• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

YamBMS JK-BMS-CAN with new Cut-Off Charging Logic (open-source)

Sleeper85

Sleeper85

Open Solar Energy
Joined
Nov 28, 2022
Messages
1,175
Location
Belgium
ESPHome application to monitor BMS and communicate with inverters supporting CAN bus protocol compatible with Pylontech, GoodWe, SMA, Victron or Luxpower (EG4).

It's with pleasure that I announce the new version of the YamBMS open-source project with the new Cut-Off Charging Logic.

Thanks to @shvm for creating the Cut-Off Charging Logic equations.

Thanks to @MrPablo for programming the Auto Charge/Discharge Current Control and Auto Charge Voltage Control functions to avoid OVP/UVP alarms.

Thanks to @arzaman for testing with Atom S3.

The old JK-BMS-CAN version with single BMS is still available here.

I hope you like it ;)


YamBMS ( Yet another multi-BMS Merging Solution )

Please => README <=

YamBMS.png

YamBMS is a multi-bms and multi-shunt application working on a single ESP32.

With the new JK-PB series only one RS485 connection is required with max 16 BMS. (15 BMS in mode2)

With the old JK-B series all BMS are connected to the ESP32.

An ESP32 supports max 3 UARTs and 3 BLE connections, it can be a mix of the two. I haven't yet been able to carry out many tests on my own but it is certain that the more BMS there are to connect, the more the limits of the classic ESP32 will be reached. ESP32-S3 or Atom S3 are preferred but I think a classic ESP32 should work well with 2 or 3 BMS.

The PVbrain2 PCB with ESP32-S3 supports 11 UARTs and 4 additional UARTs can be added with an expansion card.

Please also note that the web server is a component that uses a lot of memory and can have a negative impact (crash and reboot).


buymeacoffee.png
 
Last edited:
Thanks @Sleeper85! Enjoy your trip!

Regarding the new Auto Charge/Discharge Current Control, this feature has been introduced to hopefully prevent OVP and UVP alarms.
At a high level, the ESP will check the maximum and minimum cell voltage every refresh cycle, compare it to a BMS threshold and automatically reduce current as needed. This should reduce voltage spikes and drops.

As with most things though, the actual implementation is more complex than that!

In the YAML, there are two configuration settings, one each for Charge and Discharge control.
The values within are commented and should be fine at the default values.
2 new switches have been added, 'Automatic (Dis)Charge Current'. When disabled, maximum allowable current will be as per the existing sliders, otherwise the new logic below will be used.

First, an instantaneous allowable current is calculated based on maximum / minimum cell voltage, referencing the new parameters, as well as BMS values for cell over / under voltage release.
I've used the release values as I'd prefer to leave a buffer before a cell could hit OVP or UVP.

A moving average is then used to smooth transitions between calculated values.
After that, another sensor will return the min of the instantaneous and moving average. This means that the allowable current will always drop quickly, but rise over several seconds. This stops voltage oscillations as current values jump around.

The code used for calculating the automatic current values isn't that intuitive, so I have a link to a google sheet which will show you the impact of changing parameter values.

esphome-jk-bms-can Current Control
 
Last edited:
Enjoy your trip @Sleepers
thanks MyPAblo and the mathematicians for the hard work

i was just yesterday fighting the OVP where my system was tripping the alarms and i had to reflash to this low to get it to stop complaining
even themn it still complained a bit until i further reduced it to 15
1708681157780.png
the JK was set to 35 amp....and it seemed to ignore both its settings and the JK settings

hopefully this is fixed
and also on the BLE version as i cant get the Wire working still after trying every thing...i am beginging to think ist the BMS version.
 
chaosnature said:
Enjoy your trip @Sleepers
thanks MyPAblo and the mathematicians for the hard work

i was just yesterday fighting the OVP where my system was tripping the alarms and i had to reflash to this low to get it to stop complaining
even themn it still complained a bit until i further reduced it to 15
View attachment 197749
the JK was set to 35 amp....and it seemed to ignore both its settings and the JK settings

hopefully this is fixed
and also on the BLE version as i cant get the Wire working still after trying every thing...i am beginging to think ist the BMS version.
Click to expand...
@chaosnature, you're using a Solis inverter aren't you? What model do you have?

The auto charge current control should prevent OVP, but I know my Solis didn't always respect the requested charge voltage.
 
will the -HC jkbms version that has the CAN protocol work in parallel

I see you guys focusing on the rs458 conversion - writing the inverter protocol for it...what about writing the inver protocol for the CAN port so that the rs489 is free for paralleling?

regards
 
chaosnature said:
will the -HC jkbms version that has the CAN protocol work in parallel

I see you guys focusing on the rs458 conversion - writing the inverter protocol for it...what about writing the inver protocol for the CAN port so that the rs489 is free for paralleling?

regards
Click to expand...
We don't actually use the CAN port on the JK BMS, we only utilise the GPS / UART port.
A likely solution for multi-BMS comms will be to use a standard JK BMS for each battery, each with an ESP32 adaptor board which connect to each other via the CANbus.

Again, there's no timeline on this - implementing dynamic CVL and external shunt support is first on the list as it's easier to test and debug.
 
While we wait for the Multi setup script - this is what i have decided
my BMS is 24s, i have used up 14s from the GBLI5001 leaving 10s to use
that 10s i will spread it amongst the other battery 10s leaving 4s unbalanced

however - the maths is what i cannot do if i continue the series loop i will end up with 48v + 48v - (3.7 x4+) = 81v cells being balanced and 96 total coming out
1708705898013.png

how do i wire this up so that it stays 48v
only thing i can think of is parallel the cells that will be too many wires.
 
chaosnature said:
While we wait for the Multi setup script - this is what i have decided
my BMS is 24s, i have used up 14s from the GBLI5001 leaving 10s to use
that 10s i will spread it amongst the other battery 10s leaving 4s unbalanced

however - the maths is what i cannot do if i continue the series loop i will end up with 48v + 48v - (3.7 x4+) = 81v cells being balanced and 96 total coming out
View attachment 197791

how do i wire this up so that it stays 48v
only thing i can think of is parallel the cells that will be too many wires.
Click to expand...
I don't think you can add additional cells for balancing and monitoring without increasing the total voltage. As you say, that's not something you can do with the Solis inverter.

I'd suggest buying another JK BMS that matches your current one, setting up the second battery with that and you can potentially be the tester for multi BMS logic when I develop it.
 
let's start testing :)
JK2A24S15p x 2

as long as the logic uses the BLE as i did not succeed with the Wire lol
 
chaosnature said:
let's start testing :)
JK2A24S15p x 2

as long as the logic uses the BLE as i did not succeed with the Wire lol
Click to expand...
Ok, but it's quite probable that code won't be available until mid-March - dynamic CVL and external.shunt support is next up.

Regarding the wire version, have you confirmed that you're getting a signal on the wires themselves? You could use a multimeter between TTL ground and the RX then TX lines, just to confirm voltage, etc.

I don't know of anyone else who is getting zero data via the UART TTL port, so I'm pretty sure it's a connection issue.
 
MrPablo said:
Ok, but it's quite probable that code won't be available until mid-March - dynamic CVL and external.shunt support is next up.

Regarding the wire version, have you confirmed that you're getting a signal on the wires themselves? You could use a multimeter between TTL ground and the RX then TX lines, just to confirm voltage, etc.

I don't know of anyone else who is getting zero data via the UART TTL port, so I'm pretty sure it's a connection issue.
Click to expand...
I decided to Test Both BLE and Wire
Result below:

CAN-WIRE - When connected to JKBMS Voltage between:
rx2 (GPIO16) and tx2(GPIO17) = 0.01 (0.5v with broken wire)
rx (GPIO1) and tx(GPIO3) = 0.01 (0.5v with broken wire)
Ground and tx or rx = 3.26
# BMS Light Blinking


CAN-WIRE - When Disconnected to JKBM Voltage between - SAME Result
rx2 (GPIO16) and tx2(GPIO17) = 1.04 (0.5v with broken wire)
rx (GPIO1) and tx(GPIO3) = 0.01 (0.5v with broken wire)
Ground and tx = 3.26
Ground and rx = 2.13


To compare i did:
CAN-BLE - When connected to JKBMS Voltage between:
rx2 (GPIO16) and tx2(GPIO17) = 0v
rx (GPIO1) and tx(GPIO3) = 0.5v
# BMS Light Stable on - Can hear beep when it connects to BMS

CAN-BLE - When Disconnected to JKBMS Voltage between - SAME Result
rx2 (GPIO16) and tx2(GPIO17) = 0v
rx (GPIO1) and tx(GPIO3) = 0.5v


i then checked :
Tx and Rx from the JKBMS Voltage between:
Ground and tx = 3.29 after checking wires (0.0 Wire was broken)
Rx and T = 0.0

1708733213420.png
1708733239001.png



ESPHOMe show online
1708733593500.png
1708733857793.png
 
Last edited:
chaosnature said:
I decided to Test Both BLE and Wire
Result below:

CAN-WIRE - When connected to JKBMS Voltage between:
rx2 (GPIO16) and tx2(GPIO17) = 0.01 (0.5v with broken wire)
rx (GPIO1) and tx(GPIO3) = 0.01 (0.5v with broken wire)
Ground and tx or rx = 3.26
# BMS Light Blinking


CAN-WIRE - When Disconnected to JKBM Voltage between - SAME Result
rx2 (GPIO16) and tx2(GPIO17) = 1.04 (0.5v with broken wire)
rx (GPIO1) and tx(GPIO3) = 0.01 (0.5v with broken wire)
Ground and tx or rx = 3.26


To compare i did:
CAN-BLE - When connected to JKBMS Voltage between:
rx2 (GPIO16) and tx2(GPIO17) = 0v
rx (GPIO1) and tx(GPIO3) = 0.5v
# BMS Light Stable on - Can hear beep when it connects to BMS

CAN-BLE - When Disconnected to JKBMS Voltage between - SAME Result
rx2 (GPIO16) and tx2(GPIO17) = 0v
rx (GPIO1) and tx(GPIO3) = 0.5v


i then checked :
Tx and Rx from the JKBMS Voltage between:
Ground and rx = 3.36
Ground and tx = 3.29 after checking wires (0.0 Wire was broken)
Rx and T = 0.0

View attachment 197896
View attachment 197897



ESPHOMe show online
View attachment 197898
View attachment 197902
Click to expand...
1708733946009.png
 
Also, I'd like to add my two cents to this amazing project.

I've prepared a summary connection diagram for those who are going to use the M5Stack—a nice product that simplifies the wiring and connections, making the hardware part almost plug and play.
Both configurations have been tested and are stable

1708805763061.png
 
arzaman said:
Also, I'd like to add my two cents to this amazing project.

I've prepared a summary connection diagram for those who are going to use the M5Stack—a nice product that simplifies the wiring and connections, making the hardware part almost plug and play.
Both configurations have been tested and are stable

View attachment 198115
Click to expand...
1. Do you have a link to purchase these components
2. would one be able to use more than 1 battery set (i.e parallel 2 batteries) ?
3. which yaml code did you use for this , or this is not a ESP dev thing?
 
arzaman said:
Also, I'd like to add my two cents to this amazing project.

I've prepared a summary connection diagram for those who are going to use the M5Stack—a nice product that simplifies the wiring and connections, making the hardware part almost plug and play.
Both configurations have been tested and are stable

View attachment 198115
Click to expand...
@arzaman between the esp32 dev board and the M5 stack, agree M5 is extremely simple. Confirmed this works on my EG46000xp with 1.16.4.

Just ordered two more. Really want to get the Victron shunt reporting to CAN. From the other project I’m already reading SOC, V and current (A) on the standard ESP dev board from shunt.
IMG_4464.jpeg
 
ChrisG said:
@arzaman between the esp32 dev board and the M5 stack, agree M5 is extremely simple. Confirmed this works on my EG46000xp with 1.16.4.

Just ordered two more. Really want to get the Victron shunt reporting to CAN. From the other project I’m already reading SOC, V and current (A) on the standard ESP dev board from shunt.
Click to expand...
Do these M5stack adaptors have a 120ohm termination resistor with them? Just thinking ahead to a multi-BMS solution where there may be more than 2 BMS to CAN adaptors.
 
You must log in or register to reply here.

Similar threads

P
JK-PB2A16S-20P BMS overloading at inverter startup
Replies
6
Views
521
HRTKD
HRTKD
A
How to read Current and Charge State(SOC) from JK BMS to ESP32/Arduino ? preferably via Blue tooth connection.
Replies
13
Views
2K
DanDyse
D
S
Yet another JK BMS to CAN bus interface, this one Arduino, BLE to PylonLV CAN, on a Lilygo T-Display S3
Replies
5
Views
2K
akumd
A
C
How to Flash Atom S3 with EspHome - HA
Replies
23
Views
3K
MaikelK
M
R
a new uart protocol for a Daly Smart 150A BMS?
Replies
9
Views
2K
Tyrulf
T

diy solar

diy solar
Back
Top