Cell imbalance during (slow) capacity test: not even near rated capacity

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
Hi all,

So, I'm currently testing my 8 280Ah cells.

I charged the individual cells to a certain point, which in hindsight shouldn't have done cause it takes way too long with a 10A power supply. So when I got the charger, I did a quick charge and then proceeded to top balance as per instructional PDF from this forum, with a maximum voltage of 3.65V. I stopped the top balance when the current flowing was below 0.9A.

Then I started the (slow) capacity test, with a load of about 175W.
1627929773591.png

This went well for a long time, for somewhere around the 5000Wh mark, the cell voltage started to drift apart.
1627929970695.png

1627930036329.png

At about 6000Wh/233Ah and a voltage difference of 0.46V and a sum voltage of 23.2V (according to the Daly Bluetooth App) I decided to stop the test (for now) after running for 35 hours (with a pause during the night/when I was away).

Between 5000Wh and 6000Wh, the voltage dropped quite fast.

The Daly App in which I entered 280Ah as the rated capacity now shows 13% SOC, 87% would mean about 243Ah gone.

Did I stop the top balance too early, before hitting top capacity? What is an acceptable voltage difference between cells? Notably, "Cell 1" holds 3.179V, while the lowest cell has 2.728V.

Any pointers?


Thanks,

Martijn
 

FilterGuy

What, me worry?
Joined
Nov 26, 2019
Messages
4,435
Location
Los Gatos CA
Did I stop the top balance too early, before hitting top capacity? What is an acceptable voltage difference between cells? Notably, "Cell 1" holds 3.179V, while the lowest cell has 2.728V.
I stopped the top balance when the current flowing was below 0.9A.
.9A (900 mA) is probably too much for stopping the top ballance. I have found by experience that the balance works better if you let it go to zero or near zero (<100 mA for 4 cells...or even lower). You have an 8cell system so I would take it down to 200mA or even less) The extra time from .9A to zero is negligible for top balance and it won't hurt the cells even if you let it go to zero for a day or two.
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
.9A (900 mA) is probably too much for stopping the top ballance. I have found by experience that the balance works better if you let it go to zero or near zero (<100 mA for 4 cells...or even lower). You have an 8cell system so I would take it down to 200mA or even less) The extra time from .9A to zero is negligible for top balance and it won't hurt the cells even if you let it go to zero for a day or two.
I did another charge and top balance, this time, current was 0.5A, stopped it, waited about an hour and cells measure 3.59V.

Assembled battery, attached BMS, attached 1600W load via Victron inverter, pack voltage dropped to 25.75-25.8 quickly. After about 40Ah, voltage was 25.6V, which seems on the very low side to me. Cells can't sustain load?

1628143214743.png

More load testing today. With a 78A load, pack voltage was 25.06V.
1628143425548.png
 

Steve_S

Offgrid Cabineer, N.E. Ontario, Canada
Joined
Oct 29, 2019
Messages
5,402
Location
Rural NE Ontario Canada
LFP will ALWAYS Settle post charge, you could charge the cells to 3.650 and <0.01A taken and within 1 hour they'll settle to 3.550 and 2 hours close to 3.500+/-.
 

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
LFP will ALWAYS Settle post charge, you could charge the cells to 3.650 and <0.01A taken and within 1 hour they'll settle to 3.550 and 2 hours close to 3.500+/-.

This is called the relax Voltage :)
Good to remember if you do not want (and you really do not want) to float an already full battery ;)
 

Steve_S

Offgrid Cabineer, N.E. Ontario, Canada
Joined
Oct 29, 2019
Messages
5,402
Location
Rural NE Ontario Canada
Max voltage limits IE: 2.500-3.650 is not "Working Voltage" of 2.900-3.500 (A+ grade) or 3.000-3400 (B+,A-).
I charge Bulk to 3.525Vpc and Float @ 3.500Vpc
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
Your voltages at those SOCs sound on point to me...
I'm a bit puzzled then. According to this chart, a 90% SOC ("nearly full") should have 26.8V and 26.6 while under load. The cells I have drop way down under that (and I get about 225Ah with a 22V low cutoff and 60A load, no-where near 280Ah).
1629038363856.png

Here's another test., fully charged.
1629038616740.png
 
Last edited:

Steve_S

Offgrid Cabineer, N.E. Ontario, Canada
Joined
Oct 29, 2019
Messages
5,402
Location
Rural NE Ontario Canada
All it takes in ONE Cell to reach cutoff point to Trip LVD on the BMS. IF the BMS has no Disconnect Delay for momentary Voltage Sag and it dips too far without a delay to recover again you'll get an LVD trip. A Runner cell is not uncommon, especially with Bulk Commodity Cells, this will reach 3.650Vpc or 2.500Vpc before the rest during CFharge/Discharge cycles, they will also be noticeable with a Lead Voltage Sag. This can generally be observed by closely monitoring the SmartBMS Display of the cells.
 

HighTechLab

Small Business Owner
Joined
Sep 23, 2019
Messages
431
the capacity test discharge curve you have posted shows the exact same inflection points as the curves I see when I capacity test. I think the article you are referencing may be for the older plastic case cells, because I remember those had a much smoother curve with far fewer inflection points.
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
the capacity test discharge curve you have posted shows the exact same inflection points as the curves I see when I capacity test. I think the article you are referencing may be for the older plastic case cells, because I remember those had a much smoother curve with far fewer inflection points.
Are you saying that with only a .2C load (60A of the rated 280A), this voltage drop is in normal range?

The table and curve are from the document that I got from this forum, 'Luyuan_Tech_Basic_LiFePO4_guide_1a.pdf', there's a picture of a 'typical prismatic LiFePO4 cell' at the first page, document 'prepared by @Steve_S'.
 

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
Are you saying that with only a .2C load (60A of the rated 280A), this voltage drop is in normal range?

The table and curve are from the document that I got from this forum, 'Luyuan_Tech_Basic_LiFePO4_guide_1a.pdf', there's a picture of a 'typical prismatic LiFePO4 cell' at the first page, document 'prepared by @Steve_S'.

Did you see Andy's video and cell testing ?

You can even download this test results, compare it to yours.

But yeah the sudden fall of V in high state of charge is total normal.
Also not needed to charge the battery all the way up. Better used between 10-90% SoC. Lives longer.
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
Did you see Andy's video and cell testing ?

You can even download this test results, compare it to yours.

But yeah the sudden fall of V in high state of charge is total normal.
Also not needed to charge the battery all the way up. Better used between 10-90% SoC. Lives longer.
I understand that there is a sudden drop from 100% to about %90 SoC, but after that, this battery seems to have a lower voltage than I was expecting, as per the table I found.

Perhaps I explained this wrong -- I meant the voltage drop between rest and load, although it stays lower than expected after I cut the load.
 
Last edited:

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
I understand that there is a sudden drop from 100% to about %90 SoC, but after that, this battery seems to have a lower voltage than I was expecting, as per the table I found.

V is not an indicator of SoC in these cells. A big load and V fells from the cliffs. No load battery goes up to relax V. (so on every SoC there are 3 different V values -charge,relax,load-, and even they are changing with different size of +/- current )
That is why you need a shunt to know the real SoC of the battery (and a BMS to monitor every cell when is one of them too high or too low).

If too big V drop for load then cells will not give back all the capacity (meaning Grade B cells or used or ... )
I see you get 227Ah out of the pack. That is not good. Did you check if something is hot when testing ?


 
Last edited:

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
V is not an indicator of SoC in these cells. A big load and V fells from the cliffs. No load battery goes up to relax V. (so on every SoC there are 3 different V values -charge,relax,load-, and even they are changing with different size of +/- current )
That is why you need a shunt to know the real SoC of the battery (and a BMS to monitor every cell when is one of them too high or too low).

If too big V drop for load then cells will not give back all the capacity (meaning Grade B cells or used or ... )
I see you get 227Ah out of the pack. That is not good. Did you check if something is hot when testing ?
Thank you for the explanation. I thought it worked like that, I didn't expect the difference between v-relax and v-load to be so big, certainly not for new cells.

Well, yes, the 227Ah is the problem. I got these cells from the same shop that @Will Prowse tested, but the capacity simply isn't good enough. Nothing is hot during the 60A load test (about .2C discharge, for 280Ah cells). One connection on a terminal is slightly warm to the touch, that's all, the rest feels metal-cold ;)

Looking through the reviews, it seems I'm not the only one suffering from this, especially in more recent deliveries.

I'm currently in a dispute, AliExpress stepped in and requires videos etc, that's why I'm redoing these tests.
 

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
Well, yes, the 227Ah is the problem. I got these cells from the same shop that @Will Prowse tested, but the capacity simply isn't good enough. Nothing is hot during the 60A load test (about .2C discharge, for 280Ah cells). One connection on a terminal is slightly warm to the touch, that's all, the rest feels metal-cold ;)
Check that terminal to be sure. Hot is always a sign for bad connection. And you will get more W from the pack that causes more heat ... so problem.


Looking through the reviews, it seems I'm not the only one suffering from this, especially in more recent deliveries.

I'm currently in a dispute, AliExpress stepped in and requires videos etc, that's why I'm redoing these tests.
Will ordered the EVE cells (me too).
This ETC cells I do not know.
But if capacity is so low (81% of the 280Ah) ... get your money back.
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
Check that terminal to be sure. Hot is always a sign for bad connection. And you will get more W from the pack that causes more heat ... so problem.

Will ordered the EVE cells (me too).
This ETC cells I do not know.
But if capacity is so low (81% of the 280Ah) ... get your money back.
It's not 'hot', I can touch and hold it without problems. I doubt this could cause the lack of Ah ;) Currently charging with 40A, everything is cold.

I used the 1st link under his video about these 'VariCore' cells.
 
Last edited:

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
It's not 'hot', I can touch and hold it without problems. I doubt this could cause the lack of Ah ;) Currently charging with 40A, everything is cold.

I used the 1st link under his video about these 'VariCore' cells.

Correction, Will bought Lishen cells. Yours are ETC.
Here you can check the difference: https://diysolarforum.com/threads/noahs-ark-index-of-lifepo4-cell-manufacturers.9759/post-159492

I know it costs a bit more but better to buy from known alibaba sellers like Amy and Dongguan.
 

Martijn

Solar Enthusiast
Joined
May 6, 2021
Messages
117
Correction, Will bought Lishen cells. Yours are ETC.
Here you can check the difference: https://diysolarforum.com/threads/noahs-ark-index-of-lifepo4-cell-manufacturers.9759/post-159492

I know it costs a bit more but better to buy from known alibaba sellers like Amy and Dongguan.
Ah, so it seems VariCore is re-branding multiple brands then. The terminals of ETC look like mine, the pressure valve (?) doesn't though.

Either way, no way these cells are "brand new" as advertised. They scratched off the QR-codes under the black cover.
 

mrzed001

Voice of reason
Joined
Apr 10, 2020
Messages
721
Location
Hungary - EU
Ah, so it seems VariCore is re-branding multiple brands then. The terminals of ETC look like mine, the pressure valve (?) doesn't though.
Yep pressure valve looks like CATL, terminals like ETC ... what is this ?

Either way, no way these cells are "brand new" as advertised. They scratched off the QR-codes under the black cover.

No QR code is a nono.
 
Top