Single cell jumping 400mV on charging

[mcart117]

One cell out of eight in a 24v 8s LFP battery is jumping from 3.336V when idle to 3.791V as soon as the battery begins to charge. Any ideas as to why this might be? Is the BMS possibly giving false readings or does this cell have a serious problem?

I made a screen recording of what happens when the battery goes from an idle state to charge, but the format .mov doesn't seem to be supported, so I'll insert a couple of screen shots instead. Watch the voltage of cell 4.



I’d appreciate any thoughts on what to do next.

 

[Bud Martin]

Is this brand new setup or was it OK before and now the issue is showing up?
I would check the wire/connector of the BMS of that suspect cell to make sure it is crimped well and make good contact with the battery terminal, you can also move that suspected cell to another location and see if the issue follows.

 

[MisterSandals]

Is the BMS possibly giving false readings

You can check this with a volt meter.

And label your cells and take notes, especially if you move them around as Bud Martin wisely suggests.

Were these cells ever top balanced?

Is there a reason balancing is turned off?

Happen to have a pic of your battery wiring?

 

[mcart117]

The system has been running well for some time and is described in more detail here.

The BMS has a section of the screen which I cut out, which I think answers your question:



If there was a problem with the wiring I think it would show up as an anomolous resistance there.

 

[mcart117]

Is this brand new setup or was it OK before and now the issue is showing up?
I would check the wire/connector of the BMS of that suspect cell to make sure it is crimped well and make good contact with the battery terminal, you can also move that suspected cell to another location and see if the issue follows.

I'll try moving the cell and let you know what happens.

I have also replied to you in the general thread.

 

[mcart117]

Is there a reason balancing is turned off?

The BMS turns balancing on and off as needed, and I think it looks for a difference of 10mV, which thses cells didn't have when charging was turned off. I think the voltage of this one cells goes up so quickly that the cell over-voltage protection kicks in before the balancing function has time to turn itself on.

But it is not manually disabled, and has been working well over the last few months, as set out here.

 

[MisterSandals]

I think the voltage of this one cells goes up so quickly

That is strange. I read the other thread, thanks for posting a link. I think move it to another location and see if the problem follows the cell.

Or it looks like you have the ability to capacity test so maybe charge it and make sure the capacity is still there.

These things normally are connections but I've never seen a bad connect to cause the voltage to spike. Did you verify that the cell voltage is what the BMS is reporting? You could test the harness too.

 

[mcart117]

You can check this with a volt meter.

When I went back later this afternoon, I was able to confirm that the BMS was reading true voltages. After "resting" a couple of hours the cell was not spiking as much, so the battery continued charging for long enough for the balancer to kick in.



I then switched to discharge mode, and again this cell was running ahead of the pack, first by around 10mV, then 20mV, then 35mV, and I stopped watching when it got to 55mV (see above). All the while the voltage was jumping: up 20mV, down 20mV.



On a hunch I went back out to see what had happened while writing this post, and now the anomalous cell is jumping between 150mV and 200mV below the rest. Soon it will hit my 3V floor and the battery will switch off, as it did when it was charging.

Happen to have a pic of your battery wiring?

The cells are numbered, but on the other side, which was facing me when I put it together. You can see from the main negative and positive cables, cell 4 is fourth from the left.

I do have a couple of spare LFP cells from another supplier, so I could remove this cell for testing and replace it with one of these, but I'll have to set aside time to do that. There is another 24V 8s battery connected in parallel, and that will have to carry all the load until I get this fixed.

 

[RCinFLA]

Not sure why your first picture is showing balancing 'OFF' and you have charging 'ON' when you have a cell at 3.791v. You must have a setting in parameters preventing balancing and cell overvoltage trip set too high to still have charging enabled with a cell at 3.79v.

I think some versions of firmware only show user manual enable/disable of 'Charge' and 'Discharge' at top of status screen and some FW versions will change the top indicator if a BMS sensed error causes either charging or discharging to be disabled, so not too concerned about Charging ON indicator being on. You should always leave manual settings for 'charge' and discharge' ON unless only doing some testing.

But you should not be showing a charge current when there is a cell at 3.79v. This is likely because you have cell overvoltage threshold set too high. You might have the default battery type set LI-ION. It should be set to LIFEPO4. Warning - when you change the battery type all the setting are rewritten to firmware default so don't change it once you do all your custom settings or it will clear them all and you have to re-set them all up again.

You should post your setup screens so we can see what setup is incorrect.



This video walks through the settings for JK active balancer BMS.

 

[TomC4306]

All the while the voltage was jumping: up 20mV, down 20mV.

Something is not tight.
You seem insistent on relying on the BMS lead resistance numbers. Small tug on the wire at the ring lugs with your thumb and index finger and re-torque with an actual torque wrench all the connections.

Edit.. also, in your second image the voltage difference is reading 0.004 when actually it's 0.375.
Did anyone else notice that?

 

[mcart117]

You should post your setup screens so we can see what setup is incorrect.

The settings in the first screenshot are mine. Continued discharge current is mine. Most of the rest are unchanged from shipment.

The OP explained the screenshot you annotated was taken from a video showing the cell voltage is jumping from 3.336V when idle (ie after the cell overvoltage protection has kicked in) to 3.791V as soon as the cell overvoltage protection toggles off and the battery begins to charge again. The cell overvoltage protection then reactivates, the cell voltage drops back to 3.336V, and the process repeats.

 

[Bud Martin]

Something is not tight.
You seem insistent on relying on the BMS lead resistance numbers. Small tug on the wire at the ring lugs with your thumb and index finger and re-torque with an actual torque wrench all the connections.

Edit.. also, in your second image the voltage difference is reading 0.004 when actually it's 0.375.
Did anyone else notice that?

Good catch! I just notice that too.

 

[mcart117]

Something is not tight.
You seem insistent on relying on the BMS lead resistance numbers. Small tug on the wire at the ring lugs with your thumb and index finger and re-torque with an actual torque wrench all the connections.

Edit.. also, in your second image the voltage difference is reading 0.004 when actually it's 0.375.
Did anyone else notice that?

OK. I appreciate all replies, but we seem to wandering off topic.

I took a screen recording of the erratic behaviour of this MFP cell and tried to post it here but the forum doesn’t recognise the .mov format. I therefore posted three single frames, thinking it would be enough to explain what is happening.

A couple of people such as @Bud Martin and @MisterSandals posted good questions and made good observations, but recently there have been some less helpful posts with in my opinion somewhat supercilious observations from the single frames I posted.

I have therefore uploaded the whole video to a third party server, and anyone who is interested can watch it here.

What I am looking for is anyone who has witnessed this type of behaviour in one or more LFP cells in an 8 or 16 cell battery. Please focus on the voltage movements of cell number 4. Don’t worry about anything else the BMS is or isn’t reporting.

 

[WindWizard]

Ii sounds like the best approach as far as I can tell after reading this is to take out a voltmeter and measure the voltage drops between the studs and the plates between the cells. Keep in mind that the only time that you can perform this test is when the cells are either being discharged or charged at least 20 amps or so. If the voltage drops are consistent between the cells to the studs including the suspect cell. Another thing to check is that if there is a significant amount of resistance due to a loose stud, then it does produce heat. An infrared camera can show that as well as just feeling the terminal and making sure it is not getting warm. If the voltmeter and infrared test check out, then the only other explanation is a defective cell. The Cell would need to be removed and then checked for capacity and performance.

 

[MisterSandals]

I have therefore uploaded the whole video to a third party server, and anyone who is interested can watch it here.

I watched this a few times and its strange. I got nothing new from it.

I blew up the pic of your battery and noticed some inconsistencies with balance wires, not sure if its a material/meaningul difference or not.
It looked like some of the balance wire rings might have washers under them, some on top.
The bus bars are pretty inconsistent looking: some shiny spots, some dull spots, some black spots.
Its a blurry image when made big so these are just casual observations.

But, I'm going back to the idea that its an inconsistent connection problem. One that manifests itself when charging amps are applied to it.

I recommend disassembling, cleaning the cell terminals, bus bars and ring terminals. Move cell 4 to an adjacent location. Hopefully these cells are all labeled and have notes about this already. If the problem stays in cell #4 position, it'd make a new set of balance wires.

Weird chit like connectivity issues are known to fool electronics into erratic and unexplainable behaviors.

Good luck, i wish i had a smoking gun for you.

 

[mcart117]

I watched this a few times and its strange. I got nothing new from it.

I blew up the pic of your battery and noticed some inconsistencies with balance wires, not sure if its a material/meaningul difference or not.
It looked like some of the balance wire rings might have washers under them, some on top.
The bus bars are pretty inconsistent looking: some shiny spots, some dull spots, some black spots.
Its a blurry image when made big so these are just casual observations.

But, I'm going back to the idea that its an inconsistent connection problem. One that manifests itself when charging amps are applied to it.

I recommend disassembling, cleaning the cell terminals, bus bars and ring terminals. Move cell 4 to an adjacent location. Hopefully these cells are all labeled and have notes about this already. If the problem stays in cell #4 position, it'd make a new set of balance wires.

Weird chit like connectivity issues are known to fool electronics into erratic and unexplainable behaviors.

Good luck, i wish i had a smoking gun for you.

Thanks MisterSandals, I'll try that.

 

[mcart117]

Ii sounds like the best approach as far as I can tell after reading this is to take out a voltmeter and measure the voltage drops between the studs and the plates between the cells. Keep in mind that the only time that you can perform this test is when the cells are either being discharged or charged at least 20 amps or so. If the voltage drops are consistent between the cells to the studs including the suspect cell. Another thing to check is that if there is a significant amount of resistance due to a loose stud, then it does produce heat. An infrared camera can show that as well as just feeling the terminal and making sure it is not getting warm. If the voltmeter and infrared test check out, then the only other explanation is a defective cell. The Cell would need to be removed and then checked for capacity and performance.

Thanks for your input WindWizard. I'll have some time this weekend or over the holiday to remove this cell and run some tests. The only problem is my ZKE battery tester charges from memory at 5 amps, and as you rightly observe, the erratic behaviour kicks in when charging or discharging at 20 amps or above. I think the ZKE discharges at 20 amps, so we'll see what happens then.

 

[RCinFLA]

Looks like a lot of poor battery connections.
- one really bad on cell #4
- two moderately bad on cells #5 and #6
- three marginal on #2, #7, and #8

The no current BMS cell voltage reading pretty much proves BMS is not the issue.

The numbers can shift to adjacent cell depending which side of bus bar sense wire connected. Make sure you keep sense wires on a consistent side.

Should have about 0.2 milliohm on cell connections with terminal (0.05 mohm)-bus bar (0.07 mohm) - terminal (0.05 mohm) summed resistance.

You will get more accurate indications with greater battery current. 4.6 amps at typ 0.2 milliohm is less than 1 mV which will not be very accurate voltage measure due to low end limit of resolution of voltage measurement. Need to test with more like 20-30 amps of battery current. Voltage of cell will have an exponential additional voltage shift over 30-180 secs so make measurement quickly after applying current.

Best to use a separate DVM with at least 200 mV lowest scale and check from battery terminal to adjacent battery terminal, not on top of bus bars, to check bus bar and terminal connection voltage drop. Do for each bus bar connection.

Clean aluminum oxide build up from aluminum cell terminals with 50% diluted white vinger then clean and dry with isopropyl alcohol.

Could also be caused by old used cells with high internal resistance but I doubt you could have one so bad it would cause a number like cell #4 reading.

 

[RCinFLA]

 

[mcart117]

Could also be caused by old used cells with high internal resistance but I doubt you could have one so bad it would cause a number like cell #4 reading.

Thanks for both of those. I'll check the connections. The cells are just 3 months old, but I will check the IR of Cell 4.