Unexpected Battery Imbalance in 48V (4s 12V) LiFePO4 System After Deep Discharge. Advice Needed

DIY
, , ,
1

Good afternoon everyone,

Just looking for some insight from the community. We recently upgraded our off-grid setup to a 48V system using four 100Ah WattCycle LiFePO4 batteries wired in series. I also added a LiTime 48V active balancer. All the standard steps were followed during commissioning : the batteries were fully charged, then paralleled for 24 hours to equalize, and then connected in series.

Everything worked great while I was on site for a couple weeks. The balancer kicked in during each charge cycle, and I’d occasionally see up to 3–4% SoC difference between batteries. But within 30 minutes after the charge controller stopped, all four batteries would level out in both SoC and voltage. This was confirmed through the BMS Bluetooth apps.

However, this morning I got a call from a family member currently at the off-grid site. They had a blackout last night after six days of no sun. The main battery breaker tripped, and lights went out suddenly. I suspected the batteries had fully discharged.

After some troubleshooting, it turns out three batteries were sitting at around 0% (about 11.1V), while the fourth was showing 60% SoC and 13.2V, which seems very odd, especially with an active balancer in place. I advised not to flip the battery breaker back on and to give the balancer time to work.

By this afternoon, the situation improved slightly. The three low batteries were at ~2% SoC (12.3V), and the high one remained at 60%. They’re heading into town tomorrow to grab a lithium charger and will individually top up each battery to 100% before reconnecting them in series.

A few questions for you guys :

Any idea how this imbalance happened despite the balancer?

Is it safe to reuse these batteries in series once they’re all fully charged again?

Is there a quicker or safer way to recover from this situation in the future?

Appreciate any thoughts.

For reference, these are the system specs, should more information be required, please feel free to ask and I will fill you in.

1× Renogy Rover 60A (soon to be replaced with a Victron SmartSolar MPPT 150/45 in July)
4× 105Ah Wattcycle BT Mini batteries
1× 48V LiTime balancer
1× Victron Cerbo GX MK2 (to be added in July)
1× Victron SmartShunt (to be added in July)
1× Victron 48/800 120V Phoenix Inverter
1x Meanwell 500 W 18-70 VDC / 12 VDC converter

  • All breakers are two pole breakers for both positive and negative.

image
image1369×862 158 KB

2

Are you sure the maximum balancing power is equal or larger than the charging current? I did something similar with 2 li-time 24V 200Ah batteries in series. For my specific case it turned out there was no suited balancer on the market. So I designed my own. I am an electronics engineer and designed the PCB, wrote my own firmware and so on. It took me a couple of hours to get the thing working. After following up the behaviour I found out i needed to reconfigure the charging parameters of my MPPT to stay within the limits of the power the balancer could handle. I advise you to dive into the manuals and get the details. I used my victron smartshunt to guard the mid point voltage by utilizing the aux input. In your case (having 3 mid points) you can not do that without multiplexing. I suggest you get a qualified tech on-site. Pushing BMS’s into overvoltage protection day after day can be problematic over time.

3

Thanks for the reply Gerd, much appreciated.

I’m actually a power systems engineer based in Canada, so I’m no stranger to LV, MV and HV systems. The solar farm in question is a 6-panel setup: 6 × 250 W (totaling 1500 W). Under peak sunlight around midday, I was seeing charge currents around 25 A. The active balancer I’m using is rated for 10 A.

While I was on-site for a couple of weeks, I monitored the system closely. During daily charge cycles, the batteries never drifted more than 5% apart in SoC, and the balancer was able to bring everything back in line within about 30 minutes after charging ended, it seemed to work well.

That said, I suspect the root issue is that the batteries were fully discharged at some point. What I can’t figure out is how one battery ended up jumping to 60% SoC while the others remained low, maybe the balancer failed? To complicate things, the two-pole breaker protecting the bank failed indicating one pole open and ther other pole closed. I’ve never seen a S.E. Multi 9 breaker fail like that.

I’m heading back up this weekend (12-hour drive) to get eyes on everything and sort it out. In the meantime, they’re recharging each 12 V battery individually to 100%, and will parallel them tomorrow in preparation for rewiring.

I’ve also ordered a Victron SmartShunt. Once installed, it’ll give me much better visibility into the system’s state, and I plan to use it to implement a basic load-shedding scheme to prevent deep discharge events like this in the future. Wondering if I dont order 3 Victron Balancers, not certain it would be much better than what I already have.

Thanks again for your insight.

4

Good to hear the setup is in good hands. It is weird however that the balancer is reporting 4 different SOC’s while it’s main objective should be keeping the 4 voltages equalized. The SOC is an incremental time integration of currents that should all be the same in a series connection… The thing with LifePO4 is the steep voltage climb near full charge. In this climb the charger should jump from bulk to the next step depending on configuration. The thing is that one out of 4 packs will not cause enough voltage climb to let the charger react. I expect it to stay in bulk at 25 amp and the balancer is only able to burn/transfer 10A. For safety reasons I would for the moment mimit the charger to 10A if possible. That one BMS semiconductor in your lowest capacity pack is not the only component preventing fire. If that fails on you… A good safety system should have 2 components at least in redundancy and guarded. If you can’t figure it out feel free to shoot questions.

5

Update…

I’ve just contacted LiTime about the balancer. It would appear the unit is defective and is the cause of this very very odd unbalance. Reading online I found different reports of people seeing the same sort of results.

At this point, I may end up purchasing 3 of the Victron balancers or just not installing any balancers at all and periodically balancing them myself.