Yet another erratic SOC readout case of JK BMS :-(

[djnitehawk]

so... my first jk bms crapped out (stopped detecting voltages of 3 cells), and i got a new one with a newer firmware version.
everything works great except for the life of me i can't get the SOC % to not jump around when both charging and discharging.
you can see in the screen recording below, that it jumps from 28% to 95% even when there's no significant change in the cell/pack voltages, let alone being anywhere close to the OVP value.
i have cycled... "calibrated" between OVP and UVP multiple times causing it to go in to protection mode on both top and bottom ends.
but it still does this huge jump during both charging and discharging.
i've tried clearing cycle counts by changing pack capacity values but it doesn't seem to help either.

what do you guys think? could this be a firmware issue or am i doing something wrong?

btw, the recording was done using this app which basically reads the same data as the jk mobile app but via rs485 port. it just wasn't practical for me to do the recording on a mobile phone.

 

[chrisski]

I’d like to hear from other users, but I have given up on using a BMS to determine SOC. I like the other features like individual cell voltage and over all voltage and amperage, and of course the cutoff, but don’t pay attention to the SOC. Mine does not jump around, one of two BMSs is inaccurate.

I only have an Overkill, which is rebranded JBD; the other BMS are tiny with no way to access readings.

For me, the SOC is only correct on 1 of 2 BMSs. It does not jump around, but does read consistently low. I’d hoped there was a “learning process” to calibrate it, but does not work. I disconnected it and it got went from 20% off to 10% off

 

[burgerking]

I have long abandoned SOC as well.

Simply put, you cannot count electrons and expect all electrons are accounted for.

Just imagine: if your BMS has an accuracy of 99.99% or 0.01% error.
In 1 hour time it has accumulated 0.01% error; in 24 hours 0.24% error.
How much error do you think it would have accumulated in 10 days?

The only way it will self-correct is when you full charge or full discharge the battery wherein the SOC% will reset to 100% or 0% respectively. In my case my JK was doing great the last 2 months fully reporting SOC% (and AH) correctly because I get to full charge them everyday... and just last hot Sunday my JK overheated (>39C) and all of a sudden my battery did not get full charged and SOC% skewed.

Battery Voltage is the fall back to all of this and I'd go back to using Batt Voltage hence forward.

 

[hwse]

I am curious about your settings.

3.4v is not even up to the high knee and would not be considered to be at full charge.

Starting balance at 2.8v is pointless because the cells will only show an out of balance condition at the high or low knee. When I first connected my cells, they showed 3mv imbalance until one cell went over 3.45 and then that cell rapidly went to 3.65v and shut down the pack with almost 300mv imbalance. I then reduced charge to 2A and the JK continued to charge until all cells were balanced at 14.0v with 3.50v/cell.

Your log shows several over and under voltage protection events. Are these during testing? I have had a total of one OVP and one UVP. I use them as an emergency safety measure. I have OVP at 3.60v and charge my 12v system to 13.8v with float at 13.4v

 

[djnitehawk]

@hwse my cells are B grade bloated cells, so i'm not charging to 3.65v ever.
i've chosen my working voltage range to be 3.0v to 3.4v.
and i don't have communication between the inverter and jk bms.
so the inverter can't successfully stop charging when the cell max cell voltage reaches 3.4v. jk is the one in control of the HVD.
on the lower end however, the inverter is set to a total LVD of 24.0v and inverter stops dischage before the UVP of the jk kicks in.
those log entries are just me trying to "calibrate" the jk by doing a full charge/dischage cycle by letting them reach UVP and OVP points.
at the end of the day, there's definitely something wrong with the firmware of the new jk as my previous jk did not have this issue.
i've also noticed that the coulomb counting/current sensing on the new fw is a out of whack. even if i calibrate the current at a steady 30A flow it's not in sync after a few days of use. i.e. after a few days, even though the battery is charging at a 30A steady current, the current detected by jk is off by +/- few amps.

the unfortunate thing is, i bought this from a local reseller and he's saying that jk has come back saying they need proof using a screen recording of the jk app instead of my own app. which is next to impossible to do since i need to keep the phone recording the screen the whole day without using my one only phone :-(

PS: the pack is bottom balanced and balancing is turned off. i only re-balance every 6 months or so. that's why balance current setting is irrelevant here.

 

[uksa007]

so the inverter can't successfully stop charging when the cell max cell voltage reaches 3.4v. jk is the one in control of the HVD.

This is not a good design, the BMS is safety device not for control of normal charging.
It's the last line of defense. Like a circuit breaker and should only operate under a fault condition.
Set your inverter to charge at the max voltage you want. eg 3.45V per cell, it should be configured to stop charging once it reaches the set charge voltage. (some inverters you configure a lower float voltage to acheachive this)
Charging to 3.4v per cell is a bit too low, and it's why you are probably having issues reliably charging.

Your SOC problems are probably because you have the OVP set so low, set it to the default 3.6v and I bet it works.
This is a result of when you charge, the higher current the higher voltage will rise, more than the cells have absorbed, so the bms thinks the cells are almost charged as the voltage it getting close to the low set OVP. You essentially have no buffer to allow for this. Hecne the SOC jumps.

The upshot is you should consider tweaking your working voltage and use your inverter to control the charging not the BMS, set the OVP back to 3.6v

 

[Ucatchmydrift]

I have 2 200amp jk bms's, and the soc is not that accurate either, but im not fussed as i use a couple of shunts, which are far more accurate. (junctek kg140f- which are cheap and good). My inverter soc suks too.