Freedom won eTowers cell balancing

Batteries
1

I started discussing this issue in another section but decided to bring it here.

I installed 2 x Freedom Won eTowers in December 2021, connected to 2 x 48/3000 Multiplus’s and 4000Wp solar. I had absolutely no issues with the system untill I started getting “high voltage” errors on my GX device November 2023. I logged a ticket with FW after which they remotely logged in and diagnosed both batteries, tracing the problem to cell #13 on the 2nd battery.

After some fiddling with the parameters byt the technician, the errors were gone for just over a month. Suddenly the "high voltage"error re-appeared accompanied by our old friend, cell #13. This time I contacted my installer/supplier who, in turn, contacted his supplier. I gave them access to my VRM site for them to check settings. I got quite agitated after I logged in and saw that my dashboard, widgets and a lot of settings were changed. Everything was set up and installed exactly like the installation manual requires and seeing that only one battery showed the errors pointed in a faulty cell direction. The “good” battery showed a cell difference of no more than 30mV while the other one started exceeding 600mV.

Today FW remotely checked everything again and again fiddled with settings and asked me to keep the batteries charged for up to three days to see if the voltage difference decreases.

The technician did instruct me to swap the batteries so the new one can be adressed as nr.1 and not vica versa like they were. After I asked him what he meant by “old” battery, he pointed out that the one battery (the faulty one) only showed 360 battery cycles on it’s BMS and the other one showed 731 cycles. Seeing that both were acquired and commisioned at the same time, this would be impossible, wouldn’t it? I never cycle the batteries below 40% SOC and only twice they ran below 20% during blackouts, before the inverters switched off at 13%. I would maybe accept the 731 cycles after 2 years but how the hell could the other one, connected in parrallel and being used for an equal amount of time on the same system, have only half that cycles?

I also found that the faulty one showed around 5% less SOC when it reaches levels below 50%. Clearly, according to the PBMS software, cell #13 on the 2nd battery is the culprit, now also giving “low voltage” errors when it nears 30% SOC.

Could it be that the one battery was faulty from day 1? I was always bothered about the usable capacity, seeing that I barely used 5Kwh before the batteries dropped below 40% SOC. I my calculations were correct I should have been able to use closer to 6Kwh before it got to 40%, 8Kwh being usable out of the 10Kwh (If drained to 20% SOC).

Anyhow… I just wanted to share my eTower experience so far. I haven’t got any cold shoulders from FW yet, but let’s see what happens if the cells don’t balance after three more days of charging. I’m down in the Karoo and guess I’ll have to courier the battery to Centurion so let’s see what they bring to the table.

I changed from lead acid to Lithium batteries because I wanted a problem free storage solution and decided to pay a premium for a locally manufactured product. I hope my decision wasn’t flawed…

2

I’ve got a problem with a unballanced cell on one of my own eTowers at the moment so let’s see if FW keeps up to their warranty promises. I once also started with Chinese thingamajigs (Axpert) but soon realized limitations and decided to ditch it before it failed. There was only one option for me that would compliment my blue eyes. When I replaced my Lead acid batteries with Lithium I was a bit thrown off my game though. After much homework and budgeting I got the eTowers but now after just over 2 years I’ve got a problem with one of them. I sommer wanna kots!!

3

Let us know how it goes. Plenty eTower users here. I only had an issue when adding another. This was solved by dropping the batteries as low as they could go. Afterwards, it balanced perfectly.

4

I never had any issues with them up to now. I think I must also discharge them untill they switch off and then properly charge them again for a day or two and see if that helps. Only one cell on one of the batteries that keeps giving the errors

5

Was one of my initial concerns. So I went the other way, DIY battery bank.

It is fun but not everyone’s cup of tea.

The reason, I sat for months and wondered … so, if I spend XYZ on a bank, 10-year/x cycle warranties, the bank not even in the market that long, a fight may be brewing if there are issues, it is big bucks we are talking about.

Or worse … “what if” the seller/importer decides to close down their shop/business?

And I like the colour blue, I can mos do what I want, no?

So I decided to take the “warranty” on my shoulders, when I paid 1/3 of the XYZ banks costs.

Big caveat though … there is a sweet spot. Smaller banks, forget it, not worth it. Similar price with no warranty whatsoever, silly move that, the “savings” are not worth the “risk” of DIYing it.

6

If I recall set Min SoC to 10%, let them sit there for a few hours then set to Keep Batteries Charged. Leave there 100% for a few hours. Sorted. Almost a calibration of sorts.

7

Okay. So FW had an online session this morning, changing some parameters on both eTowers. Something strange I noticed, both batteries were installed and used together but the faulty battery shows 360 battery cycles while the good one shows 731 cycles. How even?
Oh yeah and so far discharging and then charging again helped nothing. Will try it again but limit the charge current to say, 10A so it can charge less aggressively and give the BMS more time to balance seeing that it only balance while charging and not while discharging

Screenshot 2024-02-05 123201
Screenshot 2024-02-05 1232011104×689 31.6 KB

Screenshot 2024-02-05 123259
Screenshot 2024-02-05 1232591112×693 30.6 KB

8

Taking this to the batteries section…

9

I have combined the 2 threads :wink:

1 Like
10

So this is where I’m at. After charging the batteries at a low current of 10A the whole day, the “faulty” cell jumped from 3.45V to 3.86V in about 5 minutes, triggering the high voltage alarm (Pack jumped from 54.55V to 56.03V at the same time, SOC from 97% to 100%). Again, Cell #13 is the culprit with a voltage reaching close to 4V at times. I had to set the DVCC voltage down to 55.6V to keep the error at hand, at 55.8V, as specified, unlucky #13’s voltage jumps. as soon as the battery gives the high voltage error, the whole battery is blocked by the BMS. I think that might be where the difference in cycles originates from but if that’s is the case, the problem must’ve shown up quite some time ago, just not enough to trigger the error and warn me.
What do you slim okes think??

Screenshot 2024-02-06 162428
Screenshot 2024-02-06 1624281310×440 41.1 KB

Screenshot 2024-02-06 162454
Screenshot 2024-02-06 162454761×299 23.8 KB

11

Must go back … you can try and “fix it” adjust the DVCC volts and charge amps, but at that voltage, the cells is suspect.

My personal view is based on all my trials and errors over multiple banks.

Ps. I know of a guy who wrote a piece of code, that when a cell/cells shoots out above a preset parameter, the code reduces the charge amps and volts … but … always a but … it is only for Jaibiada BMS’es. Should be standard in all BMS’es. Maybe one day.

12

Warranty replacement! The supplier and FW have been involved enough and I think now is the time to make the big push! The SOC diff is a big worry for me as well!

1 Like
13

That is exactly how it goes. Looking at the reduced capacity, I think that cell also cuts off at the low end. If you can, download the logs with pbmstools (“memory info” tab).

1 Like
14

I did download the logs but it will only download up to 400 entries and with all the errors, 400 entries is just over 4 days worth of errors. The “good battery” only have a total of 56 entries since it was installed but non of them ever related to high cell voltages and the unit has never been blocked. The faulty one though shows a lot of “fully blocked” entries. That’s why I get the idea that the battery might have had an issue for quite a long time, explaining only 360 cycles against it’s partner’s 731 cycles. It was just not noticable enough to show up on VRM so it was never diagnosed or monitored…

Attached is both batteriy’s logs, as far back as possible. I had to convert to PDF to get it uploaded.

I would say this is where FW should stop fiddling around and doing online diagnostics and start to come to the table with a viable solution. If I wanted to have constant problems with batteries I would’ve bought some cheap Chinese stuff for half the cost. Clearly the one battery was a “dud” from day one, explaining many other issues like low capacity, overcurent shutdowns at 90A, ( while according to the spec sheet 2 of these should handle a 180A max / 162A cont. current draw.), etc.

I think it’s time to increase the tone of my voice now…

15040MemoryInfo202402041218#2.pdf (1.0 MB)
15040MemoryInfo202402041229#1.pdf (174.6 KB)

15

Just an quick update. After a comment on FW’s FB page they called me and requested me to bring the faulty battery in. Since I live in Beaufort West and they are in Centurion, this task would however not be as easy as “dropping it off”. Courier guy quoted me R3024…one way. So I told the guy that seeing that the battery is under warranty, they would haave to cover the transport cost in that case…which they were not willing to do. The next option was to get it to Herholdt’s in George, where it was originally bought, still a 240km Logistical problem for me, but seeing that it was the George old motor show this weekend we decided to take the trip and deliver the battery to Herholdt’s George. I made a proper crate/box of 18mm plywood, tightly tucked in wiith an eTower pedestal underneath and another on top, insuring that the battery itself does not come into contact with the wood itself causing chafe marks. I emailed FW the receipt of acceptance from Herholdt’s this morning and they acknowledged that they’ve received it.
I have taken on a wild goose chase and asked FW and Herholdt’s (the original supplier) for a loan battery while that one is in for repairs but have received no answer on that request yet. FW did however offer to trade in the eTowers on a bigger capacity battery (Lite range) but even if I wanted to, I just do not have the financial capacity to dash out R150k for a battery. This was why decided on stackable rackmount batteries in the first place, so I can add more capacity as it becomes financially possible.
But I’m staying positive and seeing the silver lining being that they did however come to the party and agreed to repair the battery and I’m thankfull for that.

One thing that still bothers be though…how is it possible that batteries commisioned at the same time have such a big cycle difference? 362 vs 731? That;s almost half!! And even with a battery that clearly has a faulty cell, the SOH is still 100%
For the first time I’ve had the eTowers and Smartshunt, both give me a SOC% value that differs with less than 2%,(with the good battery connected on it’s own) telling me the faulty battery have been screwing around for a very long time…

16

Unfortunately this is what my colleagues and I have predicted when we first saw them using the Tianpower BMS, as we haven’t found a battery with that BMS (seems to be one of the most popular Chinese BMSs).
The early ones could not measure less than 2A, newer ones seem to be able to measure 1A, they will only balance (10mA IIRC) whilst they measure charging current, they will also never adapt their charging currents.
We have noted several times that one can often completely discharge a battery when leaving it on a system in standby, as the BMS cannot measure that low current, so the SoC is still 100%, but the battery might be ~48V.

17

The eTower uses the Pace BMS, though.

1 Like
18

As @_a_a_a said, different BMS. But I have to add that most BMSes (at least the ones that aren’t complete rubbish) should not do what you describe here. The measurement threshold is not the worst problem in the world, even some BYD models have that same issue. What such models should do to compensate for the poor current measurement is implement voltage “waypoints”. If your lowest cell is only 3V… no way is that battery much above 35% SOC. When your lowest cell hits 2.8V, the SOC is 0%. By simply applying such corrections, a BMS can often get away with a courser current reading. The real killer in this BMS you describe, would be that it doesn’t do that.

19

The Pace BMS in the eTower only balances while there is an inflow of current so while discharging there is no balancing and cells can easily go out of balance. If your charging current is a bit aggressive (even 70Amps fully charging the batteries in just over 2 hours) it seems that the BMS cannot balance the cells quick enoug before it reaches 100% SOC. I would say adding a proper active balancer or replacing the Pace BMS completely would solve many of their problems.

Another issue I’ve come accross with the loan battery (got it with 0 cycles while mine has got 750 cycles) is that the SOC goes almost 10% lower than my old battery eg: this morning I saw that the red alarm led was flashing on the loan battery with one green led on the SOC bar while mine still had 3 green led’s on the SOC bar. I guess this can be a firmware issue. SOC on VRM showed 34% but when I plugged in my laptop and directly checked the batteries the loan battery showed 16% and my old one 32%. VRM at that time also showed 32%. Communication between the batteries and the GX device is good, the loan battery’s address set as #1 and connection from the system DC busbars are busbar to battery #1 positve and busbar to battery #2 negative. Batteries are connected to each other with the bars that came with it. Very strange. The BMS’s should equalise the batteries so I think this might be a firmware/programming issue. The loan battery manufacture date is 11/2023 while my old one is 05/2021. I don’t want to make too much of an issue about it seeing that it’s only temporary untill I get my battery back from FW

20

Not a good idea at all to do that.
I tried that… in specific cases, it can facilitate top-balancing cells if you have full control over the BMS/systems. Daily, not a chance.

The system is supposed to throttle the amps the higher the SOC gets.
BUT, even if the system throttles the charge amps, batts where the cells get out of balance, you have to intervene at times to help the BMS balance because of Milliamps balancing vs Amps charging.

To get a system to charge at i.e. 1amp for example is touch and go. 5amps with milliamps balancing, yeah right ok.

Now before anyone jumps in here … this does NOT apply to balanced cells. What is works. We all know that.

It is when things go sideways that this issue comes into play.