Strange behaviour at 3.4V always a different cell pack shoots up to 3.55V, rest 3.4V

[CaptainRivet]

Hi all,
I have a very strange behavior never happened to me before.
When my 4P4S bank hits the knee at 3.4V always a different pack shoots up to 3.55V while rest stays at 3,38 till 3,42V. So arround 130-150mV devation. But nearly every cycle its a different cell pack.
All is tight down with 6NM, lishen 272AH cells, nolox paste, measured resistance is equal. If i measure the pack each cell is equal so its not a faulty cell.
Till now its electrodacus passive balancing with max 150mA and that cannot cause the differences as 150mA is not able to shift 4-8AH in one cycle.

What can cause this?
I got an active 5A balancer, lets see if that sorts it but i wanna now what causes this behaviour.

[wholybee]

That it is sometimes a different cell/pack is weird, but could be some consequence of being 4p, not sure. If you have several cells in several 4p blocks that are out of balance, I suppose that could happen. Remember, that at lower then 3.4V, 4 parallel cells can be the same voltage, but very different SOC. So, if one of those 4 cells gets out of whack...

Otherwise, the symptoms are an unbalanced pack. You probably need to break it all apart, parallel everything, and top balance to ~3.6V-3.65V.

As for how this happens, this is often the result of not charging high enough over a long period of time. The cells need to be greater than about 3.45Vpc to balance. An active balancer that "balances" at low voltages can actually pull them out of balance. If you charge to into the curve once in a while, and let it stay there for a bit to balance, then this shouldn't happen.

Anyway. Charge them up until one cell starts to run, then take it apart, parallel them all, and put a powersupply set to 3.6V on it. Set the voltage to 3.6V before connecting it, then let it sit until the cell voltages are all equal at 3.6V. Then put it back together. You could also individually test cells when it is apart also. Maybe a cell or two has tanked and doesn't have the capacity of the others.

Where did you buy these cells? Lishen are decent cells, but often bought from Alixxx, the quality can vary quite a lot.

[Simi 60]

Have you turned it off and on again?

I had something weird happen like that a while ago
Unplugged BMS and plugged it back in again
Problem sorted

[CaptainRivet]

Originally Posted by Simi 60

Have you turned it off and on again?

I had something weird happen like that a while ago
Unplugged BMS and plugged it back in again
Problem sorted

Good idea, will try.

[CaptainRivet]

Originally Posted by wholybee

That it is sometimes a different cell/pack is weird, but could be some consequence of being 4p, not sure. If you have several cells in several 4p blocks that are out of balance, I suppose that could happen. Remember, that at lower then 3.4V, 4 parallel cells can be the same voltage, but very different SOC. So, if one of those 4 cells gets out of whack...

Otherwise, the symptoms are an unbalanced pack. You probably need to break it all apart, parallel everything, and top balance to ~3.6V-3.65V.

As for how this happens, this is often the result of not charging high enough over a long period of time. The cells need to be greater than about 3.45Vpc to balance. An active balancer that "balances" at low voltages can actually pull them out of balance. If you charge to into the curve once in a while, and let it stay there for a bit to balance, then this shouldn't happen.

Anyway. Charge them up until one cell starts to run, then take it apart, parallel them all, and put a powersupply set to 3.6V on it. Set the voltage to 3.6V before connecting it, then let it sit until the cell voltages are all equal at 3.6V. Then put it back together. You could also individually test cells when it is apart also. Maybe a cell or two has tanked and doesn't have the capacity of the others.

Where did you buy these cells? Lishen are decent cells, but often bought from Alixxx, the quality can vary quite a lot.

Cells are from my buddy who runs an NPO building powerwalls and commercial EVs. Grade A+ and matched all 24 cells. He gets all 2 weeks a container direct from EVE and Lishen.
Then split up in 12 for my cat and 12 for a buddy boat. But he could fit only 8 so i took the remaining 4 and got a 4p4s bank (will split it soon in a 2p8s and go 24V). But i had my 12 top balanced, then run them for a while charge till top and then added the 4 additional top ballanced and charge top too before i put them together, each 1 add to a 3p pack.
Maybe you are right that the 1 additional cell each pack has a different top soc. Thought in parallel if there are little difference they balance each other as they are parallel but that could be the root cause.
You are right i don't like 4p much, think limit is 3p but it wasn't planed like that. Well i soon change to 2p8S in 24V, the 2x Studer XTM 24V/3500 lying at my parents and will replace a multiplus and a edecoa boost inverter in 12V. Still hestitating to go 24V as the 12V multi plus boost inverter setup works extremely well beside that little strange cell behaviour and the 2nd boost inverter shuts on maybe once a week.

I install the 5A active balancer and steer via electrodacus to run it above 3,4V and see if he gets it sorted. At the marina so i can force it with little current from shore charger to stay long in the 3,4V till 3,65V area. Normally charge till 3,53V end of charge but you are right, mostly operates in the 60 till 85% SOC area. In sommer no prob get it to 100% SoC but in side season i hardly do.

If that or simi's pull the plug approach doesn't sort it i will pull apart and parallel all and put my 3,65V powersupply with 5A on but that takes then a while to get them balanced.

[gpeacock]

Did you sort out the issue with the cell imballance? Im just about purchase an Electrodacus to replace my melted Chargery BMS. The Chargery worked fine but I found it pulled my cells a bit out of balance because I couldn't set the min balance voltage high enough... Want to make sure I don't end up with another balance problem with the new BMS..
Thanks

[CaptainRivet]

Originally Posted by gpeacock

Did you sort out the issue with the cell imballance? Im just about purchase an Electrodacus to replace my melted Chargery BMS. The Chargery worked fine but I found it pulled my cells a bit out of balance because I couldn't set the min balance voltage high enough... Want to make sure I don't end up with another balance problem with the new BMS..
Thanks

Yes i resolved it, had partly nothing to do with the ElectrodacusBMS.
There were 3 issues: as general guideline in 4P or more paralell cells everything must be identical, smallest deviations add up. To see whats really going on you must heavly load the installation (i ran 500A) and use a temp gun and simply measure temperature on every connection, busbars. Like this i discovered issue 1 and 2. Cell 1 had 33 degrees, 2 nd had 31 3rd 28 and 4 th 25 degrees l, after 1 and 2 was done i have all at 25 when pulling 500A.
1) the cooper busbars were not even enough that 4 cells in parallel connected identical to it. Solved by using 10mmx60mm aluminium bars and put them on an engine lapping machine which made them absolutely perfectly even and took away any pontential surface corrosion/oxidation and right away coated with no-al-ox where the cells connect
2) pulling 350A-450A regularly i simply didn't do symmetrical offloading means everything in the 4P4S battery must be symmetrical. I use the U layout and simply connected the busbar straight to the main fuse...wrong 1 now have a 90sqmm short cable connecting in the middle between cell 1+2 plus another 90sqmm cable connecting cell 3+4 to the main battery shunt with cable lugs stacked onto each othee. On negative side the same. You can see that in below picture. If i disconnect each side one cable bank gets out of balance, if i connect it stays within 3-8mV deviation, cannot be more prefect for a 4P4S battery with Lishen 272Ah cells.
3) the ElectrodacusBMS passive balancer has only max 100mA, thats ok for max 550AH in my experience but its balancer is simply to small for a 4P4S 1088AH bank. An easy fix get the 5A heltec active balancer and its steered by EXTIO5 with SOC >96% put on 4. So with a hysteesis of 3% it switches on at 99% and off at 96% SOC. I need that because i will add another 12 304AH eve and i am already now hardly in the balance area...mostly 60 to 85% SOC. Yes if you daily hit 100%SOC you might get away with the internal one if you have a perfect top balance. Well a 5A Heltec active balancer is 30$ and that works fine.

[CaptainRivet]

Originally Posted by gpeacock

Did you sort out the issue with the cell imballance? Im just about purchase an Electrodacus to replace my melted Chargery BMS. The Chargery worked fine but I found it pulled my cells a bit out of balance because I couldn't set the min balance voltage high enough... Want to make sure I don't end up with another balance problem with the new BMS..
Thanks

How do you melt a chargery BMS?
You realize that ElectrodacusBMS has a different design philosophy though you have to make some changes to your design. What cells do you have in which config? 12 or 24V?
ElectrodacusBMS has shunts in the positive.
You can use the PV shunt as main charge bus and connect all charge sources and the main Shunt as load bus where you connect all loads. Like this you have a very clean install and eg ypu can simply switch all charge sources with one or 2 Battery protect 220A.smart. this has the advantage you can also use the remote on/off via BT connect and also connect a BMV712 relay output in series to the BP220 remote that switches it off as last resort in case ElectrodacusBMS fails.
I would highly recommend to get victron shunts, they stay cooler then the other high quality and the 500A shunt has the same height then a NH2 fuse holder which creates an easy install by just connecting with a small busbar the shunt directly to the fuse holder. Or the NH3 fuse holder the connector fits directly on top of the 500A victron shunt.

[gpeacock]

Nice Set up! I wish I had your attention to detail.
It wasn't the Chargerys fault for melting.. I managed to put the whole setup underwater for about 4 months. The Chargery literally dissolved into a pile of goo...
Surprisingly the batteries cleaned up fine and after a couple weeks of testing the pack seems to be staying in balance without needing a balancer. Its set up as a 4p4s pack (12V@400AH). I'll be keeping a close eye on it but the spa months didn't seem to harm the pack at all. I like your setup with the aluminum connection bars. Mine will be replaced at the end of the season (they didn't fair as well under water) and I will probably switch to what you did. Having said that Im surprised it was the bus bars causing your pack to go out of balance. I use my pack for a thruster (500A), windlass (200A) Inverter (200A) and charge via 2 alternators at 100A each. The single + and - connections to the system has worked fine for me with no issues.
Yeah, understood about the shunt needing to be on the positive with the electrodacus. Not sure why he bucked the trend there... I was actually thinking about not bothering with the shunt and just continuing to use the BMV712 for the pack meter. I really only need the voltage cut offs on the bms.
The chargery worked fine but I like the single box style of the electrodacus and I dont need the extra balance power that the chargery has. In fact I found that because the highest balance voltage setting was 3.4 it actually pulled my pack out whack more often than not when turned on. The extra switches that the electrodacus has might be nice in addition to its potential to play nice with victron products.
Thanks

[CaptainRivet]

Originally Posted by gpeacock

Nice Set up! I wish I had your attention to detail.
It wasn't the Chargerys fault for melting.. I managed to put the whole setup underwater for about 4 months. The Chargery literally dissolved into a pile of goo...
Surprisingly the batteries cleaned up fine and after a couple weeks of testing the pack seems to be staying in balance without needing a balancer. Its set up as a 4p4s pack (12V@400AH). I'll be keeping a close eye on it but the spa months didn't seem to harm the pack at all. I like your setup with the aluminum connection bars. Mine will be replaced at the end of the season (they didn't fair as well under water) and I will probably switch to what you did. Having said that Im surprised it was the bus bars causing your pack to go out of balance. I use my pack for a thruster (500A), windlass (200A) Inverter (200A) and charge via 2 alternators at 100A each. The single + and - connections to the system has worked fine for me with no issues.
Yeah, understood about the shunt needing to be on the positive with the electrodacus. Not sure why he bucked the trend there... I was actually thinking about not bothering with the shunt and just continuing to use the BMV712 for the pack meter. I really only need the voltage cut offs on the bms.
The chargery worked fine but I like the single box style of the electrodacus and I dont need the extra balance power that the chargery has. In fact I found that because the highest balance voltage setting was 3.4 it actually pulled my pack out whack more often than not when turned on. The extra switches that the electrodacus has might be nice in addition to its potential to play nice with victron products.
Thanks

upsi...really suprise the cells survived the pool treatment.
a lot of load for a 4P4S 400AH bank, if you didn't have an issue keep it...never change a well running system unless you have an issue.

Electrodacus:
you must have at least the main shunt on positive so electrodacus works, you can spare the 2nd PV shunt. without the shunt you fry the current measuring amp which you also need for the calibration process, just simply add another victron shunt into positive. actually thats good to have shunts on both terminals because the shunts are very stable and actually also take the mechanical load of the main battery cable. its also good to have both electrodacus and BMV712 delivering the bat data, it really helps getting it properly calibrated and you really see an issue very early as the measure approach is differently.
design:
well was the 3rd iteration and had 3 month time anyhow till my hulls dried out and i could restore the underwatership. so did it right with the experience on my vessel with my loads and be finally done with it. its under my owners bed so to work there for longer i need to sleep in the guest cabin. all i don't need to touch when system is running is below bed but inverter or electrodacus is in a cupboard with door for easy access behind bed.
regarding busbars i am not suprised thinking about it and knowing how soft but heavy cooper is and they are delivered originally in 6m length in 5mmx30mm. so if they lift them eg in the middle its already bent and not straight anymore, then pull if from the stoarge rack and it gets bend in the other direction. not a problem for 12 or 14cm long busbars but if you look at mine its actually two >60cm long busbar connecting 8 cells. the big alu bars in 10x60 are delivered in 4m length and even just picked up in the middle that won't bent at all as alu is also much lighter.

the big alu busbars where actually the only choice i had here, tined cooper in needed dimension is not avalible here in canaries but we have an absolute genius inox and alu store here which has nearly everything you do want in inox or alu. so i researched and asked which specs of alu profil they have in stock or could order and choose the best from electrical side and biggest they had and fitted...the beefy 10mmx60mm busbars where just 40Euros alltogether
but with my experience now I will use these ones in the future for other builds, especially if using long busbars. you never stop learning.



suggest to use my U-setup of cells and not 16 in a row as
a) 16 is just too much mechanical force on terminals due to expension in different temps...massive busbar work well with 8 in row max but loads of people had problem with 16, then you need the flexible busbars.
b) use also 2 threaded rods in the middle so you can clamp them evenly in 2 rows, 16 the force on the outside to middle cells are also quite different so the outside get to much while inside to less. here 4 is best, 8 work well too that force on each cell body is quite similar.