The numbers are off....

[piper18o]

I posted this info on another question and a follow up but did not get a reply. I have new info now that hopefully will allow for those smarter than me to answer. On my solar system, I just added three more 48 volt batteries for a total of six, and one of the new batteries charges slowly and discharges slowly. (on the picture, it is the top dark blue bar.) All six are lifepo4 eve 3.2 cell x 16 batteries, giving me two banks of three 16 cell batteries. (that is a total of 96 3.2 cell batteries for clarification) My inverter is a Sol-ark 15K. before adding these last three, one of the first three was a little faster charging and faster discharging than the other two but not that far off. That one (Red on the graph) still charges a little faster and discharges faster than the other two on it's bank but at least it is close. In the attached picture, the top three are the new batteries, and the bottom three are about 4 months older. My bus bars are 4 position bars with the main cable on position 2 while the battery cables are on positions 1,3, and 4. You would think that it is the 4 position battery that might be charging and discharging slower since it is the farthest from the main battery cable, but that is not the case. It is the 1 position on it's busbar. On the other bank, the slightly faster charge and discharge battery is in the 3 position. All cables are the same size and length.



I found that when I isolate the batteries that seem to charge and discharge slower, the wattage is wrong. For example, with only the top battery online, Solar Assistant said that the battery was pulling about 500 watts but my inverter was using about 1000. This was at night with no solar input and is an off grid setup, so power could not be coming from anywhere else. When I isolated each of the other batteries, Solar Assistant said that the battery pull was pretty much matching the inverter load of about 1000 watts. Is this a problem with one of my BMS's? All six are JK B2A24S20P. Does anyone else have an idea as to why this is occurring?

 

[NVS]

My first thought is how good are the cells balanced within each 48v battery? Did you top balance all the cells prior to assembly?

My next concern/ thought is are the jk bms able to be paralleled? You maybe better off building 1 16s4p battery with only 1 bms. Or Two 16s2p batteries.

 

[rodrick]

If all of your cables are the same length swap one of the fast charging batteries cables withe the slow one and see if the problem follows the cable or remains the same this will give you a place to start looking

 

[piper18o]

My first thought is how good are the cells balanced within each 48v battery? Did you top balance all the cells prior to assembly?

It still doesn't explain why solar assistant is only showing 500 watts from the battery when the load is 1000 watts and when the other 5 batteries are disconnected. (battery breakers off) That is what makes me question the bms numbers because that is where Solar Assistant gets its numbers, right?

 

[DougfromdaUP]

I’d start by verifying the voltages and currents with a meter.
There’s no way six properly connected batteries should have voltages that differ by .4 volts!

 

[Vigo]

I agree. If all 6 batteries are in parallel they should not be able to achieve that .4v spread without some kind of resistance problem in the wires/connections between. I like Rodrick's idea of simply swapping the best and worst batteries' locations so each uses the others 'wires' and see if the problem follows the batteries or is based on location (which would point to wiring issue). A voltage drop measurement across the leads under heavy current, or an IR/thermal imaging camera, should also be able to locate a cabling resistance issue.

 

[piper18o]

’d start by verifying the voltages and currents with a meter.
There’s no way six properly connected batteries should have voltages that differ by .4 volts!

Here is another confusing issue to me. Where is the missing about 25% power coming from?

I agree. If all 6 batteries are in parallel they should not be able to achieve that .4v spread without some kind of resistance problem in the wires/connections between. I like Rodrick's idea of simply swapping the best and worst batteries' locations so each uses the others 'wires' and see if the problem follows the batteries or is based on location (which would point to wiring issue). A voltage drop measurement across the leads under heavy current, or an IR/thermal imaging camera, should also be able to locate a cabling resistance issue.

Here is another confusing issue to me. Where is the missing about 25% power coming from? This is off grid with no solar coming in and only battery power at this hour. 150 watts has to be coming from somewhere, or the readings supplied from the bms's are wrong.

 

[Kenny_]

Yep. I have similar discrepancies.

 

[Vigo]

When I isolated each of the other batteries, Solar Assistant said that the battery pull was pretty much matching the inverter load of about 1000 watts.

solar assistant is only showing 500 watts from the battery when the load is 1000 watts and when the other 5 batteries are disconnected. (battery breakers off)

Ok, so to attempt to answer the question i have to reconcile those two statements because i interpret them as opposite. The top one says when one battery is connected and other 5 are disconnected, Watt readings from inverter and BMS match. However, the bottom statement makes it sound like even when all 5 other batteries are disconnected, the Watt readings DO NOT match. The last pic you posted also shows a discrepancy.

Does the first statement mean that you tested the other 5 individually and they all matched the inverter reading, but the 6th battery does not?

This is all pointing to a BMS measurement error, IF that is where this watt reading is coming from. Does the total battery voltage shown by the BMS match what you can measure across the terminals on the outside? If you can find a discrepancy in the voltage readings, then you could try cleaning and reconnecting all the voltage taps and see if that corrects the measurement error. If it's not a voltage problem it would have to be a current problem and im actually not sure how the BMS measures current on-board or if there is any practical way to recalibrate or repair that. Probably depends on the specific BMS. If they used an external shunt of some kind it should be easy to clean and reconnect those just to see if it helps accuracy. Added resistance across a shunt would appear as higher voltage drop = higher current = higher watts.

 

[DougfromdaUP]

The bms is my lifepower4 batteries doesn't report any current flow under 2 amps. I often draw 300 or 400 watts from the five I have and none of them report any current when I monitor them with bms_test.
(edited to correct typo)

 

[piper18o]

Ok, so to attempt to answer the question i have to reconcile those two statements because i interpret them as opposite. The top one says when one battery is connected and other 5 are disconnected, Watt readings from inverter and BMS match. However, the bottom statement makes it sound like even when all 5 other batteries are disconnected, the Watt readings DO NOT match. The last pic you posted also shows a discrepancy.

Sorry for not being clear. I ran 6 checks, one for each isolated battery. The load at the time was about 1000 watts. All batteries when individually tested, showed a wattage draw very close to the load except for battery number 1. That one showed a 500 watt draw with the 1000 watt load. Battery 6 was off a little as well but not as much, and I forgot how much. You can see on the first screenshot that in didn't seem to have as much taken out of it as well. The next time I took a screen shot of solar assistant was a couple of days later. The load was just over 600 watts. All batteries were online. The discrepency was less, but if the wattage is off on BMS #1, by about 50%, it would corrolate to somewhere near the discrepency, as the numbers fluctuated from 100 to 150 watts difference.

Very heavy draw affects the charts as well. We had overcast all day today, and we had a light load draw all day today, as nothing major required very much power. Here is what it looks like right now.



Batteries 1 and 6 actually appear very similar right now, about one fourth to half the draw of the other 4 batteries individually. That fluctuates a little as well. The first three batteries are hooked up to the left terminals on the sol ark, the last three are hooked up to the right set of terminals.

 

[Vigo]

That definitely clarifies things. But, it does seem like unless you can 'recalibrate' the BMS's internal current sensing, you would either have to just live with this, or perhaps try replacing the BMS.

It's a very annoying situation because while i probably wouldn't go to the hassle of getting a battery or bms swapped out over this issue, it does make you wonder if the overcurrent protection function of the BMS would function in the event of a problem, or if it would let the thing pass enough current to smoke its own BMS board (heat/fire hazard?) because it thought current was lower than it actually was. Im probably being dramatic, but just expressing how annoying your situation is because it forces you to decide how much you REALLY care. At least when things are 100% broke the decisions are easy!