What's the best way to diagnose/repair this cell (green line) that does not want to come to the party. Its one of three Sunsynk batteries. I've been keeping them fully charged to try correct the unbalance. Is it worth removing the other two batteries and charging this on its own and then cycling?

I noticed this because my batteries would force charge from the grid early in the morning and you can see the reason below when that cell voltage drops to 3V. My initial thoughts are it must be faulty and I have no choice but to book the battery in for repairs and wait the 6 weeks?

You can see how the SOC rapidly dops and increases with the grid charge

25 minutes ago, slipx said:

What's the best way to diagnose/repair this cell (green line) that does not want to come to the party. Its one of three Sunsynk batteries. I've been keeping them fully charged to try correct the unbalance. Is it worth removing the other two batteries and charging this on its own and then cycling?

I noticed this because my batteries would force charge from the grid early in the morning and you can see the reason below when that cell voltage drops to 3V. My initial thoughts are it must be faulty and I have no choice but to book the battery in for repairs and wait the 6 weeks?

You can see how the SOC rapidly dops and increases with the grid charge

You could try removing the other 2 but lithium batteries at different SOC do balance when charging. This seems like a bad cell as your active balancer should bring it higher if it would allow further charge.

@slipx ,

Taking a wild guess here. IF the Sunsynk batteries use a Pace BMS (saying so purely by the way the balancing1 & 2 is displayed), then I would suggest one needs to lower the cell voltage threshold when balancing kicks in, By default, it is 3.45V, so the green cell never gets there, so it doesn't get a chance to balance.

Hope my logic is correct.

Easiest way to check (PACE) BMS is to see if PBMStools will populate the screen with RS232/USB cable, then you can set the balance voltage.

I have access to PBMS Tools and yes they do use Pace. I'll post a screenshot and maybe you can help guide me

@slipx , Check the Offgrid Garage Youtube https://www.youtube.com/watch?v=fT40MtzaaLQ

He shows the procedure from 16:00 onwards.

Checked PBMS Tools and the balance threshold is set to 3.4v and the balance cell voltage delta is 30 (less than my current reading of 42). I'm assuming it will use a combination of these two settings to balance.

Set the balance threshold to 3.32 and now all the cells are balancing except the "weak" cell. Will see how it goes overnight.

Edited July 18, 2025Jul 18 by slipx

3 hours ago, Scorp007 said:

You could try removing the other 2 but lithium batteries at different SOC do balance when charging. This seems like a bad cell as your active balancer should bring it higher if it would allow further charge.

I don't think PACE makes a BMS with active balancing in the 100A verianty. I am open to being corrected though, or perhaps this battery has an bolt on active balancer.

A passive balancer won't be able to bring that cell back in line.

There is no active balancing on the BMS and I was not able to correct that cell even after lowering the balance threshold.

Apart from the weak cell 14 on battery 2, looking at all the batteries over a 24 hour window I've noticed that cell 1 on the third battery is also always below the others. Its this something to be concerned about and is it a sign of a weaker cell? Does not cause the battery 3 SOC to deviate much?

Well I have a similiar situation with one of my batteries where one cell was consistanly lower than the rest.

When the tech from the battery company came out to look at my batteries for a warranty claim he said that they will need to rebalance that cell for me and it is not something that can be done on site. I'm still waiting on the batteries to be returned though so not sure what the final outcome was when they took it back. My other battery had 4 bad cells that needed to be replaced.

The green line is the SOC of the above battery over 24hrs and according to the tech that one cell could account for that odd SOC behaviour at the middle suddenly first running higher then dipping below the others. The Red one is the one with 4 bad cells that was running on reduces capacity already.

Not sure if this is helpful but will confirm final outcome once I get my batteries back.

I had a similar issue with my newest purchased Leoch 48100-S pack. From the beginning cell 15 showed significantly higher voltage at charge and lower at discharge. This being a brand-new pack I contacted the official importer and seller, Averge, with a guarantee claim. But they didn't want to do anything about it. I asked to allow me to open the pack without voiding the guarantee to check the cells connections. They wrote me the permission and so I did. I tightened all M4 hexagon bolts of the cell bridges with a socket wrench. Some allowed up to a half turn. After reassembly and re-connection the issue with cell 15 no longer persists. It looks like all cells run the same voltage within a few mV. This proves to me that all cells are in good health.

On 2025/07/20 at 11:14 AM, slipx said:

Apart from the weak cell 14 on battery 2, looking at all the batteries over a 24 hour window I've noticed that cell 1 on the third battery is also always below the others. Its this something to be concerned about and is it a sign of a weaker cell? Does not cause the battery 3 SOC to deviate much?

A 40mV difference in the flat part of the curve is rather substantial. The graph is not clear enough to see what happens in the "knee" of the charge curve (above 3.45v).

If it was me I would have it looked at. To what voltage do you charge to? Does each cell reach at least 3.5v ever so often during charging?

Edited July 21, 2025Jul 21 by I84RiS

Battery 2 was sent in for repair. When it comes back I need to sort out battery 3. Cell 1 is not holding voltage.

Edited July 24, 2025Jul 24 by slipx

@slipx what method/integration are you using to get the Sunsynk BMS cell metrics into HA?

Hi @Gambit

I'm running this python script on a raspberry pi https://github.com/Tertiush/bmspace/tree/main/pace-bms-dev

5 hours ago, slipx said:

Hi @Gambit

I'm running this python script on a raspberry pi https://github.com/Tertiush/bmspace/tree/main/pace-bms-dev

Thanks. Was also looking to go this route but was hoping to find an esphome external component. Was also contemplating writing my own.

2 hours ago, Gambit said:

was hoping to find an esphome external component

Here you go
https://github.com/syssi/esphome-pace-bms
I tested this with my pacebms and esp32 works perfectly.

11 hours ago, WannabeSolarSparky said:

Here you go
https://github.com/syssi/esphome-pace-bms
I tested this with my pacebms and esp32 works perfectly.

Thanks so much. I found this one a while back, but it is for the RS485 Modbus interface. Even though the Sunsynk batteries have RS485 ports, I was under the impression that the BMS metrics were only available over the RS232 interface on these batteries. I see there is a GitHub issue with more information on this, so I will have another look at this.

10 hours ago, Gambit said:

the RS232 interface

@Gambit
the rs232 on most bms's is usually for configuration software like pbms etc that connects to your pc/laptop
rs485 is usually for the monitoring software and inter battery comms etc
If your battery has a pacebms/jkbms/seplosbms etc then definitely the rs485 port will work for most monitoring systems

On 2025/07/28 at 7:30 PM, WannabeSolarSparky said:

@Gambit
the rs232 on most bms's is usually for configuration software like pbms etc that connects to your pc/laptop
rs485 is usually for the monitoring software and inter battery comms etc
If your battery has a pacebms/jkbms/seplosbms etc then definitely the rs485 port will work for most monitoring systems

@WannabeSolarSparky Thanks for the advice. I managed to get all the BMS metrics from my 3 x Sunsynk 5.32 (SUN-BATT-5.32) batteries using a RS-485 transceiver on an ESP32 with ESPHome, but it turned out to be a bit of a mission to create the frankenstein RS485 cable to link all 3 batteries and the Sunsynk inverter. Unfortunately, you can't get the BMS metrics directly from the master battery like you can over RS232 (or I didn't find an easy way to do that) and instead have to setup an RS485 bus between all the batteries to connect to query each battery BMS directly. What makes it tricky with the Sunsynk batteries is that they use RJ45 sockets, which also share the CAN bus so I had to crimp a weird cable with 5 x RJ45 plugs and split off the CAN connection to the Inverter. Super happy it all worked first time but I'm wondering if there was an easier way to cable it all up.

I now have a single ESP32 microcontroller powered by the inverter's 12V rail pulling metrics from 3 batteries and 1 inverter and pushing to Home Assistant.

4 hours ago, Gambit said:

What makes it tricky with the Sunsynk batteries is that they use RJ45 sockets, which also share the CAN bus so I had to crimp a weird cable with 5 x RJ45 plugs and split off the CAN connection to the Inverter

Hey if the diy works then that's great :)
Usually you can cascade all the batteries together.
usually there are 2 rj45 ports used for rs485 to link all the batteries and then there are usually also rj45 can port to link to inverter.
but sometimes the diy way is more satisfying anyway if it works :)

On 2025/07/20 at 11:14 AM, slipx said:

Apart from the weak cell 14 on battery 2, looking at all the batteries over a 24 hour window I've noticed that cell 1 on the third battery is also always below the others. Its this something to be concerned about and is it a sign of a weaker cell? Does not cause the battery 3 SOC to deviate much?

@slipx would you mind perhaps sharing that plotly card config you use, I'd like to try and setup something similar now that I have all my batteries back up and running instead of going through the history screen and selecting each cell every time.

Thanks

@-cK- Sure. Here you go

type: custom:plotly-graph
title: Battery 1 Cell Voltage
hours_to_show: 24
ha_theme: true
refresh_interval: 10
entities:
  - entity: sensor.battery_pack_1_cell_01_voltage
    name: 1
  - entity: sensor.battery_pack_1_cell_02_voltage
    name: 2
  - entity: sensor.battery_pack_1_cell_03_voltage
    name: 3
  - entity: sensor.battery_pack_1_cell_04_voltage
    name: 4
  - entity: sensor.battery_pack_1_cell_05_voltage
    name: 5
  - entity: sensor.battery_pack_1_cell_06_voltage
    name: 6
  - entity: sensor.battery_pack_1_cell_07_voltage
    name: 7
  - entity: sensor.battery_pack_1_cell_08_voltage
    name: 8
  - entity: sensor.battery_pack_1_cell_09_voltage
    name: 9
  - entity: sensor.battery_pack_1_cell_10_voltage
    name: 10
  - entity: sensor.battery_pack_1_cell_11_voltage
    name: 11
  - entity: sensor.battery_pack_1_cell_12_voltage
    name: 12
  - entity: sensor.battery_pack_1_cell_13_voltage
    name: 13
  - entity: sensor.battery_pack_1_cell_14_voltage
    name: 14
  - entity: sensor.battery_pack_1_cell_15_voltage
    name: 15
  - entity: sensor.battery_pack_1_cell_16_voltage
    name: 16
layout:
  showlegend: true
  hovermode: x unified
  height: 500px
  yaxis:
    tickmode: auto
    nticks: 20
    fixedrange: true
  legend:
    bgcolor: rgba(0,0,0,0)
    itemsizing: constant
    font:
      size: 11
defaults:
  entity:
    show_value: false
    hovertemplate: "%{y}"
    line:
      width: 1.5