New Pack of 12 Prismatic Cells Install Questions

[Dave McCampbell]

Hi All,

Because of problems with our earlier 8 RJ house bank prismatic cells we are expanding our house bank on our catamaran from 540 to 840 ahrs rated, using new 12 EVE LF280K cells from Docan Power.  These cells are all from the same batch.  They have been closely tested and matched at Docan Power with capacities at 308-310 ahrs each, internal resistances at .15-.16mohm and are resting at 3.261-3.263 vpc.  Those figures are exceptionally close and much closer than my earlier RJ cells.

We currently plan to assemble the 12 cells into one compressed bank of 4 series sets of 3 parallel cells each (4S3P) for a 12 volt nominal bank, one Electrodacus SBMS0 BMS and an external NEEY active balancer to help with balancing.  We need these cells to remain in service for at least 10 years and as closely balanced as possible.  Charging is primarily from two nearly equal sets of 24v nominal solar panels, total 1560 watts with two Victron 50 amp MPPT controllers.

Here are a couple of questions to confirm my thinking before assembling this pack:

·       Given the matching and testing that has already been done, and very tight capacities and internal resistances, is there a need to recheck the capacities and top balance the cells before we put them in service?  We do have equipment to do that, but I have concerns about unnecessarily stressing these already well matched cells.

·       Other than using a 10 amp active balancer with single balance wires on each of the 4 sets is there some other way to ensure that the parallel cells within the sets remain in balance?

·       Should I give any consideration to arranging the 12 cells in 3 separate batteries with 3 BMSs and active balancers?  With BMSs like the SBMS0 I don’t see how that could work at HVD/LVD without the danger of suddenly disconnecting the separate batteries from the charging and load busses.  Using direct control or smaller external relays to disconnect the charging or load sources from multiple batteries if a cell fails seems like an unworkable plan to me.  I guess that can work with internal Mosfet BMSs disconnecting individual batteries but not external relay BMSs like the SBMS0.  Also, there are Dacian’s concerns about multiple batteries in parallel and the significant expense of extra equipment required.

Thanks in advance for your thoughts and advice.  Dave SV Soggy Paws

[Marinepower]

Docain would know way, way  more than me, and Im sure he will chime in.

But for what it's worth....

I have 4p4s setup with the electrdacus and its been working fine for a couple of years now (admittedly only 105ah cells).

The safest way to do it would be 3 separate batteries and 3 BMSs, but I would be tempted to try the 3P4S set up and monitor closely for a few months to see if yo can get away with it.  I dont know how big your typical loads are, but i think big parralle packs stay better balanced if your average loads are smaler than 0.2C or less and you charge back up 100% on a fairly  regular basis

I also have a cheap active balancer (5 amp) wired but not connected to my 4P4S pack.  My idea was to only connect it if the cell balance started to drift significantly.  So far, to my surprise, that has not happened & I have never had to use it.  I have heard mixed results with active balancers connected 100% of the time, which is why i avoided it (except for monitored maintenance.). The Neely may be better designed for this pourpose however.... 

If you are going to 3P4S, I would make sure you have beefy busbars btw the  parallel cells with very clean, tight connections.  Also, be sure your cell layout is optimal for current sharing btw the parallel packs. 

I would not skip the cell top balance, but I do not think you need to go past 3.55v per cell to get a good top balance with Lipo.  IMO you should not do damage to cells if you get it up to that voltage fairly quickly and then discharge them to a lower the SOC with in a short time once it gets there.

MP

 

[sailingharry]

I am going down a very similar path, but upgrading from AGM. I have 12 314Ah cells sitting in my basement.  The resting voltageEs and the IR are very closely matched.  I won't (as of now) be using an active balancer, but that's something I'll be watching for.  My peak loads are ~200A, and my average daily demand is ~250Ah (ie, a very lightly loaded bank -- maybe even an over-designed bank...).   I've thought about much of your questions already.

* I am not intending to top balance, or anything else, before installation.  My cells are nicely at a low SOC (by voltage, maybe 30%).  To top balance would require a compression frame, and a low voltage charger, and a LOT of time (I'd have to pump in about 7,000Ah at 3.5V to bring 12 cells from 30% to 100%!!!).  My intention is to fully assemble, charge to 100%, and watch/wait.  If a few days in, the balancer is still struggling, well, I'll have to reconsider.  I could buy an active balancer or I could "manually balance" (a wire and a resistor between cells as needed).  I do have shore power available, so I can charge (and balance) 24 hours a day.  And I'll have about 3 months between connecting the charger and my first sail, so I have the luxury of time.

*  Keeping parallel cells in balance.  The classic way to "balance" cells is to put them on a high quality charger, in parallel (usually the entire pack re-configured as a 12P 3.5V pack), bring them to ~3.5V, and leave them for a while.  In our final configuration, every single time our batteries are near-full, the 3 cell groups will all be sitting at 3.5V per pack, in parallel.  Sounds like top balancing to me!  It is fundamentally impossible for 3 cells, in parallel, at >3.45V, for over a couple hours, to not be in balance.  An active balancer will only address the balance between each set of 3, not the cells within the 3 (unless I misread your entire question here!).

* I've considered 3 banks.  It makes my head spin, and I chose not to.  I consider an "allow to charge" signal to be a absolute MUST HAVE, and that pretty much rule out all FET BMS.  I want to disable all charge sources BEFORE a disconnect.  But, in theory, you could design a system that works -- even with 3 Electrodacus.  Each battery would have it's own disconnect, and each 'Dacus would control that disconnect directly.  However, the design of the 'Dacus is such that it should NEVER happen.  Then, take the signal wires (Type 1, etc) in series.  So, for instance, a single high cell in a single bank calls for "stop charging" and opens the Type 1.   ALL charging stops.  This is not a really big deal, as all 3 banks are in parallel, and all 3 banks are probably over 95% SOC before this happens.  They will sit there, at 95% SOC, and patiently wait until that outlier gets balanced back down.  Similarly, if you set a low SOC signal (Type 3?  I forget) at say 20% (my intention on mine), the first bank to get below 20% will trigger the "loads off" signal (in my design, that is ONLY my inverter, with an audible alarm so I can take action). BUT, the ROI doesn't work for me -- 3 'Dacus, 6 shunts, 6 Class T fuses, 6 contactors, a few more large cables, ugh.  You could possibly get by with 2 shunts -- the shunts only provide "information" and not actual battery management, so if all 3 'Dacus were wired to the same shunt, each display would show current, SOC, Ah, etc for the "bank" and not for each battery.  Also, since the fuses are after the shunt, you could probably get by with only 2 (either with 6 shunts, or only two).  If you went with only 2 shunts, you would have to have only 2 disconnects (the disconnects have to be after the shunt) -- meaning you would wire the Type 5 (I think) in series, and any 'Dacus calling for a problem in its battery would take the entire bank off line. 

*  Marinepower cautioned on the size of the paralleling bus bars.  This was a struggle for me.  The series bars are the real challenge -- bus bars that are ACTUALLY capable of 400A are rare (they are "rated" for it, but take the dimensions to for instance the Copper Institute's bus bar table, and it's a fail).  I finally solved my concerns with series.  The parallel bars "in theory" take no current.  In practice, they do.  But I'm very comfortable with using bars "real rated" at 200A for this, as I just can't see meaningful current.  Even at the ends, my connection at the center cell means 1/3 of the current comes from the center cell (no bus bar current), and 1/3 from each side (meaning 133A each).  Even this can be mitigated by taking a big cable from each cell to the shunt -- I chose not to.

The fundamental problem with 3 batteries is that you have to monitor all 12 cells, and if ANY cell goes over/under voltage, you have to take action.  The FET style take action by shutting off the battery -- a "bad thing."  The 'Dacus turns off the offending problem (charge, load).  So to do it well, you need 3 BMS, and they have to have external signals.

[Dacian Todea (electrodacus)]

Dave,

Parallel cells will always be in perfect balance as they can not have a different voltage. The groups of 3 parallel cells will be no different from a single larger capacity cell.

The 12 cells including compression plates may be to heavy to handle by a single person unless you are fairly strong so maybe split the pack in at least two 2S3P just to more easily be able to move and install.

The 1560W array will be fine and you will be limited to around 100A * 13.5V = 1350W just slightly under the array power so perfect.

100A charging for that pack is in no way a problem as I charge a 8S2P battery with same 100A.  Discharge for me is also around the same 100A peak and is all protected with 120A breakers.

In my case I split the pack in 4  2S2P to have just 4 cells about 23 to 24kg each (around 50lb).

Then all 4 are installed on a board that has two wheels at the end so I can take it out fairly easy.

The SBMS0 is able to keep my battery in balance for over 2 years now but is your choice if you want to add external cell balancer.

 

On Thursday, January 9, 2025 at 10:55:18 PM UTC-6 da...@svsoggypaws.com wrote:

[Dave McCampbell]

Many thanks to all that responded to my post with questions regarding testing 12 new EVE LF280K cells from Docan Power and how to best to configure the new LFP house battery pack. 

From the responses it’s clear that not all agree on how best to do this.  But having learned quite a bit from our first effort 4 years ago, we have reached some conclusions.  We now believe that our original cells’ failure was due to several possible factors including original cell quality issues, connection and wiring resistance problems, significant balance variances that our small passive balancer was unable to resolve, and our own inability to recognize and address these things while underway much of last year.  

So we have spent quite a bit of time recently researching these issues.  From that effort it is a bit clearer now than it was then that some equipment and build methods will work better for us than others.  And your responses have helped to confirm what we now think is best for us.   As overseas cruisers in a marine environment our situation is different from others like off grid and RV builds.

Regarding testing, we plan to take all the new cells to a top balance of 3.55vpc which we think is a reasonable voltage to represent 100% SOC without stressing the cells.  Since there is very little additional capacity above that, we see little reason to go higher and this ensures all cells start life in balance.  Additionally we plan to only capacity test a sampling of 3 of our 12 cells between 3.55vpc and 2.8vpc.  This is what several experts recommend and where the defaults are set by Dacian at Electrodacus for HVD and LVD.  This will give us a good idea of the accuracy of the factory testing and our equipment.  We plan to use our excellent ZKETECH power supply and load tester for this work.

Regarding configuration, we plan to go with a 3P4S arrangement that puts 3 cells in parallel and 4 of those sets in series.   All cells will reside in one custom fiberglass box with compression and be monitored and controlled with one Electrodacus SBMS0.  We will use our excellent NEEY 10a Smart active balancer which, unlike most others, has fully adjustable parameters to ensure the sets remain well balanced.  The key to this arrangement is to ensure we start with well-matched cells.  And we need to be very careful with assembly of the wiring and cell connections so that resistance is kept to a minimum and the parallel cells can easily remain in balance.  This number of cells would allow us to reconfigure to 2P4S or even 1S fairly quickly in the very remote chance that there was some developing problem with a cell in the future. 

We think that there are just too many issues and not enough reason to compel us to make the changeover to parallel batteries now.  For us these issues include the added expense and complexity of reconfiguring to multiple batteries and internal Mosfet BMSs in order to protect our equipment in case of a HVD/LVD event.  Also, we don’t see the rational for multiple batteries as backup when we can easily reconfigure a multiple cell single battery if needed.  And then we also have the capability to switch to a start battery if we have a temporary emergency.  Evidence shows that well matched, quality grade A cells that are well cared for just don’t degrade suddenly, especially after the first few months of use.  And as is mentioned elsewhere, there is no way parallel cells will get significantly out of balance as can multiple batteries, unless there is an internal short, which will quickly be visible even on the series cells monitoring. 

Finally, since we already have high quality external relay BMSs and other equipment for a complete system with backups, there is no compelling reason we can see to make these changes.

Again, thanks for all your advice.  Dave and Sherry, SV Soggy Paws, Still In SE Asia

[Dacian Todea (electrodacus)]

A separate balancer is not needed to keep that battery in balance. If that cell balancer does not know if battery is charging or discharging it will just be counterproductive (meaning it will create imbalance instead of balancing the battery).

As an example say one of the 4 series cells has just slightly higher internal resistance than the other 3 that say are perfectly equal.

So say the 3 equal cells are 0.3mOhm while the 4th one is 0.4mOhm

While charging with say 100A and internally cell voltages are say 3.2V at the terminal of the cells voltage will be slightly higher due cell internal resistance.

All 4 cells will read 3.2V while no current flows trough them but as soon as 100A of charge current flows in to cells the cell voltages will be:

0.3mOhm cell * 100A = 30mV drop on internal resistance so externally the BMS and balancer will see 3.230V

0.4mOhm cell * 100A = 40mV drop on internal resistance so externally the BMS and balancer will see 3.240V

Now the balancer depending on settings may start to balance the cell with highest voltage trying to top balance and keep the cells equal and that will be fine.

But if now the battery is discharged with 100A the cell voltages will look quite different.

3.2V - 30mV = 3.170V

3.2V - 40mV = 3.160V

So now the 3 equal cells are 10mV higher than the cell with higher internal resistance and a balancer that has no idea if battery is charged or discharged will try to balance the 3 cells resulting in creating an imbalance.  The SBMS0 will know if battery is charged or discharged and will only balance the battery during charging thus not imbalance the battery during discharge.

[Chris R8]

Sorry guys what you have are vendor matched good grade B cells but that's nothing close to manufacturer matched garde A+.

They maybe much better then what you previously had, which is also that generally the cell quality went significantly up to when you bought your first cells but these are not even close to real factory matched cells.

Grade A+ factory matched all cells have the exact same internal resistance over the whole SOC curve (!)and capacity I'd also identical to max 1AH difference.

You need to 

A) charge them all up and do capacity test all cells

B) put the weakest and closesest together in one pack, means the weakest 3 cells form one parallel packthen the next and so on

C) you need to do a top balance on all cells before mounting them into the packs

D) the bank must be absolutely symmetrically build and the busbars massivly oversized otherwise they interfere with their resistance of the cells, 4p with the 280AH cells is a 32cm long busbar and the resistance difference between cell1 and 4 adds to the cell resistance making even cells uneven or if unlucky the difference in resistance between the weakest and best cell is growing.

E)  the active balancer a 5A heltec is more then enough just need to solder 2 wires to the remote on/off and steered by an Extio 3 put at 96% SOC, this switches active balancing on when at 99% SOC and off at 96%. Even when you get a floating SOC that's still enough. You can keep the electrodacus balancing as thats super small and doesn't hurt anything.

Contradicting what Darcian said you need an active balancer above for >600AH bank capacity because the balancing current of SBMS is simply to small for >600AH. The problem with >2p banks is that also the parallel cells can also only equalize themselves in the knee to 100% so you need either a longer absorption time (which with SBMS doesn't work) or you can use an active balancer that also during discharge can restore the top balance and helps equalizing the parallel cells.

I have real grade A+ 272AH Lishen factory matched, all had 273AH and 0.12mohm resistance at 3.35V resting voltage.

Doing it Darcians way without an active balancer I got massive imbalances with a 4P4S bank over time even with perfect cells.

Then I add the balancer with type Extio 3 at 96%SOC. That restores the top balance in under a week being daily at 100% SOC in summer.

Now my bank us at 3 till 8mV deviation at 3.50V when latest 100% once all 3 weeks and also on lower voltage. No pack is peaking or the deviation rising when I run 500A loads.

Also 57 days without being at 100%SoC when reaching 3.5V at day 58 the deviation was 14mV and immediately corrected to below 8mV through the balancer. Most of deviation was that the cell pack needed to equalize itself in the knee.

I also have cutoff of SBMS at 3.55V but main charge source solar goes in absorption at 3.45V or 13.8V with 30min absorption and then 13.3V float. Means SBMS cuts all other charge sources but solar gives the 4P bank the necessary time to equalize and the active balancer to restore top balance. That's needed by high capacity multiple parallel cell banks.

This gives me rock solid top balance and deviation between 3 and 8 mV from.3.0V till 3.55V on a 4P4S bank with 1088AH that sometimes see 500A so 0.5C loads.

[Chris R8]

Sorry Darcian, 

With >600AH you need an active balancer that works ONLY in the knee, this can be done by using Extio 3 with 96%SOC or the balancer measures the cell voltages itself and only switch on if cell are above 3.4V during charge and switch off when cells are below 3.35V during discharge. If that range is a bit exceeded that doesn't matter, important is its not 100% on. Heltec has now a 5A version that switches insteadtself on at 3.4V and switches off 3.35V discharging.

The electrodacus balance current is by far too small to be able to balance 3 parallel 280Ah cells=840AHand is not capable to deliver enough current so a top balance can be restored. Also the fact that by cutting through electrodacus BMS there is simply 0 absorption phase which parallel cells need to equalize and a passive balancing cannot work during discharge, an active balancer can do this properly.

The voltage ranges you stated there is no balancing so you don't run into the problems you stated.
 

I am running now my bank like this and it fixed all the massive problems I had with imbalances.

Put an 5A active balancer from Heltec on you bank and run it with EXTio3 and 96%SOC and you will be surprised what this does to your bank which is in my eyes strongly out of balance, deviations of 40mV is by far too much for a 2P8S bank and requests actions.

electr...@gmail.com schrieb am Mittwoch, 22. Januar 2025 um 17:55:16 UTC:

[Dave McCampbell]

Hi Dacian.  Thanks for your input on balancing.  I understand the difference between balancing while charging and discharging and your enlightening computations.  Does that apply to both active and passive balancers?  

We have never balanced during discharge as the SBMS0, as you mention, does not allow this.  My concern is that my new house bank of about 900 ahrs of prismatic cells, even well matched and top balanced, is just too large for the 200ma passive balancer to keep up with.  At least that was my experience last year.  And I can't afford a repeat while underway for the Med next year.

My new balancer is a 10 amp NEEY Active Smart balancer.  It is much smarter as far as being able to set its multiple balancing parameters than either a common passive or active balancer like the Heltec series.  Several users have done extensive testing and the unit is now in its fourth generation.  It is an interesting and very smart piece of equipment used by several professional installers with good results.  Anyone not familiar with it can find substantial information on the internet.

I don't know conclusively what caused the demise of my earlier 8 cells, but I suspect it was a combination of problems described above.  This time I want to ensure I can cover all the possibilities so have invested in extra equipment like the balancer and new much better cells.  These are not Grade B cells (the same as Dacian's from EVE) and are very well matched, so they should be easier to keep in balance.  We will see when I install them on our boat in about 6 months.

Meanwhile, the additional 760 watts of solar panels and new DSSR50s are in place and working well.  Next the DEXT.

Dave

[Dave McCampbell]

Hi Kampfsc.  I have to disagree with you on the cell quality of our EVE cells.  I don't know where you get your cells from, but I know what constitutes a Grade B cell.  Maybe your definition is different.  These are advertised worldwide by a very reputable dealer as Grade A and factory matched.  That includes a QR code and listing of each cell with ALL their individual specifications provided with them.  I've never seen that before.  In any case they are the same as Dacian's and based on that, plenty well matched for my purposes.

Our new configuration will be 3 cells in parallel sets and 4 of those in series not 2 or 4 in parallel.

Thanks for your other comments on building and managing the bank once it is installed.  I will consider them very carefully when the time comes in a few months to install the bank.  I do agree with you that I probably need a larger and active balancer.

Dave

[michael clark]

michael clark

6:49 PM (1 minute ago)

to Dave, electrodacus

good evening to all, 

just so you know. i have a 4P8S EVE 280 Ah cell build and it is passively balanced by the SBMS0. I also have a 2P8S Lischen 272Ah build passive balanced by the SBMS0 ( I built two of those; I also built a 8P8S Eve version 3 cell build with 280 Ah cells without a BMS but added a Neey 15amp active balancer which is only periodically used....

each of the builds the cells in parallel seem to help regulate themselves some.and the SBMS0 does balance quite well. often there is only 5mV differentials sometime 2mV differentials.

why some work better than other???? all speculation and application.

if I encounter any problems it was always due to a loose connection.

mine are all stationary builds and not on a boat or moving RV...

8p is 280 x8 = 2240 AMps no balancer most of the time....although I periodically check the NEEy to see the pack voltage and each of the parallels.( 280 Ah EVE cells 64 of them)

4P is 280 x4 = 1120 Amp >>> passively balanced on charging by the SBMS0 (280 Ah EVE cells 32 of them)

2P Lischens is  272 x 2 = 544 amps  >>>> passively balanced by the SBMS0 (2 sets of 16 of each of the Lischen 272AH cells)

all are 24 volt build so 8S configuration after being paralleled 1st...

good travels to all.

the Lischens have been in service since 2019, the EVEs for about 1.5 to 3 years at present...with the 8P8S EVE version 3 cells in service for about 1.5 years

I read and follow many people but the final decision is always up to the builder to decide the best course of action.

monitoring is crucial at times and other times is not necessary, sometimes I do not look for more than a month, other times daily!!!

Michael

[Dacian Todea (electrodacus)]

Dave,

Even with no cell balancing at all (disabling the SBMS0 cell balancing) the cells will not be damaged so whatever it was that got your old cells damaged has nothing to do with cell balancing.

If there is no cell balancing the usable capacity may decrease over time due to increase imbalance but it will not damage the cells.

Quite some time ago I did a test on my old 100Ah GBS battery that had fairly unequal cells by disabling the cell balancing for 4 months and then I did a manual cell balancing by charging the pack then each cell manually charged to 100% and the total imbalance was around 2Ah maybe a bit more than 2% imbalance after 4 months of normal use with no cell balancing at all.

Likely your battery is way better but even at this extreme 0.5% cell imbalance per months that will be about 6% imbalance so 6% capacity loss after a year (not permanent loss just the loss due to cell imbalance that can be recovered after balancing).

Your battery capacity is much larger so 0.5% for a 900Ah battery means around 4.5Ah imbalance in a month so at 200mA of cell balancing that will require 4.5Ah/0.2 = 22.5h of cell balancing so about one hour per day for SBMS0 to keep that sort of battery in balance. Likely much less required for a new EVE battery pack in normal operation.

A pack that has higher delta between cells than my old GBS is either a pack containing used 18650 cells or a pack that was not properly installed meaning there is high resistance on some of the connections points.

The way a pack will get damaged or degrade much faster will be one with bad connections that heats up including heating up the cell and that will significantly increase degradation or improper installation that allows the pack to be overcharged or over discharged where damage can be immediate.

Cell balancing is one of the least important features and will not help with battery longevity almost at all. Maybe get an extra 0.5% better capacity after 15 to 20 years of usable life when using cell balancing vs not using cell balancing.

The main role of cell balancing is to keep the cells top balanced so that you do not overtime lose usable capacity and require you to manually re balance cells maybe once a year to recover the 5 or 6% capacity unavailability.

Even with cell balancing there may be a 0.5% imbalance so 0.5% unavailable capacity but that will never increase over time. I can if I get some time disable the cell balancing on my EVE pack and maybe check the imbalance after a few months to see how it compares with the old GBS battery but I'm fairly certain it will be way better much less than 0.5% per month so I will need to run the experiment for at least 6 months to a year to get an accurate average.

[jedics]

How well has this compression method worked? How do you measure how well it did or didnt work?

Deciding on what to do with my LF638's.

[sailingharry]

It's hard to assess the effectiveness of an individual installation, as there is nothing to compare to.  However, Eve is pretty clear on their recommendation to compress, so I took their guidance.

To do what is in that picture is pretty easy.  EVE gives a "pounds of force per cell" standard.  Look at how many wide you are compressing (how many deep -- one after another -- doesn't matter).  Say the standard is 200 lbs per cell (I forget).  Say you have 2 wide (my install), so you need 400lbs of compression.  My install has 6 rods (top and bottom sets, one left, one middle, one right).  So the pressure per rod is 400/6 = 66lbs.  I took the springs I bought, set them on top of a bathroom scale on my drill press, and pressed down with the drill press to 66lbs, and then measured the compressed length of the springs.  On assembly on the boat (with 8 cells, I wasn't carrying an assembled pack!) I tightened the nuts until the springs were the right length.

For my installation, I firmly attached one compression panel to a bulkhead, installed the batteries, compressed the springs, and then installed some very stiff downward compression across the tops.  They haven't moved a micro-inch since!

[jedics]

IF they havent moved at all then why do I need springs, istn the same thing achieved if I just tighten the nuts?