JKBMS SoC is drifting

[PVJunkie]

Hi,
I´m using a JKBMS since 1 year.
To test the SoC I discharge the battery until undervoltage warning occurs (lowest cell < 2.65 V).
At the beginning (after some full cycles) the SoC showed 7% at this Point. Sounds good for me.
During summer I had a discharge limit of 10% and it worked well.
But with start of the dark season I got undervoltage warnings at 10%. I increased the value. But it happend again and again. I retestet the Soc and now it´s 41% when the undervoltage warning occurs.

Is this a known Issue?

I can´t see a possibility to reset the SoC calculation. Will it be resetted at the next time the battery would be completely charged?

 

[DIYrich]

Do you fully charge the battery to 3.40v? That is when the bms recalibrate. Not at the bottom end.

 

[PVJunkie]

It wasn´t fully charged since 3 months. So it´s clear that SoC isn´t recalibrated since 3 months.
But I think this is normal operation in many countries

 

[DIYrich]

It wasn´t fully charged since 3 months. So it´s clear that SoC isn´t recalibrated since 3 months.

That's your problem.

 

[sunshine_eggo]

It wasn´t fully charged since 3 months. So it´s clear that SoC isn´t recalibrated since 3 months.
But I think this is normal operation in many countries

Not sure what this means. ALL battery monitor devices require periodic synchronization with a full charge, or their accuracy will drift. This is just a fact and independent of anything like "normal operation in many countries".

A discharge test should always follow a charge to full when practical. It's the only way you can actually measure the capacity.

 

[DIYrich]

If you can't fully charge at least every 2 weeks, ignore SOC and go by voltage for low limit.

 

[PVJunkie]

"normal operation in many countries" means that in my opinion the bms should be able to handle a few month without recalibration because there are a lot of countries with only 1-2h of sun over months.
A friend of mine runs a Pylontech system and the SoC calculation is much better after 3 months without full charge. He has a deviation of ~5% instead of ~40% at my JKBMS.

What is the exact trigger point of SoC reinitialisation. Is it when the first cell reaches "Cell OVPR(V)"?

 

[PVJunkie]

If you can't fully charge at least every 2 weeks, ignore SOC and go by voltage for low limit.

I check the cell voltage, too. It´s completely safe.
But it would ne nice to see a working SoC

 

[sunshine_eggo]

"normal operation in many countries" means that in my opinion the bms should be able to handle a few month without recalibration because there are a lot of countries with only 1-2h of sun over months.

Opinions don't matter. Reality is that the BMS keeps track of SoC by counting current in and out of the battery. This is inherently inexact, and it can't possibly take into account things like cell self-discharge variation. Even Victron, who is generally regarded as a quality company, indicates their battery monitors need to be re-synched "a couple of times a month" to maintain accuracy.

If you're on 1-2 hours of sun, you must be using an alternate source. Battery maintenance should include periodically using that alternate source to charge to full. All batteries have some maintenance requirement. A battery that tracks SoC needs to be periodically re-synched to 100%.

A friend of mine runs a Pylontech system and the SoC calculation is much better after 3 months without full charge. He has a deviation of ~5% instead of ~40% at my JKBMS.

How does your friend know how accurate the BMS is?

What is the exact trigger point of SoC reinitialisation. Is it when the first cell reaches "Cell OVPR(V)"?

It depends on the BMS. Typically between 3.50 and 3.65V/cell. OVPR certainly will trigger 100%.

 

[Lorax_UK]

I am in a very similar situation, running an off-grid system with 32kWh of batteries, controlled by 2x JK inverter BMS units.

I previously used PylonTech batteries, which in the years I used them, did not seem to suffer from SOC drift like the JK BMS units do.
My reason for saying this, is that unless the JK system gets to 100% regularly (every 2 weeks or so), the SOC becomes very inaccurate. This shows itself when the SOC shows, for example 45% capacity, then immediately shuts down due to low cell voltage (2.65v).
The PylonTech units never suffered this problem, even when they didn't get charged to 100% for weeks/months.

The JK BMS units I'm using are PB2A16S20P, with Firmware version 15.29
I will try updating the firmware to see if this helps.

I'm not sure about the narrative that SOC is inherently difficult to measure without regular charges to 100%, as it does not explain how PylonTech managed to do it.

 

[Lorax_UK]

This shows my system yesterday, with battery SOC apparently at 51% at 11pm, with almost zero consumption, before low cell voltage protection cuts the power all together, showing that the SOC is almost zero.

 

[Detritalgeo]

I feel this fact needs to be more prominent;

Drift occurs any time charge or discharge is less than 0.5A in most 100A BMS's. That's per BMS, so multiply that by how many batteries you have.

Current below this cannot be sensed or tracked. Imagine your car odometer not increasing as long as you drive less than 10 mph...

Charging to full permits an adjustment of the SOC to a known value to correct forthis drift. Some people have had better results with relying on a single smart shunt to to minimize the drift time. @Adam De Lay has a video illustration of this.