diy solar

diy solar

JK BMS issues (suspect current sensor faulty) for 48V inverter setup

cjdell

New Member
Joined
Aug 31, 2022
Messages
17
Hello,

I've been running a DIY battery storage system (with the main intension of timeshifting grid consumption) for approximately 1 year with the following components:

- JK BMS B2A20S20P
- EVE 280Ah Cells (Grade B from Fogstar in the UK)
- Growatt SPF5000ES
- 3kWp solar (10 panels 1 string - not really relevant)
- ESP32 to convert native BMS protocol to Pylontech (thank you uksa007 for the code to do this)

I've had a number of niggles with the BMS that weren't showstoppers such as:

- Current sensor sometimes showing odd readings flipping between positive and negative values (apparently this is due to ripple from the inverter and the BMS not sampling frequently enough)
- SoC calculation going way off course unless I top up the cells all the way to 3.45V so it can recalibrate almost *every night*
- BMS suddenly realising at about 30% SoC (for example) that it's bottomed out the battery and then causing the inverter to panic. Annoyingly the inverter would cut output power when this happens (even though the grid is still available to switch to)

Today I just had an alarming notification that the BMS short circuit protection had triggered. Good news is there was no smoke. Bad news is the inverter wasn't able to take any power from the battery without re-triggering the short circuit protection. Has anyone else had this happen?

I thought at first my inverter had fried, but then I noticed that there wasn't any current drain registering on the BMS at all (literally 0 amps). Normally there would be at least 0.5A or so even with the inverter in standby.

So I'm thinking the JK BMS is faulty and perhaps the clues were there along. I've seen others report issues with the current sensor over the last year so I've completely lost confidence in JK as company. Interested in hearing thoughts from other JK BMS users...

Thinking of getting the Seplos Mason kit from Fogstar which gives me the Seplos BMS V3 along with a much nicer enclosure than the one I cobbled together:

Is this a good plan or is there still hope for my JK BMS?

I welcome your opinions!
 
I always thought that Seplos Mason looked like a cool package, just make sure you understand Seplos has an RS485 version and a CAN bus version (unless they might currently have a module that supports both methods of front-end communication), so you can get the right one for your Growatt's physical comm type (providing they speak the same language on the protocol too). @houseofancients knows a lot about this, with Seplos and RS485 or CAN.
 
Hello,

I've been running a DIY battery storage system (with the main intension of timeshifting grid consumption) for approximately 1 year with the following components:

- JK BMS B2A20S20P
- EVE 280Ah Cells (Grade B from Fogstar in the UK)
- Growatt SPF5000ES
- 3kWp solar (10 panels 1 string - not really relevant)
- ESP32 to convert native BMS protocol to Pylontech (thank you uksa007 for the code to do this)

I've had a number of niggles with the BMS that weren't showstoppers such as:

- Current sensor sometimes showing odd readings flipping between positive and negative values (apparently this is due to ripple from the inverter and the BMS not sampling frequently enough)
- SoC calculation going way off course unless I top up the cells all the way to 3.45V so it can recalibrate almost *every night*
- BMS suddenly realising at about 30% SoC (for example) that it's bottomed out the battery and then causing the inverter to panic. Annoyingly the inverter would cut output power when this happens (even though the grid is still available to switch to)

Today I just had an alarming notification that the BMS short circuit protection had triggered. Good news is there was no smoke. Bad news is the inverter wasn't able to take any power from the battery without re-triggering the short circuit protection. Has anyone else had this happen?

I thought at first my inverter had fried, but then I noticed that there wasn't any current drain registering on the BMS at all (literally 0 amps). Normally there would be at least 0.5A or so even with the inverter in standby.

So I'm thinking the JK BMS is faulty and perhaps the clues were there along. I've seen others report issues with the current sensor over the last year so I've completely lost confidence in JK as company. Interested in hearing thoughts from other JK BMS users...

Thinking of getting the Seplos Mason kit from Fogstar which gives me the Seplos BMS V3 along with a much nicer enclosure than the one I cobbled together:

Is this a good plan or is there still hope for my JK BMS?

I welcome your opinions!
I think in your case you would have 3 options

1. Buy another jk BMS and go forward
This will be your cheapest and easiest option

2. Buy a different BMS like the pace. Uses the same balance leads , but seems less prone to issues ( used by sok, jakiper, gobel and many others)

3. replace your BMS by a seplos one, and add a mason box that includes a 200a one.
Would double your capacity, but is also a huge added cost.

On a personal note, I'd make sure you order one with a V2 ( aka 10e ) seplos BMS.
Been around longer, less quirks, and IMHO a much much better BMS.
If you need active balancing ( which neither V3 or V2 natively has, both have passive balancer) , add a neey 4a active balancer if that is what you want/wish
Skip on the seplos balancer, as it is not configurable, thus will hurt you balance more that it will help, plus again, doesn't seem up to the task

Just my 2 cents
 
Possibly loose or oxides under the buss bars or terminals to the battery. Absolutely the first thing to check. Could easily be the now you see it, now you don’t cause for weird things.
Also the cheapest fix if that’s what it was. Go into the JK app and have a look at the event log and see if there’s any clues. Hopefully you don’t have a dud JK.
 
Back
Top