Through Andy’s well-known YouTube channel (Off Grid Garage), I came across Victron, but also the JK BMS. Both are supplied by NKON. The aim is to build an ESS and, while we’re at it, add backup power functionality for three phases.
But what I hadn’t realised is that Andy only has experience with a completely off-grid setup, and that’s not what I want. I’m now reading here that there can still be some communication issues between the JK BMS and the Victron system, especially during charging?
So my question is whether there is much experience with the JK BMS here, and whether there might be better alternatives. I want to build two or three battery banks based on the well-known EVE MB31 prismatic cells.
Andy’s Garage is only off grid since a few months.
I experience ‘currents-jumps’ with my jk v19 and Multi RS 6000 when charging, I have no clue about the origin, others report this behaviour also, but numbers are very low.
Diagram show current jumping measured by the Victron Smart shunt, at the time PV power connected to the RS was about 5000 watts.. absolutely no need to draw power from the batteries (eve 314ah)
When this happened pv power is likely always high, and bms soc never reach 100%
The JL-BMS does not reliable calculation the SOC when sensing small currents. Thus can result in SOC jumps.
The BMS is sending charge current limitCCL and charge voltage limit CVL to Victron. After SOC is 100% , the BMS might lower the CVL. If the CVL is lowered to much, Victron will discharge batteries until battery voltage is equal CVL.
I see this with my PACE200 BMS as well. I need to fine tune CVL.
Use a custom advanced graph including battery voltage, current, CVL and SOC.
Many people combine JK-BMS with Victron SnartShunt. The SmartShunt calculates the SOC more reliable. I would start e/o SmartShunt.
Check out the NKON ESS Pro batteries. You might not be able to it cheaper on your own and they have a fire distinguisher included.
Diagram is from the smartshunt, it happened once that in the morning soc was 100% and 6 hours later 98%…with 3000 to 5500 watts of pv power and less then 600 watts of loads. Furthermore this jumping is bad for the lifespan of the battery, you don’t want fast polarity changes in the battery chemistry.
For the ts I advice to use a battery and bms brand that is fully supported by Victron.
The small down peaks of 25 A might be related the change of CVL when BMS is changing from abdication to float voltage.
I thing to remember Venus OS has some hysteria when reaching 100% SOC and having surplus PV. But I cannot remember.
55A or 1300W jumping is normal for many PWM like loads like elective stove or coffee machine on partial load. My coffee machine switches 1800 W every few seconds.
This should not harm LFP batteries.
Much worse are partially cloudy days while cooking. Then my system jumps in minutes to second from +150A to -150A (on a 1120 AH battery)
But surely this behaviour seems correct to me, doesn’t it? It’s comparable to my solar panels, which produce 1,000 W, while a heating plate, boiler or coffee maker briefly switches its heating element on.
Charging a car shouldn’t cause any problems either, should it?
I can imagine that a car battery is subjected to a different kind of load, though. It’s more one-directional and constant — i.e. either charging or driving.
You say you’ve heard this before—was that on this forum or through Andy’s garage?
I certainly haven’t seen everything from him, but I’ve seen a lot, and I haven’t heard of this here before. Has this problem been identified only on the Multi RS, or on the MultiPlus as well?
This is not a JK specific issue.
As Björn already proposed : a diagram with CVL, Vbatt, CellMax could help.
Had the same issue with a WECO 20kWh pack, one cell was > 3.65V, BMS reduced CVL from 58.4 to 53.2 V and Victron starts discharging to 53.2 V ignoring even max “Limit Inverter Power” limit of ESS. and feeds in >10kW on a sunny day, only “System-Feed-in” was limiting the discharge power.
I would reduce the absorption voltage of the inverter/charger and/or the DVCC/CVL to 55.4 volts and reduce start balancing voltage to 3.46 V to resolve this problem.
Thanks for your honesty. May I conclude that you wouldn’t choose to build a DIY battery again using a JK BMS?
There are a few reasons why I would prefer to build the pack myself. The cost; if something fails, you can repair it yourself (in the worst case, replace the BMS); and the form factor. I want to build everything into a 2000 × 800 × 600 mm steel cabinet, and many standard packs don’t fit properly (they’re too wide or too deep).
JK seems to be (I say this with some caution) the best-known BMS. Victron’s website mainly mentions other, mostly non-Chinese brands, but I don’t get the impression that these are widely used. That’s why I’m also a little surprised that Victron itself doesn’t recommend JK, even though it appears to be a bestseller. The same may apply to Seplos. Unfortunately, it’s a fairly uncertain market, with many unknown—and above all Chinese—brands that may no longer exist in a few years. One supplier warned me not to focus too much on the warranty: one or perhaps five or six years was still reasonable, but longer periods were considered of little value. There’s a fair chance that these brands won’t still exist in ten years’ time.
You’d rather not have to replace roughly 32 kWh worth of batteries.
I understood, NKON uses standard JK-BMS and standard EVE cells in their ESS Pro.
You can change every component you own.
Pls contact gem and let them confirm.
I have Gobel Power SR-GP1-PC200 rack batteries. The BmS as well as the prismatic cells can bechanged. This is the case with every rack battery with Standart BMS and cells.
You now Andis Offgrid Garage.
Some important websites about batteries and other gear: www.diysolarforum.com see particular: DIY solar general discussion section; topic: “Yixiang battery owners: any problems so far”
Check also the Youtube channel of Harold Halewijn (battery maker and Victron installer)
Check www.helionenergie.nl and his YouTube channel (former Victron employee)
You want to install all Victron gear and batteries inside a closed cabinet… its far better to install all the Victron gear flat on a fire resistance plasterboard (like knauf Diamond board on top of 18mm plywood), the batteries can be set up inside a (tiny) room made of layers of Diamond board, equipped with smoke detection and ventilation, this is also the way I have install the two 16 kwh batteries.
There is only a small chance that a Victron inverter will catch fire, less then a fusebox (groepenkast), important that you install a class T fuse holder with Adler EF3 fuse on every battery,
With diy batteries you have zero warranty, buy ready made batteries from helionenergie.nl or Nkon.
The V15 bms from jk was more a fluke then a blessing, the V19 is pretty good, the upcoming v22 isn’t any better, could be worse on heat build-up inside.
That is probably a subjective impression. From the many posts on this forum, cost is a major driver for a lot of DIY users, not quality. DIY batteries are also the largest percentage of “strange” issues - Seplos and JK are prominent. Like all things it takes skill to properly build and configure a battery, something lacking in way too many of those who undertake the task having been empowered by online videos.
You can’t beat the 10-year warranty of a commercial battery and at least it is tried and tested. The ones on the list have been around a long time, fly-by-night manufacturers don’t go through the effort required to qualify their products.
The battery is the center of the ESS universe, cutting corners there rarely results in a good outcome.
Agreed and as a result the battery is often actually undersized among other things for both the inverter and the loads they run.
Which then results in the following ‘problem’, that really is not a problem if the battery and DC set up is correctly sized.
Although I have not experienced this sudden max amps discharge down to float level, if the battery and BMS is a good size it actually is a non issue anyway.
I have my own home system with a JK BMS have used it for many years. So not speaking with zero experience here.
That is a horrid BMS configuration.
To stop/limit charging you would slightly lower CVL, not drop it through the floor.
The system will naturally want to dump as much power as fast as it can.
I would have returned that battery to sender and got one with sensible programming.
Not something you see in other batteries, so a specific, and poor quirk. Not the best basis for developing a rule of thumb for ESS design.
If you like very fine grained control over charge parameters and even use different BMs then checkout “battery safety controller” on GitHub.
Code and hardware are open source but also orderable.
Cost is indeed a factor, but if I’m prepared to invest around €1,500 in a steel cabinet for protection and to keep out the dust (there’s building work going on in the garage), then cost isn’t the deciding factor. However, I don’t want a system that has a problem every few weeks. That would still make it an expensive system. I had something similar with a Domoticz home automation system on a Raspberry Pi. In principle, it worked, but after every power cut or update there was always something else that took another evening to sort out. I now have a Homey Pro, and it has worked flawlessly for a few years. That cost a few hundred euros; now we’re talking about €8k and upwards.
