diy solar

diy solar

Strange BMS cell voltage readings

squowse

Solar Enthusiast
Joined
Jan 4, 2021
Messages
572
I have put together a 16S 272Ah Lishen cell pack for testing. I neglected these cells a little so not sure what to expect. Some of them were stored for quite a few months empty.

Doing the capacity test, I've rigged up the computer to the UART interface of Jiabaida (JBD) 6S-21S 150A BMS. Model AP21S001.

Early in the test at approx 18.6A I noticed that cell 3 was close to 3.0V already. When checked with voltmeter it measured about 3.17, lower than the others at approx 3.28, but not as drastically low.

Then - even weirder the computer display showed it up at 3.28ish. But the voltmeter still read it close to 3.19!

It's continuing to fluctuate up and down, approximately every minute or so. Actually mostly reading 3.27 now. I will be able to get the plot once the test has concluded. Possibly early, as the BMS may well trip out for low cell voltage.

I've checked the other cells and all read near enough the same on voltmeter as on computer. All 3.28ish with minimal variation so far.

Has anyone experienced something like this? If it was a bad connection (ive checked tightness) how would the voltage end up reading higher?

If the BMS voltage ADC was bad onthat channel - it's a coincidence it happened on the "bad" cell?

69EBC67F-F85E-4E13-89DD-05B9F43F4835.jpeg
55C83512-7691-4A4D-B4BB-596F76D72014.jpeg
074C9EA8-9150-4CCB-BD4D-EBAF09C79BC1.jpeg
F58642ED-8E06-4702-A06E-34D1D8D4B222.jpeg
 

Attachments

  • E297DF10-6961-452D-BE90-3C646A718A22.jpeg
    E297DF10-6961-452D-BE90-3C646A718A22.jpeg
    470 KB · Views: 5
So I decided to switch cells 2 and 3 and now it's behaving itself. So far. Unfortunately it's done the thing where it resets the SoC to its bad guess based in voltage. So I really need to charge back up again so I can do the full cycle to initialise the BMS SoC measuring. Very annoying you can't manually input the SoC and force it to use the shunt only.
 
Charging back up to 100% and cell2 is now having a fluctuating reading - up to about 100mA either side of the multimeter reading, about every 4-5 seconds now. I wonder if this could be linked to the BMS balance lead wiring? I've checked for loose connections.

When I first switch it on, it auto adjusts itself to 17 cells. I have followed the instructions and also this diagram -
My model is slightly different to others in that the first connector has 15 pins, and the second one 7 pins.
 

Attachments

  • AP21短接线图英文版_12.JPG
    AP21短接线图英文版_12.JPG
    92.3 KB · Views: 9
I think that may have been the balancing, I could see how you might get a variation each end of the balance cable if it's doing something with it at the BMS end?

Anyway Im back on the discharge test, and now the cell 12 reading is fluctuating. It's the lowest cell but the reading drops 200mV for no apparent reason.
 
If cell voltage readings go significantly different with load current, there are a several possible reasons.

First are some poor terminal connections. A cell terminal connection with higher resistance will add to voltage reading when current is applied. Make sure the sense wiring is done consistent so there is not more than one bus bar between adjacent BMS sense wires.

Cells with low state of charge will have greater internal impedance so will drop more in cell voltage with load. Same applies to charge current.

A poor condition cell, even at same state of charge, will have greater cell internal impedance causing greater cell voltage slump with load current.
 
If cell voltage readings go significantly different with load current, there are a several possible reasons.

First are some poor terminal connections. A cell terminal connection with higher resistance will add to voltage reading when current is applied. Make sure the sense wiring is done consistent so there is not more than one bus bar between adjacent BMS sense wires.

Cells with low state of charge will have greater internal impedance so will drop more in cell voltage with load. Same applies to charge current.

A poor condition cell, even at same state of charge, will have greater cell internal impedance causing greater cell voltage slump with load current.
How does that fit with the voltmeter readings being normal? I even put the probes on the balancing lead lugs.
 
Some of them were stored for quite a few months empty.
Depending how low the SoC got, this can cause conductive dendrites to grow creating leakage in cell.
Early in the test at approx 18.6A I noticed that cell 3 was close to 3.0V already. When checked with voltmeter it measured about 3.17, lower than the others at approx 3.28, but not as drastically low.

Then - even weirder the computer display showed it up at 3.28ish. But the voltmeter still read it close to 3.19!
Assuming you did not check the voltages after you stopped the load current this would likely indicate connection issues. Check if that cell's bus bar is warm. It takes about 3-5 minutes for a cell to reach equilibrium voltage after a cell current change so if you change the current the cell voltage will time shift to its new current equilibrium. You need to wait until equilibrium is reached to get a stable cell voltage reading.

The BMS shares each sense wire connection between two cells. If you have a bad bus bar connection, the voltage drop of the poor connection is either going to be subtracted or added to an individual cell actual voltage reading on BMS depending how the sense wires are distributed. You should always put the sense wires on the same side of cells so each cell gets a bus bar in the series voltage measurement. The first or last cell will not be able to have the additional resistance of bus bar so that cell will read different under significant load current than other cells which have the voltage drop of the bus bar connection included with BMS voltage measurement.
Good battery connecitons.png
Bus bar connection testing.png
 
Last edited:
Thing is that the voltage fluctuates on the BMS display while my voltmeter connected to the balance lead crimp rings shows steady voltage. I think it may have something to do with capacitance of the voltage sensing circuit.
 
Right. What does the voltage read on your multimeter between the crimp ring and the wire the crimp ring is connected to. You know that crimp that's my point.
The voltage of the cell apparently is not actually fluctuating (DMM) it is only the device that is measuring the cell that is fluctuating (BMS).
So look at each and every single contact point. Terminal to busbar, bus bar to ring log, ring lug to wire, wire to pin at the BMS. PIN to pin....
Something's loose.
 
Right. What does the voltage read on your multimeter between the crimp ring and the wire the crimp ring is connected to. You know that crimp that's my point.
The voltage of the cell apparently is not actually fluctuating (DMM) it is only the device that is measuring the cell that is fluctuating (BMS).
So look at each and every single contact point. Terminal to busbar, bus bar to ring log, ring lug to wire, wire to pin at the BMS. PIN to pin....
Something's loose.
Yeah it would need to be between the ring and the BMS. My meter was between two crimp rings. Unless the BMS measures between the top or bottom and each individual ring? Its done this with two different BMS that each had their own set of balance leads.
 
I finally built my battery pack after the ice storm took out the tranmission lines and the house drop a couple of days ago. Things were going fine until I tried connecting to the 16s 100A JBD BMS. The OverKill app would not connect at all. The JBD app on the Android App Store found it. But it was showing random voltages, fluctuating between 0.5 and 6 volts for each cell (i.e., the same cell would have changing voltage). All of the cells were at around 3.4v. The BMS never turned on (probably because of the odd voltages), so it's not a loading issue. Double-checked the order of the connections and they seem correct. Any ideas? The black connector wires seem a bit odd, but that's how they came. I think the base PCB is just made for up to 20 or 21 cells and they only populate the number you order.

I eventually broke the pack down and hooked up 4 cells to a Yeti 1250 and used a cheater cord to backfeed the panel to run the furnace and fridge (with the main breaker off, of course)
 

Attachments

  • PXL_20230225_151416421.jpg
    PXL_20230225_151416421.jpg
    214.3 KB · Views: 6
Back
Top