Dropping voltage EVE LF280K What I have done wrong?

[perryfranz]

Hi!
After 1 year of use my 24V battery with 8 LF280K + JK B2A8S20P has a strange behavior.
When i was using my AC split for heating it stopped working because the BMS has shuttled down for under voltage protection.
I thought because it has ended the capacity so i turned off the inverter and let the battery charge for the day.
After a few days the same behavior happened.
I tried to understand why and I thought it may be something wrong with the BMS.
Those were the data of my cells some days before all this happened when i let it charge fully to see if i have loose some capacity.



For me those are great value for an LFP cell (the MPPT was not charging anymore the battery and the BMS was balancing it).

3 days after the battery was shutting down looking at the data I thought that maybe it was some problem with the BMS.
Keep an eye on cells n°2…

I let the battery charge for the morning and I have checked it. (I have no screenshot because I panicked)
The battery was fully charged after 4h (my system has 1200W of solar power, it’s impossible to charge 7kWh in 4h).
I reset the BMS pressing the lifepo4 button on the settings.
And after I have set the capacity the BMS showed me the real capacity of the battery:

I thought that maybe the cells we’d unbalanced and I let him balance all the day and forget about it.

The day after the Same things happen but i was late for work and i only took some screenshot of the settings:




I reset the parameter again and went to work.
The day after same issue:
After reset the setting->

I tried to run some test with an heating element to see if the cell was dropping voltage during use or if there were other issue with the thermal camera:
the battery was at 78% with a average cell voltage of 3,39V (cell n°2 3,396V and cell n°5 3,401V). I tun on the heating element for 1290W and waited 10min (all the days before the AC unit was shutting down after 10 min of use).
Before turned on the element Turned on the second heating element
After a few minutes and then I turned off the heating element .

I’m not seeing some drastic voltage dropping, maybe the load was on not long enough (the AC unit do not use more than 800W in heating during spring in Italy).

During the test I have also took some pictures with the thermal camera but i do not see anything dramatic.
.
One busbar it seems to be hotter but is only reflecting heat due to a plastic shield i have applied to insulate a bit the busbar (maybe useless).


For your experiences it’s a cell fault or a BMS fault?
Do I have to run some other tests to understand better what’s going on?
I have also the YR1035 for the internal resistance and the DL24EW tester to run some capacity test.

 

[sunshine_eggo]

You are showing us cell votlages consistent with a balanced battery when fully charged and when nearly fully discharged you are showing us cell voltages indicating the cells are also well matched at the bottom end.

A LFP battery at 2.9-3.0V/cell is at a VERY low state of charge, less than 10%.

The evidence presented suggests you are simply consuming the entire battery capacity.

Have you confirmed your BMS cell readings with a separate meter?
Have you evaluated your consumption to see if it's excessive?
Are you able to fully charge your batteries every day?

Your PV produces a daily amount of energy. If you use more than this, you will begin to eat into your battery capacity and eventually the battery will deplete.

Running a heat pump (even mini-split) with only 1.2kW PV and 7.2kWh of battery capacity means you have a limited amount daily of run time unless the unit is VERY VERY small.

 

[perryfranz]

Hi!
I have not tried yet to measure the cells with a second meter.
The consumption are not high because this is a “secondary” system (I have already a 6kW + 10kWh on grid system for the whole house) and I run the laundry / ac unit.
The AC unit (Samsung digital inverter dual split 12000btu) is on only during morning (1h average) and evening (1h average) and the laundry (washing machine - heat pump dryer) are on only during weekend.
When the battery is discharges (20%) it needs 2 days to charge fully) with the 1200W string (6 Jinko solar 240W) and now I’m waiting for the second MPPT to connect the other 3 panels (300W Bisol), all those panels are used (jinko) or saved from the junk (Bisol).
It’s a constantly evolving system and I add all the part a bit at time. I do not want to depend from it but I want to build something that can help me in the future also for an electric vehicle.


I will upload the measurements.
Currently I’m waiting for the Victron 150/35 MPPT (for the 3 Bisol), Cerbo GX and smart shunt to have all the reading of the system remotely. Now the Epever MPPT has a remote monitoring but it’s not reliable and I do not have the knowledge to connect the JK online with other system.

 

[Steve_S]

It appears that you have ONE battery pack of 24V/280AH. That has 7,168Wh or 7.1kWh.
You do not state what the Start Surge is for the Mini-Split and THAT is important !
You also do not state what it is using while running in "normal fashion" as you use it.
Please try to get screen captures of both of those conditions.

Note that when any device like a Fridge/Freezer/AC including Minisplits start they have a heavy start surge (momentary) which can result in Voltage Sag and if long enough the BMS will cutoff discharge. All batteries will sag on a heavy load initially, some recover faster than others, LFP is actually pretty good on this point. In fact I would not be surprised that if it is a 120VAC unit, it pulls a full 15-17A for startup (AC). If the unit has a heater that inductive load will be heavy but that varies because different makes & models do things differently.

Quick Envelope Math that you can do.
A 120V/15A circuit is pulling 1800W and once through inverter with losses etc that translates closer to 1900W.
(Volts X Amps = Watts.) (Watts ÷ Volts = Amps). 1900W ÷ 25.6V = 74.2A At Battery Pack (Nominal voltage LFP not 24V for lead-acid.)
Further on the Envelope Math, 7168W (Batt Capacity) ÷ 1900W (pull by Inverter) = 3.77 Hrs runtime.
Of course the Mini-Split likely is NOT running at full 15A but likely closer to 7.5A when running, so at best you "may" get 6 hours runtime.

Your JKBMS settings are pretty much what I recommend to people and there is no problem I can see there except for ONE thing... Shutoff Voltage at 2.500 means you "could" brick the BMS requiring you to put it into Emergency Mode to charge it if it's below the recovery point.

If possible, provide Make & Model of your MInisplit and preferably with a link to it's specs. That will be more telling. Additionally, have you ever checked it & other heaviest like fridge etc with a Kill-A-Watt type meter to see what they are "Really & Honestly" pulling. Fridges & AC systems that get dirty and their coils "gunk up" with "stuff" their efficiency decreases & they end up using more power over time. So many forget to clean the could & screens and they get dirty fast.

Hope it helps, Good Luck

 

[740GLE]

He states it’s an inverter based mini split so inrush should be minimal.

With JK having a “voltage calibrate function”, the BMS could be displaying 27.6v for the battery bank in the app, but actual terminal voltages could be 26v.

Trust me, I was playing around with that calibrate setting and realize is just a scaling factor that can scare the heck out of you unless you know what you’re doing.

I second Eggos recommendation to reverify terminal voltages never just trust the BMS app.

 

[perryfranz]

t is a 120VAC unit, it pulls a full 15-17A for startup (AC

No, here in Italy we have 230V and the startup current is less than 1A (let say 2A) for the inverter unit (on off unit in Italy are out of market from the late 2000) and when the house is “well heated” it says on 150-200W. Inverter AC units do not have high current like the on-off one but it makes an upward ramp where it increases the compressor rpm by varying the frequency of the compressor inverter.

The washing machine (old Miele from 1993 and permanent magnet motor [max 500W spinning]) has a bigger surge start current than the inverter AC unit.

Now I do not have the model of the AC unit but i will update with it.

 

[740GLE]

What inverter do you have feeding the AC loads? What does it have for idle consumption? Is that the 68w shown in one of your screen shots?

 

[sunshine_eggo]

What is the actual load of the mini split? I would estimate 600-700W .

Do these events happen more typically a few days after the dryer run?

Flat panels generally don't perform well unless it's late spring to summer to early fall, or you're in the tropical zone.

What inverter do you have feeding the AC loads? What does it have for idle consumption? Is that the 68w shown in one of your screen shots?

Good point... 68W * 1.63kWh - 22% of battery capacity consumed daily.

 

[Steve_S]

here in Italy we have 230V

KEY CRITICAL INFO ! Had I known that I would have written differently and not wasted my time.
GOTO you membership profile and add the "Closest City" and Counttry for your location. We have folks from Australia to Zambia here and every powergrid/system is different on their own.

 

[740GLE]

I mean the guy does have the time, effort and tools to post thermal images of the battery and BMS working, but a simple DMM to measure actual battery pack and cell voltages against the app. . . .

 

[perryfranz]

Make & Model

The AC unit is a Samsung dual split, the external unit is the AJ040NCJ2EG/EU max current is 900W for the external unit, each internal unit let say that consume 100W each one, so 1100W in total in the worst condition.
I have ever see the machine draw over 600W.

happen more typically a few days after the dryer run

Never happened after a drying cycle and last time i have checked the dryer was connected to the grid (or the house, not the off grid system).
The heatpump dryer is an Hisense D901GE (based on the Asko model) and the max power required is less than 600W. Unfortunately not the inverter compressor.

guy does have the time

It can be not easy to understand but it took me less than 40min to capture the screenshot during the small test and plug the portable thermal camera into my smartphone. And i have done everything after a day of work and studying. It can happen to forget to do something even if fo someone else it’s an immediate thought. I wrote in fact “Do I have to run some other tests to understand better what’s going on?” for this purposely because I knew that i was forgetting something.

What inverter do you have feeding the AC loads?

It’s an Edecoa 3500W, the standby power is 13W.

 

[upnorthandpersonal]

From what you have posted and from what I see on the sceenshots, it just looks like a battery cycling and using all the energy. Maybe you can run a long time screen recording session to figure out your actual consumption, or use a kill-a-watt like device on the output of your inverter. Measure your actual generation as well.

 

[perryfranz]

it just looks like a battery cycling and using all the energy

I’m thinking the same thing at this point, probably i have overreacted but it never happened before.

Now I will follow the suggestion of @Steve_S and @740GLE to measure with another meter and the “kill a watt” to read the output.
This evening i can take my time to run everything without any rush.
I can see from the remote monitoring of the Epever MPPT, the battery is basically fully charged so i can simulate a complete discharge.
I will keep you updated and thank you all for the support.

 

[740GLE]

Let us know what the DMM reads of the battery when the BMS says full.

If you pick up a victron smart shunt it’ll estimate the remaining time depending on the current load. It’ll also track 100%

 

[perryfranz]

Now I’m still running the test.
Before starting it I have read all the paper and notes that I have made when I built the battery in 2022.
One of the notes were “set the correct voltage of the battery on the BMS before using it, it may be not calibrated and need to be adjusted” I took all my 3 multimeter and have done the measurement and set the correct voltage.
Now I have drowned 2.18kWh (from the killawatt meter) in 1h and the battery is at 65% with 182Ah remaining, the average voltage is 3,238.
The meter on the battery is reading 25.93V and the BMS (under load) 25,9V.
I’m screen recording the jk app just in case and filming the measurement.
I’ll let you updated.

 

[740GLE]

How bad was the voltage adjustment off by?

Also you measured and adjusted that calibration with no load right, not even the inverter idling current.

 

[perryfranz]

How bad was the voltage adjustment off by?

It was off of 1.09V, when the BMS was reading 28.15V the multimeter was reading 27.06V.
Now after a week of charging and discharging the max difference that I saw was 0.11V average.

Yes, i disconnected the inverter and MPPT (fisically from the battery) and then I had mesured everithing.

Now I can monitor the battery remotely with the victron smart shunt and I have not seen anymore the previous behavior,.

2 Lessons learned:
- when something is working just fine, do not touch any parameter (only when a fault occour)
- do not panic and read everithing done in the past and trust the comunity.

Thank you all!

 

[perryfranz]

After the calibration I run the test suggested:
At the beginning I had (on the BMS 26.86V) and it was connected to the inverter:
- 5W Led lightblub
- Killawatt meter (mine is a Meterk I bought from Amazon years ago)
- electric heater (rated 2000W at 220V, it has over 30y, the label it's not so precise) the kill a watt meter says it was draining 2200-2400W.

I had it run for 2:43h and it drained the battery from 98% to 9%.
The kill a Watt meter had measured a total consumption of 5.90kWh only from the electric heater.
Let's add 12.5kWh for the bulb
and some heat dissipation from the cables ( I have to replace the disconnecting switch of the inverter because it heats itself a lot under heavy loads).



I have all the screen recording from the BMS, i have only to speed the video from 2.4h to someting quickier, but i can attach the 2 screen from beginning ad end.



Looking at the data, probably i need a better meter to calibrate the BMS because from 3.1V avg i could have drain something more from the battery and not the only 22Ah remaining that BMS says.

I'm absolutly happy with this setting, i have to power only some non vital appliances and if I can avoid stressing the battery with deep discharge I'm happy with it.

 

[740GLE]

It was off of 1.09V, when the BMS was reading 28.15V the multimeter was reading 27.06V.
Now after a week of charging and discharging the max difference that I saw was 0.11V average.

Yes, i disconnected the inverter and MPPT (fisically from the battery) and then I had mesured everithing.

Now I can monitor the battery remotely with the victron smart shunt and I have not seen anymore the previous behavior,.

2 Lessons learned:
- when something is working just fine, do not touch any parameter (only when a fault occour)
- do not panic and read everithing done in the past and trust the comunity.

Thank you all!

Glad to hear is was an easy fix.

I am surprised that parameter is so easily messed with. IMO I can’t think of a reason why it’s even there, JBD don’t have that.

 

[perryfranz]

parameter is so easily messed with.

If you press the @Lifepo4 button on the "setting" page (after you have entered the password, so it's not something that you do intentionally) the BMS rest all the parameter and also the adjustment done in the past.
So, lessons learned and now I will touch the parameter only if it will have a falut or getting replaced for a 48V battery... And I hope that is gonna happen soon.