Seplos bms jumping SoC?

[Dorokun192]

I would like to ask those who are currently using seplos bms v2 16S 200A. I bought mine locally here in the Philippines and this is my 2nd time that I've installed this BMS. The issue is that there are instances where the SoC of the battery pack jumps from 22% all the way up to 99% in a matter of seconds as well as during discharge it would jump from 45% to 7%. I checked the voltage of each cell and they are all equal up to 3.21xV. I also tried to calibrate the SoC via full charge and discharge but I can't seem to do it due to this issue. I've attached some screenshots from my Deye app showing the SoC of the battery (green) in reference to the power generation (blue) as well as consumption (red). How do I resolve this issue?

 

[houseofancients]

Seems your cells grossly out of balance..
BMS will jump to 100% when an ov protection is hit.

1 Top balance all your cells ( see resource section for a how to)
2 double and triple check the BMS lead connections to the cells

 

[impactkaiser20]

Hello, did you solve the problem? I have the same problem here in the Philippines

 

[ddonq1412]

Same issue. Trying to top balancing can report back in about 9 days.

 

[ddonq1412]

Did anyone get this sorted

 

[houseofancients]

Did anyone get this sorted

Are your cell top balanced already ?

 

[ddonq1412]

I spent over a week charging the cells individually and then putting in in blocks of 4 and getting them over 3.6v and it drained well but then start flip flopping again.

 

[DiySolarU]

I have the same problem with 2 out of 4 packs in parallel with the same BMS. My cells are very well balanced. The SOC jumps from 95% to 10.7% within seconds, but the cell voltage is at 3.33 volts and the cell difference is only about 3mV.

Is there any solution to resolve the Problem?

 

[ddonq1412]

Charge to 100% grab reading of cells. See if they close to 3.4 after charge ending. Disconnect the Comms and try a drain of battery see what is drains. The inverter might have to go to lead acid mode to stop the errors. Failing that strip down and manually charge the cells

 

[DiySolarU]

The cells are not the problem. They are all between the same cell voltage. Only 3-4mV difference. I have 4 packs in parallel and the problem is on 2 of them at the moment. The cells are also always balanced when the cell are over 3.40 volts with a Neey active 4A balancer to 2mV. I load the Pack up to 55.20 Volt (3.45 Volt/cell). I use 2 x DEYE 8kW Hybrid Inverter in parallel with Voltage Regulation to 55.20 Volt (no CAN). Works perfect for me.




I use EVE 304Ah Cell A-Class perfect TOP Balanced. I measured each cell individually with 150A discharge current before I built the battery (EEL battery). I think it must be a firmware BUG. It is physically not possible that the SOC drops from over 90% to under 11% within seconds when the cell voltages are all at 3.3 volts...

 

[ddonq1412]

That is slightly different have you tried disconnecting them and run as individual battery see if issue remains. Have you got home assistant or look at monitor logs to see what each cell does it and under what load.

 

[houseofancients]

The cells are not the problem. They are all between the same cell voltage. Only 3-4mV difference. I have 4 packs in parallel and the problem is on 2 of them at the moment. The cells are also always balanced when the cell are over 3.40 volts with a Neey active 4A balancer to 2mV. I load the Pack up to 55.20 Volt (3.45 Volt/cell). I use 2 x DEYE 8kW Hybrid Inverter in parallel with Voltage Regulation to 55.20 Volt (no CAN). Works perfect for me.

View attachment 189288


I use EVE 304Ah Cell A-Class perfect TOP Balanced. I measured each cell individually with 150A discharge current before I built the battery (EEL battery). I think it must be a firmware BUG. It is physically not possible that the SOC drops from over 90% to under 11% within seconds when the cell voltages are all at 3.3 volts...



View attachment 189279

check the connections of your voltage sensor leads

 

[DiySolarU]

I think i will write my own Software to log all the Data in a Database (one Sample per Sec). Then i will see what happens...

 

[DiySolarU]

@houseofancients

When I built the ELL battery, I saw that the wires for the voltage measurement are in a very bad condition. I replaced them all with 2.5mm flexible wire. Each screw is correctly tightened with the torque wrench and fixed with lacquer. This means, it can not be the problem in 2 Units at the same Time. I am not very happy with the Seplos BMS.

Currents below about 1A are not registered at all and are not taken into account in the SOC calculation. The two Deye inverters draw around 150 watts (approx. 2.88 A) from the 4 batteries when idling (e.g. at night). However, the 4 Seplos BMS show 0 A standby. The PC software is very poorly made. And the Bluetooth connection keeps breaking off or doesn't connect at all. I will never again build a battery with a SEPLOS BMS...

 

[houseofancients]

@houseofancients

When I built the ELL battery,

so this isnt a Seplos box ?
maybe PCB's are broken ?

I saw that the wires for the voltage measurement are in a very bad condition. I replaced them all with 2.5mm flexible wire. Each screw is correctly tightened with the torque wrench and fixed with lacquer. This means, it can not be the problem in 2 Units at the same Time.

or you made the same mistake on both boxes.
This bms is a smart bms, and the leads are also used for resistance calculations, which you may have screwed by soldering new wires on..

I am not very happy with the Seplos BMS.

i think you will not be happy with any server rack based bms, but look forward to your evaluation of others

Currents below about 1A are not registered at all and are not taken into account in the SOC calculation.

this is the case in ALL server rack bms'es ( as far as i am aware of)

The two Deye inverters draw around 150 watts (approx. 2.88 A) from the 4 batteries when idling (e.g. at night). However, the 4 Seplos BMS show 0 A standby.

same here

The PC software is very poorly made.

agreed

And the Bluetooth connection keeps breaking off or doesn't connect at all.

never had an issue with that, on any of the total of 38 boxes i have build , however you need to be close to the boxes as they are faraday cages

I will never again build a battery with a SEPLOS BMS...

up to you

View attachment 189291

 

[DiySolarU]

I have built batteries with DALY BMS and JK-BMS. Both can detect currents below 1A without problems and perform very accurate measurements and calculation of the SOC... I have been using a 250A DALY BMS with RS-485 interface on my own DIY battery for 3.5 years. It works without any problems.

 

[ddonq1412]

seplos v3 might have given you different results. you could try and flash the bms or check to see if that are all using the same software version.

 

[DiySolarU]

Because the SEPLOS software is so bad, I have developed my own dashboard for my 4 batteries. I read all the data from the 4 BMS every second and write it to a SQLite database. This allows me to analyze what happens when the SOC of the BMS flips around. I do my own SOC calculation in the dashboard Software.



All this data from the 4 SEPLOS 2.0 BMS is recorded every second in an SQLite database so that it can be analyzed later.

 

[Quattrohead]

The battery pack must hit the OVP before the BMS will set to 100% SOC.
I have 4 packs in parallel, 3 of them I got to OVP and they set to 100%, the other has a runner so it would not reset to 100%
You can cheat a little by setting the relevant OVP slightly lower. The default settings are rubbish.
I think this is how these work, check on Andy's off grid garage. He spends a lot of time figuring these things out.

 

[DiySolarU]

@Quattrohead
Yes, I know that. The problem is that when the battery is still full (approx. 60%, approx. 3.3x volts per cell), it can happen that the SOC drops to approx. 10.7%. Without a heavy load. I'll see what the voltage of each cell will be when it fails. I'm logging the data now and after a few days I'll do a redout and see what happens. If I find out anything I will post it.