R's are in series in it. Current drops when it passes through a resistance. So there is no way that the same amount of current that enters the battery may get past to the next cell having passed through a resistor, stated simply. Start with whatever current you want on either side and move through and calculate it out. Current drops through each resistive load. Not true? Or another way to put it would be that each cell consumes power, which means if you have a fixed amount of power entering the battery, as power is consumed in charging a cell, there is less power available at the next cell. How can you consume electrons and do work with them, and not reduce the overall volume of electrons with a fixed input of electrons to begin with?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted