JKBMS Inverter Edition Problems/Issues | No Support / Help to fix major issues. - DO NOT BUY ! Warning (as of Oct.12.2024)

[kommando]

SMA wrote the first Canbus protocol but also programmed their Inverters to ignore 100% SOC and continue to charge until the BMS Target V was met. Pylontech cloned the SMA Canbus making enough changes to cover their tracks. A lot of the Chinese AIO's use Pylontech but did not copy SMA and decided to stop charging on receipt of a 100% SOC Canbus message.

So having closed loop is one thing, the Inverter and Battery working together the way you want is another.

 

[Steve_S]

12,000W Inverter will pull 234A (nominal) +/- from a 48V (51.2V nominal) battery system.
Seems that "someone" forgot, 1 JK 200A BMS can deliver 200A and up to 350A in Burst mode for a few seconds (I think it is 3 sec).

If you have ONE battery Pack, it's output is capped by the BMS and what it can deliver.
If you have 2 or 3 batteries in PARALLEL, then load is proportionally divided between battery packs. Assuming 3 Packs with 200A Capabbility, they can deliver 600A cumulatively and take 300A Cummulatively to charge.

NOTE: Proportional ! it is all relative to size/capacity
100AH, 200AH & 300AH CAN coexist in a bank. If you pull 600A from that, it will be delivered as 100A, 200A, 300A which is a result of the BMS but also the chemistry limitations (by design) as well.


NOW - SORRY FOLKS BUT THIS IS NOT A TECH SUPPORT THREAD !
TOPIC IS:

JKBMS Inverter Edition Problems/Issues | No Support / Help to fix major issues. - DO NOT BUY ! Warning (as of Oct.12.2024)​

 

[empowered]

SMA wrote the first Canbus protocol but also programmed their Inverters to ignore 100% SOC and continue to charge until the BMS Target V was met. Pylontech cloned the SMA Canbus making enough changes to cover their tracks. A lot of the Chinese AIO's use Pylontech but did not copy SMA and decided to stop charging on receipt of a 100% SOC Canbus message.

So having closed loop is one thing, the Inverter and Battery working together the way you want is another.

so whats the 'fix'?

 

[empowered]

12,000W Inverter will pull 234A (nominal) +/- from a 48V (51.2V nominal) battery system.
Seems that "someone" forgot, 1 JK 200A BMS can deliver 200A and up to 350A in Burst mode for a few seconds (I think it is 3 sec).

If you have ONE battery Pack, it's output is capped by the BMS and what it can deliver.
If you have 2 or 3 batteries in PARALLEL, then load is proportionally divided between battery packs. Assuming 3 Packs with 200A Capabbility, they can deliver 600A cumulatively and take 300A Cummulatively to charge.

NOTE: Proportional ! it is all relative to size/capacity
100AH, 200AH & 300AH CAN coexist in a bank. If you pull 600A from that, it will be delivered as 100A, 200A, 300A which is a result of the BMS but also the chemistry limitations (by design) as well.


NOW - SORRY FOLKS BUT THIS IS NOT A TECH SUPPORT THREAD !
TOPIC IS:

JKBMS Inverter Edition Problems/Issues | No Support / Help to fix major issues. - DO NOT BUY ! Warning (as of Oct.12.2024)​

my bms is the 300amp

 

[kommando]

so whats the 'fix'?

JK changed firmware to keep reporting 99% SOC until the Target V is met.

 

[Steve_S]

my bms is the 300amp

JiKong - JKBMS "Proper" inverter model BMS is MAXXED at 200A with up to 400A momentary surge handling.
The Funky Clones labelled as JKEMS had a supposed 300A model but THOSE ARE NOT JK by JiKong.
Pushing ANY FET Based BMS to the Max Rating or above is almost like Jumping Up & Down, demanding that Murphy's Laws get applied to you ! in simple worlds, NEVER push them to the Max. We tell everyone to DERATE so you never touch the limits.

Read the Specs & Docs, and make NO ASSUMPTION ever ! Making assumptions, leaping to conclusions from there serves no purpose of any value at all.

AGAIN: This is NOT a Support my product thread ! This is about the ongoing troubles, lack of support/communications for owners of these JKBMS Inverter model BMS' which is quite plainly outlined from Page 1 Post 1 & 2 and onwards with the occasional sidetrack (as always hap[pens).

 

[empowered]

JiKong - JKBMS "Proper" inverter model BMS is MAXXED at 200A with up to 400A momentary surge handling.
The Funky Clones labelled as JKEMS had a supposed 300A model but THOSE ARE NOT JK by JiKong.
Pushing ANY FET Based BMS to the Max Rating or above is almost like Jumping Up & Down, demanding that Murphy's Laws get applied to you ! in simple worlds, NEVER push them to the Max. We tell everyone to DERATE so you never touch the limits.

Read the Specs & Docs, and make NO ASSUMPTION ever ! Making assumptions, leaping to conclusions from there serves no purpose of any value at all.

AGAIN: This is NOT a Support my product thread ! This is about the ongoing troubles, lack of support/communications for owners of these JKBMS Inverter model BMS' which is quite plainly outlined from Page 1 Post 1 & 2 and onwards with the occasional sidetrack (as always hap[pens).

Amazon.com

the app is jikong, all the paperwork and box says jikong.

 

[JJJJ]

@empowered you have a number of questions. May I suggest starting a new thread? It is quite likely that you will get better answers to your questions. On a personal level I have found this approach very helpful.

 

[Reebcord]

Мне лично понравилась поддержка ГНФАР.
Они ответили на мой запрос поддержки от 8 июля вчера и сегодня. Который не есть, но поддержка

Приветствую вас, сэр, у меня проблема с моим sofar hyd6ep и bms jk, когда soc достигает 100 процентов заряда, зарядка сразу отключается и нет плавного уменьшения тока, как на моем старом аккумуляторе с bms daly, помогите мне, если вы знаете, как это исправить

 

[HRTKD]

Приветствую вас, сэр, у меня проблема с моим sofar hyd6ep и bms jk, когда soc достигает 100 процентов заряда, зарядка сразу отключается и нет плавного уменьшения тока, как на моем старом аккумуляторе с bms daly, помогите мне, если вы знаете, как это исправить

Translation:
Hello sir, I have a problem with my sofar hyd6ep and bms jk, when the soc reaches 100 percent charge, the charging immediately cuts off and there is no smooth decrease in current like on my old battery with bms daly, please help me if you know how to fix this

 

[Reebcord]

JK changed firmware to keep reporting 99% SOC until the Target V is met.

what does it mean? That is, there will not be a smooth decrease in current as it should be at the end of charging 100℅) Will it just hang at 99 percent longer and wait until it reaches the rcv point? right

 

[RCinFLA]

As Columb counter builds errors over time and the discrepancy in reported to actual SoC grows, the full charge referencing point that resets the Columb counter will usually always have a step shift in reported SoC.

When Columb counter cumulative error results in excess SoC reporting, some algorithms just hold at 100% readout until full charge trigger locks in the new 100% charged reference point. Others, dumbly, just keep reporting growing SoC beyond 100% until full charge trigger is tripped causing readout to abruptly to jump to new 100% SoC refreshed reference point.

This can also happen with under reporting of SoC error where readout jumps from a lower % SoC to abruptly 100% when full charge trigger is tripped.

The one good thing about seeing an abrupt SoC % jump is you know 100% full charge has been tripped resetting Columb counter and washing out cumulated errors.

The main difficultly is algorithm that detects 100% full SoC.

In real world usages, there can be variable inverter use loading that causes battery current to jump between charging current and discharge current randomly. PV charging current varies as clouds go by. All these things create variable overpotential voltage based on cell current that greatly complicates determination of when 100% full charge is reached.

Designing an algorithm to determine full SoC that is based on a rather sterile laboratory condition of stable constant charging current with well-behaved absorb voltage set point with gradual charge current taper off will likely have difficulties in real world conditions.

Best determination of full SoC is looking for a full charge battery voltage when current in and out of battery has been low for some time assuring cell voltage equilibrium has pretty much been reached.

There also needs to be a backup algorithm when use case activity prevents enough low current time to establish equilibrium. Cell voltage look up table is a good indicator of SoC when cell is at equilibrium, not having significant in or out current or some period of time.

You don't want Columb counter reset to be too strict as recalibrating too infrequently can be worse than too much error in determine full charge reset of Columb counter.

Another complication is trying to make BMS idiot proof by having automatic AH determination. Adding error prone 0% SoC determination to error prone 100% SoC determination only makes matters worse with an incorrect battery AH capacity for SoC to reference against.

I would also like to have a manual user triggered 100% fully charge trigger to reset Columb counter manually. Better yet, give me the ability to punch in a SoC % number then manually set Columb counter to that level. -- and turn off automatic AH determination.

 

[kommando]

what does it mean? That is, there will not be a smooth decrease in current as it should be at the end of charging 100℅) Will it just hang at 99 percent longer and wait until it reaches the rcv point? right

Yes.

 

[TechSree]

Did anyone faced Bulk to Absorption change issue in JK Inverter BMS ?. is it a new bug with firmware 15.38 ?. using mobile app 5.0.1

T

Thread 'JK Inverter BMS Firmware 15.38 Stuck on Bulk- not going to Absorption and Float. new bug?'

Apr 2, 2025

Using JK ESS BMS and firmware version 15.38, Mobile app 5.0.1. When checking the status of charging its not always working correctly as per the new charging algorithm. I tried with the following values
RCV - 3.460, SOC 100% - 3.450, RCV Timer - 1, RFV Timer -6 - In some days it will go to Absorption after 4 or 5 hrs of of reaching all cells to 3.450V
RCV - 3.450, SOC 100% - 3.440, RCV Timer - 1, RFV Timer -6 -In some days it will go to Absorption after 4 or 5 hrs of of reaching all cells to 3.440V
Also noticed that when changing the RCV from 3.450 to 3.457 then back to the old value after...

• TechSree
• Replies: 8
• Forum: DIY Solar General Discussion

 

[RCinFLA]

'Stuck in Bulk' is really a lack of achieving absorb voltage level on battery, as far as BMS observes.

Most likely cause is calibration difference between BMS absorb voltage and inverter/charger absorb voltage, where inverter/charger's absorb voltage regulation is slightly below BMS's absorb voltage calibration so BMS never thinks absorb voltage is reached.

Biggest issue in my opinion is a random inverter load demand dropping battery voltage just before BMS's absorb timer calls for a voltage check. If BMS finds battery voltage below absorb level it sends BMS back to beginning of charge cycle again with another full absorb time. It could do this indefinitely.

I think an improvement would be when absorb timer times out and absorb voltage is still a bit below RCV, give it two or three more chances to achieve RCV over the next fifteen to twenty minutes with a recheck for RCV being achieved every five minutes.

There also needs to be a safety abort in the number of successive times it sends the algorithm back to a full absorb timer sequence which can be damaging to battery.

 

[TechSree]

'Stuck in Bulk' is really a lack of achieving absorb voltage level on battery, as far as BMS observes.

Most likely cause is calibration difference between BMS absorb voltage and inverter/charger absorb voltage, where inverter/charger's absorb voltage regulation is slightly below BMS's absorb voltage calibration so BMS never thinks absorb voltage is reached.

Biggest issue in my opinion is a random inverter load demand dropping battery voltage just before BMS's absorb timer calls for a voltage check. If BMS finds battery voltage below absorb level it sends BMS back to beginning of charge cycle again with another full absorb time. It could do this indefinitely.

I have verified that and used the following settings
RCV - 3.460, SOC 100% - 3.440 and not connected load on inverter but still its not going to absorption but when changing the RCV to 3.450 and then back to 3.460 then gain change to 3.450V will trigger the absorption.

 

[Steve_S]

I have verified that and used the following settings
RCV - 3.460, SOC 100% - 3.440 and not connected load on inverter but still its not going to absorption but when changing the RCV to 3.450 and then back to 3.460 then gain change to 3.450V will trigger the absorption.

You are overdriving the cells a bit - the Working Voltage is 3.000-3.400 and LFP will "always" settle due to the chemistry and is normal & expected. Adjust the values down "Just a bit" and you will be rewarded with well balanced fully charged & saturated cells within each pack.

All my packs reach 100% by the time Absorb transitions to Float and within 15-20 Minutes of entering float, All Cells in All Packs are within 0.005mv of each other across the bank (6 packs in parallel).

Solar Controller Notes:
ABSORB is set to 27.8V / 55.6V (3.475 vpc) for a Max of 1 Hour.
- rtpically only runs 30 Minutes before EndAmps is reached, at which point it transitions to Float.
FLOAT is set to 27.7 / 55.4V (3.4625 vpc)
- Trickles in for balancing till BMS' go into REST Mode as Float can service ongoing loads while solar is sufficient.
ENDAMPS/TAILCURRENT for 280AH Packs = 14A. This is calculated as 100AH X 0.05 = 5A or 280AH X 0.05 = 14A
- Use the largest pack in a bank for calculating the EndAmps/TailCurrent.

Settings from the PC-APP. The BMS' are all V15 Hardware 200A models with 2A Active Balancing.
Balancing NOTE: Do not start too early, it is self-defeating. Once the 1st cell hits 3.42 it begins to balance out the cells and keeping the cells within the spec'd range.

* My bank has V1 (2018) through to V3 (2024) Cells in the packs (All EVE LF-280) and a mix of Bulk, B and A Grade cells and this works across ALL of them. It also worked with my previous setup with 1x105AH, 2x175AH, 3x280AH in one Parallel Bank.

 

[batmax]

what is the official link of jk bms ?

 

[dmsims]

what is the official link of jk bms ?

I beleive https://jikongbms.com/

 

[AshleyL]

I have verified that and used the following settings
RCV - 3.460, SOC 100% - 3.440 and not connected load on inverter but still its not going to absorption but when changing the RCV to 3.450 and then back to 3.460 then gain change to 3.450V will trigger the absorption.

First.......any difference in voltage at BMS and the inverter itself?
Sound to me, you didn't calibrate the voltage between the BMS and inverter.
Triple check again.
When the inverter reports 55.36v (RCV 3.460V) at the end of the charging, what is the JKBMS voltage reading? Calibrate the BMS voltage to match the inverter voltage reading at the end of the charging.
I got a feeling your inverter is reading 55.36v, but your BMS is reading 55.2v to 55.3v instead and hence it will never trigger the RCV Timer.


Second, your SOC100%volt is too close to the RCV value, make sure there is a spread of 0.040v between RCV and SOC100%volt. Any sudden fluctuation in voltage due to load might causes RCV Time to be extended.

Third, you mentioned there are two batteries in parallel. Ensure the second batttery (not the master) has lower RCV time than master battery. Example, if you set RCV Timer of 30 minutes at Master pack, you must set 24 minutes at slave battery to avoid race condition.

 

[TechSree]

First.......any difference in voltage at BMS and the inverter itself?
Sound to me, you didn't calibrate the voltage between the BMS and inverter.
Triple check again.
When the inverter reports 55.36v (RCV 3.460V) at the end of the charging, what is the JKBMS voltage reading? Calibrate the BMS voltage to match the inverter voltage reading at the end of the charging.
I got a feeling your inverter is reading 55.36v, but your BMS is reading 55.2v to 55.3v instead and hence it will never trigger the RCV Timer.


Second, your SOC100%volt is too close to the RCV value, make sure there is a spread of 0.040v between RCV and SOC100%volt. Any sudden fluctuation in voltage due to load might causes RCV Time to be extended.

Third, you mentioned there are two batteries in parallel. Ensure the second batttery (not the master) has lower RCV time than master battery. Example, if you set RCV Timer of 30 minutes at Master pack, you must set 24 minutes at slave battery to avoid race condition.

Yesterday after reading post from @RCinFLA I have calibrated the bms since there was a 7 mV difference with inverter reading, today its working fine and working as expected. I didnt given more proper attention to the 7mV difference since all cells are above 100% soc Volt but it looks like we need to keep it at RCV volt. Thanks @AshleyL ,@RCinFLA for pointing out this and special thanks to @Steve_S and all other members those who helped to solve the issue.

​

​

 

[TechSree]

You are overdriving the cells a bit - the Working Voltage is 3.000-3.400 and LFP will "always" settle due to the chemistry and is normal & expected. Adjust the values down "Just a bit" and you will be rewarded with well balanced fully charged & saturated cells within each pack.

All my packs reach 100% by the time Absorb transitions to Float and within 15-20 Minutes of entering float, All Cells in All Packs are within 0.005mv of each other across the bank (6 packs in parallel).

Solar Controller Notes:
ABSORB is set to 27.8V / 55.6V (3.475 vpc) for a Max of 1 Hour.
- rtpically only runs 30 Minutes before EndAmps is reached, at which point it transitions to Float.
FLOAT is set to 27.7 / 55.4V (3.4625 vpc)
- Trickles in for balancing till BMS' go into REST Mode as Float can service ongoing loads while solar is sufficient.
ENDAMPS/TAILCURRENT for 280AH Packs = 14A. This is calculated as 100AH X 0.05 = 5A or 280AH X 0.05 = 14A
- Use the largest pack in a bank for calculating the EndAmps/TailCurrent.

Settings from the PC-APP. The BMS' are all V15 Hardware 200A models with 2A Active Balancing.
Balancing NOTE: Do not start too early, it is self-defeating. Once the 1st cell hits 3.42 it begins to balance out the cells and keeping the cells within the spec'd range.

* My bank has V1 (2018) through to V3 (2024) Cells in the packs (All EVE LF-280) and a mix of Bulk, B and A Grade cells and this works across ALL of them. It also worked with my previous setup with 1x105AH, 2x175AH, 3x280AH in one Parallel Bank.

View attachment 289665

View attachment 289666

I think float value(RFV) of 3.4625V is a bit high since the resting volt for LFP seems to be around 3.375V. Using 54V(3.380V) as float. using RCV 3.450V as 100% soc volt and changed to 3.440V to check wether its going to Absorption or not but now its seems to be working after doing the calibration.
My Current settings are RCV 3.460, 100% Soc 3.440V, RFV 3.380V, RCV Time 1 hour, RFV Time - 6hours
May I know why you are recomending higher Float value of 3.465?.

Thanks for your help

 

[Steve_S]

I think float value(RFV) of 3.4625V is a bit high since the resting volt for LFP seems to be around 3.375V. Using 54V(3.380V) as float. using RCV 3.450V as 100% soc volt and changed to 3.440V to check wether its going to Absorption or not but now its seems to be working after doing the calibration.
My Current settings are RCV 3.460, 100% Soc 3.440V, RFV 3.380V, RCV Time 1 hour, RFV Time - 6hours
May I know why you are recomending higher Float value of 3.465?.

Thanks for your help

LFP's "Working Voltage Range" is 3.000-3.400 and most will use from 2.800-3.400 to allow for a buffer against Voltage SAG's when a load starts.
I float at 3.465 to allow for full saturation (deep charge) while keeping the packs cells all balanced. As soon as the system starts pulling juice from the batteries (as float decreases and can no longer provide "running current") the packs do settle and usually sit at 3.405/3.410 for a little while as by then they are drawing down. Do remember that FLOAT is Constant Voltage - Variable Current and will trickled down to 0.001A or even 0A input when the batteries are truly at 100% and they cannot take anymore juice and go into REST mode. If there is no output demand at all and no input from any source, the cells will drop to 3.410 +/- 0.005 and sit there for hours from that float voltage.
* VPC = Volts Per Cell

With Grade-A Matched cells this is very consistent and in fact, for the most part the cells all stay at or below 0.005mv differential until they start to get to 3.100 +/- 0.050v VPC

With B-Grade (just voltage matched - not Matched & Batched) cells, the deviations are typically around 0.010-0.015 as soon as 3.150 VPC is reached. With BULK Grade that differential get's up to 0.025-0.030 by 3.100

I always reference manufacturers specs & guidelines and stay within spec, including new cell initialization rules and do not take short cuts. I have a LOT of kWh of Battery Packs out in circulation and only a few of the early ones had minor issues.

 

[samir101ng]

Seems there is a new fireware V15.39 on Andy's Drive. Any change logs or anybody tested the new firmware ? The fact that new firmwares are still coming out despite a new V19 BMS board means that at least certain improvements are being made. Might be the last though as attention shifts to the new kid on the block.

 

[ChrisG]

What RS485 to USB (amazon quick purchase hopefully) do you all use to update the JK Firmware?