I think the problem comes from the chemistry of the cells, they are not the same brand, therefore different manufacturing, the chemistry must vary. It would be necessary to make a charge/discharge curve of one of the cells of each pack and check if the absorption duration varies between the cells.
Andy look to your Mosfets - Not all Mosfets are equal - BMS with Mosfets have different forward resistances and/or different numbers of Mosfets in parallel will have different resistances - older cheaper Mosfets tend to have higher forward resistant which get worse as the heat up - a crude test would be to check which BMS get hot under similar high loads - Hugs from UK
I agree with Trevor. Also, the on resistance of Mosfets is non-linear over current. In other words, you will get different resistances at different currents.
@@marklockwood7592 That's so right- there have been a lot of advances in Mosfet' design and material purity in recent years - I imagine not all manufactures have caught up - some e-bikers have pepped up their bikes by the simple expedient of changing the Mosfets in their controllers.
I've got to say, these JK Inverter BMS are nice, I've got 3 running in parallel right now, soon to be 5. Andy, you have helped me a lot in understanding these BMS and how to connect them and get them working right, thank you from sunny Nevada USA.
I’m doing something similar but 7s 24v 280AH my BMS is a Seeplos 3.0 7s 24v 150A I have a stand - up battery case ! You are 1 of the top people on earth in knowledge about this !
First, please notice the charge voltage and the voltage at the battery At 8:09 they match for each battery (approx), - the 55.2V to 54.8V offset is due to the bus-bar-to-battery drop at 80A? V1 V2 V3 Batt V 54.84 54.82 54.8
I think it is due to the balancer, the middle shelf will have a more even cell capacity and so they all reach 100% at about the same time, meaning it can absorb more energy, the other two shelves will be more imbalanced so once ONE cell is at 100% the overall current will drop because of them all being in series, even though the other cells may only be at 95% or so.
Andy, I believe that the middle battery pack is better balanced than the other two, which causes it to have a slightly higher resting voltage than the other two banks. When you slowly discharge them, the middle pack then discharges more as it has a SLIGHTLY higher voltage overall for the first few minutes/ Ah discharged When you do a higher discharge rate, the overall voltages of each pack are dragged down into the flatter portions of the curve, which will then make the battery packs all supply similar current, hiding the higher resting voltage. Since you are working with 4 sets of variables here: cell size (304 vs 280), cell busbars (CU vs AL), bmses and the effects of longer cabling for the lower packs, I propose that you fit 3x identical BMSes for 1-2 weeks - preferably connecting them to your raspberry Pi / VRM to allow for logging and totalizing - This is "easy" since you already have all these new JKbmses which were bought at the same time and can run the same firmware. My hypothesis is that the higher trailing charge rate for the middle pack will disappear after 3-5 cycles
The voltage difference at charging is the key give away.. Top and bottom pack have a higher pack voltage when charging, there for are less balance. Furthermore, as you paralleled 3 packs, once the difference between the packs becomes more than 2 mv, the batteries will try to balance out ( from my own exp.)
@@houseofancients It looks like the top and bottom batteries have a lower delta voltage in the cells. at 9:58 the top and bottom battery have a delta of 25mV and the middle battery has a delta of 62mV. Although that might be because of the higher current in that battery. I wonder if the battery cells were balanced if this issue would be solved.
Aaaannnnnd. Balancing is wasting some energy right? 😊 So little bit here and little bit there. My opinion is that with such count of variables this "problem" not worth another try/guess. It may still end up with a wrong conclusion. Thanks Andy for all the good work!
Have a look at cell 11 in the middle shelf. At 87% SoC it is already above 3.45V. I believe this single cell causes the internal resistance of the whole battery bank to be higher than the other banks, resulting in this "problem". I have two Seplos banks (with NEEEEEEEY active balancers) where I observe similiar behaviour, when charging with 100-140A (50-70A per bank on average) cell 9 in bank 1 goes to much higher voltage than the others. When discharging at moderate Ampeages the voltages are fairly equal, but when discharging with >100A the same cell has the lowest voltage. Highest voltage when charing and lowest when discharging means the weakest cell. I also observe that the bank with the weak cell charges slower than the other bank.
Hi Andy Really enjoy your video info and style and may have partial answere to above glitch (not seen in comments). Following things I have learnedfrom LIPO ebike builds. First, assume things you can't change are valid/the same (cell chemisty, internal resistance, etc.). Next, make similar the controllable variables such as: cell temps, BMS model, BMS voltage calibration, lead length/size, cell SOC, etc. This would mean manual "balance" to actual full capacity or SOC, replacing vertical bus bars with equal length cables (maybe only for test) or at least connecting to bottom of one of them), measuring internal resistance at actual usage high current dc level vice ac impedance, replacing aluminum bars with same copper, and NO balancing during test. This should point to most of cause if ONE variable at a time changed. Or, since that would take a long time, just ignore it but keep in mind for next time 3 or more banks need to be in parallel. Can't find it right now but correct way to parallel 3 batteries/banks really looks weird (SLA but LFP also due to flat charge/discharge curve). Keep up the good work please and cold one sent., Dick
I have some Jakiper JK48V100PRO batteries. With one of the packs, I observe some of the same as you do. That is the oldest pack and it has slightly different cells inside. Different model cells and possibly from a different manufacturer. Possibly with a slightly different charging curve? But as you said, it is not a problem.
Interesting Topic, ... as I have three 24v 280 Ah Eve battery banks (purchased one or two years apart), that all have the same 3 BMSs; with one battery bank charging or discharging at lower amps. I have suspected difference in individual battery cell chemistry and internal resistance; not sure. Was going to switch battery bank positions and keep bms positions while cleaning all bus bars and bms wire loom connects (someday), .. as a test for learning more all the time ... for later. Not immediate concern as my setup is working good. Look forward to your next discovery on this topic. 🐾🐾🎶🌠 15:26
OK, so, if your banks are from different age, they may charge a bit different. That will be quite normal. Calendar age takes its toll. But yeah, swapping the positions can mitigate that.
That is why i use different solar and charger array for each battery bank, and combine the battery banks through diodes so no power is transmitted from one battery bank to the other. Will be making some videos about my system soon. Amazing work Andy. I love your channel and your expertise. Keep it up 👍
The balance current from one bank to another is not a problem with LiFePO4. It is very small or even non existent. Depending on your system size, you can install separate chargers and arrays for each battery as you did.
With my system it is better with the diodes because the BMS's control solar charging for each battery bank but all BMS's control the inverter, so if one cell would fail without the diodes all power would go to the failing cell and fire/explosion could erupt.
@@StarJackerSRL That will not happen. that's what the BMS is for. If one cell fails, the BMS will disconnect this bank from the system. The diode will not help.
@@OffGridGarageAustralia 123SmartBMS does not do that on it's own. My BMS only controls the inverter and incoming charging from solar panels, that is why i use Victron diode battery combiner.
I have a similar situation between my 4 parallel batteries. Two of them are CALB329Ah and two of them are envision 306Ah and 308Ah. The Envision batteries seem to get full first and under certain conditions they can even backfeed energy into the CALBs. I assume it has to do with slightly different voltage curves And as you said it's not an issue in practice.
The explanation might be easier to explain using the charge curves of the batteries. When we build these batteries we found that matching charge curves for the 16 cells is more important than the internal resistance of the cells (importance of using same manufactured batch). What you have is a slightly different charge curves beween the batteries at the 3.35V-3.48V cell voltage range. Thus at the "knee" of the curve the middle battery is just a bit slower to get to the bend. It actually means that the cells used in that battery is better matched. The benefit is not huge but if you do the calculations the capacity would be 0.25-0.8% better than the bottom battery.
I have noticed the very same thing with my system. In particular my 1 bank of 3P x 16s Identical cells (EVE 280's), identical BMS x 3 packs wired similar to yours. All batteries were from a single batch, BMS's from same production run. Active balancers also same production run. All were installed and commissioned within days of each other. In my case (and yours) cell capacity matching of the banks is the only difference. The best "capacity matched" bank, takes the longest to discharge and recharge as compared with the others. In your situation, you could also add the small differences within your different bms's.
Glad you had a great family Christmas Andy, you deserved it. 👍 Happy 2024, and have great year. Interesting… I’m still watching the video but my speculation: maybe different BMSs are not such a great idea??
I think is the BMS not the batteries, I think the JK will get more current when you are almost fully charged (say 55.0). But we will see that in your next video. Thank you for doing all the hard work for us !
Measure the voltage drop across the positive connections and then the negative connections (input to output) of each bms with the same current flowing through each to see if the BMS is causing the issue (maybe the FETs have a higher forward bias)
Happy new year Andy and community. Can you also measure the battery bank itself, this way you have your bank and BMS resistance. There could be a difference. Maybe also validate with the same clampmeter what the real current and voltage is, set the BMSes accordingly. Keep up the good work this year! Thanks for all your efforts in debunking all myths and issues related to batteries, chargers etc🎉
Thanks Edward and Happy New Year to you as well. Yes, some have mentioned that here that the BMS could interfere with the measurement. I'll do more testing.
The one symptom I noticed was the middle shelf having a slightly lower voltage than the other two when charging above 3.4V, which explains why it charges for longer. But as to why it has a lower state of charge, I guess you'd need to look at each variable in isolation, such as swapping out the cell busbars or the BMS with one of the other batteries etc.
There could be a small difference in measuring the voltage across the three different BMS. Some here have seen a slightly higher voltage, you see a smaller one... I have more insight in one of the next videos...
Just to the side of the discussion. I use the YR1035 to quickly test the internal resistance of LiFePo4 cells. I then tested the same cells in a combination for battery banks via the EBC-A20 PC software and finally the same application with the EBC-A40L. All these tests were performed on dozens of Winston 60Ah cells and 16x EVE 304 Ah cells. All the Ohm values were completely different from each other even for single cell. Often even orders of magnitude different. You perform this task with the YR1035 for all three banks, so the implication of the instrument will most likely be very similar for all measurements. However, please remember that the real internal resistances will MOSTLY be quite far away from the results you define.👋Anyway Thank you.
How long until the middle bank catches up? I can't imagine it being more than a minute or so? Do you see the same if you charge to 3.5V? If you do, how long does it take to equal out now?
Hi Andy! I have been figuring out the daisy chaining batteries and the different voltage losses in those parallel connecting cables. If you think about battery voltage difference with almost empty and almost full battery bank. That voltage difference is much more than the voltage difference caused by the non equal battery cables. So the small voltage difference caused by the cables is not so big deal. The batteries will become almost empty and almost full :)
I would say around 10min until it catches up with the other two banks. At an average charge current of ~25A that could be 5Ah. Seems excessive to me... I don't really have a device which can measure the charge curve of a whole bank... wait.... maybe...💡
I think the different charging current is as a result of the availability of internal active balancer in the JK BMS. All the cells in the JK BMS bank are balanced and takes the all the energy they can store. I suggest you verify the overall Watts hour stored in each bank and compare with the rated capacity. Thank you for all the good work you are doing.
Sweet setup! Your garage is the electricity and free solar energy Valhalla. So I bought a BMS active balancer that doesn't have any visual or bluetooth notifications. I connected it for a few days and think it worked bc the cell voltage difference is minimal. Do you suggest leaving it on? Doesn't it interfere with mppt or inverters/buck converters..?
Hi Andy, try to measure the voltage drop across the battery bank BMS during discharging and charging. Maybe the BMS internal resistance changes/different during discharging and charging.🙂
Andy regarding alu or copper the thickness is important but also the contact area and material of the bolts count. I have a similar situation, need to find time to investigate since it stress more one of the batteries.
Well, the contact area is not too important if the torque is sufficient. Whe have tested this here and could push over 230A through the tiny contact area of terminal and bus bar: th-cam.com/video/VPQvs-2YYzA/w-d-xo.html And I'm still running this exact setup in the top and bottom battery.
Hey Andy, Happy New Year! Just wanted to ask about the 3D printer update and if you share any of your STL files you designed or find useful? Always enjoy your videos and congratz on the boat!
Is the middle shelf the 280K and the bottom 280 not-K and of course the 304's are also non-K type. Just a subtle difference in electrolyte and therefore a slight change to the charge curve? Like you say, its not causing problems, and other than swapping round BMS's or busbars eliminate those variables its the only difference other than capacity - unless the middle shelf were B-grade, or a different factory or batch is maybe all it takes to shift the charge curve a little.
Hi Andi, das JK misst eine niedrigere Voltzahl, während die anderen beiden bereits die 55,2 Volt erreicht haben. Man kann doch das JK von Hand kalibrieren. Wenn man das JK bei beispielsweise 54,7 Volt auf 55 Volt kalibriert, dann wird es das erste BMS sein, welches die 55,2 Volt erreicht und auch als erstes runterregeln. Sind es diese kleinen Zusatzboxen, die die BMS mit dem Cerbo verbinden, welche für das Herunterregeln verantwortlich sind? Normalerweise sind doch die Solar-Charge-Controller für das Herunterregeln zuständig und nicht das BMS? Das hat sich doch erst mit den kleinen Boxen geändert?.
So, what is the title of the video that answers the question presented in this video? The next videos appear to discuss the "reset" issue with the JK Inverter BMS.
Merry Christmas and a happy hangover. Did you zero the low ohm meter? How repeatable is the zero measurement? How repeatable are the resistance measurements? How matched are all the cells?
I think the difference are caused by the measuring accuracy of the BMS. The BMS measures the voltage and decides if the battery is full or not. You have determined that this happens at 55.2 volts. But each BMS has its own accuracy. We are talking about small differences, as you note yoursel, so a measurement accuracy of 0,1% could mean that one BMS trips in reality at 55.25Volt while the other trips at 55.14Volt.
The BMS only measures the voltage but does not control or regulate the charging. Even if one BMS shows only 50V, the battery itself will charge exactly the same. (Obviously, the OVP of the BMS will not work correctly ins such a case)
First see if it follows the BMS when swapped. If not, I would set up a high current flow and measure for voltage drop in the milliamp range on both the positive and negative cable paths of the middle battery. That is how I found a weak connection on my solar system.
Hi Andy, what if the problem was the real capacity of each battery? maybe as the years go by they have different abilities that bring you this problem. I don't think the problem is the BMS.
Hi, the different impedeance would cancel out during an discharge/charge cycle, but you have not taken into account: a) if discahrged, (and then leaving idle during the night) the banks will equalice by voltage and so charge the bank witch was discahrged most du to lower resistance. b) your top bank has higher capacity, and so lower resistance provinding more power
Could it be that because of the lower total resistance, bms included, of the center bank more ah were removed and therefor mire AH were required to charge the cell? The voltage was lower during charge, after all.
I guess the Battery Shelf 2022 might deserve its own video about the performance and cycles in the past 2 years I guess. And then it might also deserve a protection well known as DENKMALSCHUTZ cause it is so iconic cause everyone can recognice the long and winding road back then. And it is clean and nice - compared to the collection of batteries that appeared seconds later. After a long and winding road with EVE LF160 cells grade B we finally got a set of 16 x EVE LF280K grade A from a german electrician shop in Hannover. Grade A means in this case: all produced on tthe same day same production site and a row of serial numbers with a 305 Ah result for each cell over 3 single discharging test with EBC A40L in a heated room with 25°C according to EVE 's test procedure described in their 20 page long description. 305 Ah means they are also running quite close from top to bottom with the Seplos BMS balancing close to nothing cause max deviation is about 0,002 V but only 2 are better cause 13 are 0,001 . And that is most likely not a cell difference but caused by the position in the battery cause when I mixed it up the results were not the same and the worst became the best. When I put it back in the previous order I got the same result as before, this time again the worst cell in the battery, whicch means from best to worst it is just a question where the cell sits in the battery cell chain. And the top balancing is a rather short procedure till its done. This is a complete different result and world compared to the LF160 Grade B cells (have been lasered and produced within about 7 months).
Did you calibrate the Voltage measurement of the JK BMS? There is an option in the settings to calibrate it. It could be your just thinking it is not fully charged?
good afternoon andy andy question, I'm a little confused about the communication via can bus between the cerbo gx and the new jk bms, you have to modify the internet cable or you can use an original one without modifying it, that's my question, I'm confused and thank you in advance, andy
Andy, good afternoon, thank you for the response, but do you have a diagram on how to make the cable? I would appreciate it a lot.@@OffGridGarageAustralia
Different BMS, different circuitry. I have 4 batteries and made them with identical cable lengths, cells and BMS, I do have some variation between the batteries during use, but only minimal - up to 5% of the charged status maximum.
Sorry that sounded a bit rude, I'll try again.... Hi Andy, I think it could be Calibration of the different BMS's If you check voltage at each battery bank voltage will be same. but JK BMS is showing lower hence will keep charging as it thinks it's at a lower voltage. I have a 5 bank that did the same thing. And keep up the great work Best channel on here.
Hi there, but the actual voltage on all BMSs is the same - and non of the BMSs is limiting the current or stops charging the battery. The measurements of the BMS does not matter at all. What do you think about that? Greetings
Hallo Andy, ich weiß nicht ob ich es abschließend richtig verstanden hab, aber den Stromfluß zu deinen Batteriebänken bestimmt grundsätzlich mal der Spannungsunterschied. Die Zellen haben beim Laden eine bestimmte Spannung erreicht (z.B.55,0V). Der Unterschied oder die Differenz zur eingstellten Ladespannung (55,2V) beträgt nur noch 0,2V. Wären jetzt alle Innnenwiederstände gleich, würde überall der gleiche Strom fließen. ABER, die Spannungen der Zellen (Bänke) sind unterschiedlich, verrechnet mit dem jeweiligen Innenwiederstand bekommt man einen anderen Stromfluß bis zur vollen Aufladung. Am Ende der Aufladung haben alle Bänke die gleiche Spannung. Aber bis dahin ist das Verhältnis aus unterschiedlichem Innenwiderstand und Spannungsunterschied der Bänke mit der Ladespannung entscheidend. Nur ein Spannungsunterschied treibt den Strom durch einen Widerstand. I=U/R also "Spannungsunterschied, geteilt durch den Innenwiderstand = der aktuelle Strom" oder 0,2V/0,011Ohm=18A, was deinen Beobachtungen nahe kommt. Kannst gern auf deine Batteriebänke umrechnen. VG Jan
I don’t know if one of the battery cables has one more strand of wire or a thickness and size of the aluminum bars. But I will check out a cold beer 🍺 and get back with you 😊
Hi Andy, Switching bms could be a good test to see what is happening. My 2 cent is simply cell. You don't have 3 batteries but 48. Over 48 cells, they are suffering differently over the time.
As the charging voltage applied to the shelf is the same for the three banks, for more current to flow through the middle bank its resistance must be lower. Using ohms law actually calculate it rather than dicking around trying to measure it. Use your amp meter to confirm readings and Swap the shunts around after checking all connections to confirm that there is no variation between them..
Dear Andy, what I found in my shelf, that when charging with Jk weeker cell goes higher voltage, then active balancer takes out energy, but when current go's down at 55.2 this weeker cell starts to use more ampers and, and active balancer put back energy to weeke for absorption. While pasive balancer was not able to make so big trouble for weeker cell. So, my prediction is that please change active balance current from 2A to 0.1A and look for a few days. You will be surprised that weeker current is enough.
I have the same but i am using 3 identical jk bms. But my cells have a different manufaktorer. My CATL cells still take Power even if the eve and the rept stopped.
My take is that the middle battery has more losses for some reason. Hence, when discharging it is discharging a bit more than the other two, and it has to recover more energy on its charge cycle.
I’m a bit puzzled: By not measuring the impedance from each busbar across the cables into each MCB, through each MCB, and across each battery, you still have 3 unknowns (at least, unmeasureds) in the system-the Positive busbar to MCB positive output and negative MCB output to negative busbar for each of the 3 batteries. This might explain higher resistance and lower voltage of the middle shelf battery. A second source of variation in resistance could be your busbar connection configuration. I’d have to go back and see the video(s), but I think you did use a diagonal config and kept the MCB to busbar positive cables equal length (to each other) and negative cables equal length (to each other). In theory, each of the unmeasured cable/busbar cumulative resistance should be the same if you used diagonal wiring and had same distance between each battery cable. Unmeasured Sources of variance could be torques, washers, crimps, different wire chemistries, corrosion, and-outside the wiring-the MCBs when closed. If, after you have done the swapping of the BMS’s (I’m assuming YT production times and you’ve already done the swapping), and the slower charge/load rate doesn’t follow the middle BMS, you might look at these things which are currently assumed.
Hey Andy, when you switch the BMSs, you should check the Resistance of the Breakers too (measured after the Main Breaker). It would be also interessting if the effect can be observed if you charge to a higher voltage. I think it could also be a slight difference in chemistry of the cells. Looking forward to see what you can figure out there! 🐸
Yes, breaker could play a role but I doubt that would make such a huge impact. Charging to a higher voltage shows the same with the middle battery still charging longer.
In my case the 100Ah set says "battery fully charged" at 55V and disconnects charging. EVEs of 300+Ah stabilize at 55V and don't absorb anymore, while 300Ah+ CATLs get to 55.2. Maybe something to do with cable lengths also? Cable lengths are pretty random between ~25cm and 2m. I'm thinking of a small V-drop across them.
The higher the current the higher the battery temperature and lower the internal resistance. Lower the resistance, current gets even higher. So one battery always wins the charging race. Adding series resistance (thinner cables) will help to balance the charging rates.
Why is there only 91k subscribers? He is super informative!
4 หลายเดือนก่อน
We seem to have a lack of solar and battery nerds that enjoy the Australian/German flair of dear Andy.. I saw a channel with a female who is quite a bumbler when it comes to putting batteries together as she has had sparks flying and was giving poor advice but yet had a huge number of subscribers.. Maybe Andy needs to get his beautiful wife more in the videos as this would seem to boost subscriber counts?? His advice is backed up with first person testing and he has grown with this knowledge over the past several years IMHO to be an authority/expert from doing..
hi, would you be able to do an experiment with a 15 cell battery (like a pylontech or home made) and a 16 cell with a pacebms. I have 80kW with the Pace (8 batteries) and 37kW with pylontech (11 x us2500C); what happens for me is: when both are connected, only the pace discharges unless the load is extreme (like 10kW, then the pylontech will take a very small % of the load, maybe 5%), If i turn the Pace off and allow the Pylon to discharge as its the only bank, when i turn the pace back on, it dumps whatever it has into the pylon to take that back to 100%. the batteries are connected as follows: 6 pylon, 4 pace to a lynx power in, 5 pylong 4 pace to a second lynx power in, then a lynx shunt , then a lynx distributor which has 3 inverters (voltronic 11kW) the shunt is the aggregator for the battery stats, with the pace connected via can to a cerbo and the pylontech connected to solar assistant for data, i also use SA to monitor the inverters. any insight appreciated.
Hi Andy, I have the same question as you! I connected two battery packs, each with 16s 280Ah, using the same BMS (which don't communicate with each other), cables of the same length, an equipotential bar, and the same brand of cells. The only difference is the production date of the cells. I also tested the resistance by measuring voltage drop. But nothing! I just can't figure out why the cells are behaving so strangely. Sometimes they seem perfectly synchronized, other times one discharges much more than the other. I think there are many factors at play, including cell temperature. However, it seems more like a theoretical question because, in practice, the system works. Overall, the cells seem to work similarly over time, so I assume aging will be similar. Thanks and greetings from Italy!
I have EVE LF280k from different production lines and only a few days apart. They have in average around 4Ah difference in capacity, but they have diffrent charge curves. Seems there are big tolerances in materials and manufactoring.
The charge curve could be what is different here. Even having a 100Ah and 280Ah in parallel did not show such a result as the 280Ah bank always charges/discharges faster and they meet again at the end.
@@OffGridGarageAustralia In reality, the behavior seems completely random, as batteries often reverse roles. Is it not related to the quality of the cells? It appears to be linked to random factors.
Hello Andy, and happy new year.... Just a down side to things......here in UK ....our days are getting longer and yours are getting shorter......ha ha (Don't pay it back in 6mths) I take it all your BMS have same settings, but as you said in last minute of video..... You have different BMSs and all 3 are working on 3 different banks.... Ok top back is 305ah Bottom 2 are same but with different bus bars and different BMSs.......for get the top one. Now the lower 2..... different BMSs, different bus bars , different temperatures (due to hight day night thing) Different age. I'm thinking it be something simple.... forget data thing......also middle one is first of the 280 ah cells connected to dicarcarge point??? Is it as simple as this bank is next low ah bank and so discharge a little more everyday than bottom bank???? I know you have built a fantastic battery bank and its well perfectly well laid out..... You do have a Mish mash of cells BMS and bus bars..... If it was all the same.....then I would ask the question as you have. I'm a fitter on plant equipment......you can have two machines.....but both are not alike... Never over look the basics.....forget data. I don't think it will help, but thats how end up fixing faults. Sorry for going on Andy. Look fwd to out come.
Yes, it has an active balancer. It uses a lot less energy as it 'recycles' the energy rather than burning it off as in a passive balancer which the other two BMS have.
I think manufacturers have varying charging curves so your results don't seem too odd. Would be more odd if you were using the same BMS brand across all three batteries with mirrored settings. Are all of your BMSs set to the correct chemistry?
My goodness. Battery current diverging during charging as each battery comes up to fully charged at a different time due to variations in manufacture, wiring etc etc, leaving the one with a lower relative SOC to soak up potentially undesirable levels of current in a system that does not monitor individual battery current? That can't be right as it's been stated a few times in the past by the channel operator that he's checked and his batteries don't have this problem. I do remember some guy banging on about this in the past and that seemingly cluey guy also said that it happens to pretty much every battery system out there which is why the commercial setups monitor this sort of thing. We're not quite there though as apparently it's not really a problem, but give it a couple of years and we'll see how the story evolves. It is true what they say. You can lead a horse to water but you can't make it drink. You just have to leave the horse there and hope it eventually works it out for itself.
Yeah swap bms would tell if bms is the cause. But i guess it is because of the active bms method. Maybe it will waste more energy than passive and take longer to fully charge. 😅
passive balancing discharges higher voltage cells via resistors - throwing away energy as heat. Active balancing discharges a higher voltage cell into a capacitor, and then uses the stored energy to charge up a lower cell - less heat overall
@@LibertyDIY BMS power consumption by itself is in the order of a few milliamps. These packs are cycled daily and the behaviour is highly repeatable one day to the next -even a 20mA consumption/wastage difference over 24 hours is only a 0.48Ah difference, which would be replenished in less than a minute at this 40A charging rate difference
The only thing I can think of is there is a slightly higher self discharge on the middle battery. Is there a difference if the time between full charges is long or short?
My heltec bms screen shows like your jk-bms. I wonder why ? Referring to toe video appearance ! It identified my heltec bms as JK_BD6A17S8P . Is Heltec using jk-bms software ?
Voltage Current Calibration of the BMS! I use 3 Pace BMS after calibration oft the BMS on 3,65 Voltage everything is spot on! Also The Neey Balancer can be calibrated… I tested 10 Neeys on 1 Battery at 58,4V they are all slightly off 😉 after calibration everything is good… balance Leeds calibrated is also possible!
Would measuring and recording the voltages across each individual bank with an accurate DMM while simultaneously recording the amperages, either charging or discharging, allow you to use ohms law to determine the overall resistance of each bank? I imagine each bank uses different sensors to evaluate the amperage going in and out of the battery banks? There might be some significant variance between them as well, particularly at high or low charging rates. Their response may not be as linear as one would like. It would also be interesting to check the voltage differences at different locations on the bus bar - at high amperages, even small resistances could cause significant, and therefore measurable, voltage drops.
Did you ever get an answer to this? I was thinking the "Profile" that each BMS has in the firmware.
4 หลายเดือนก่อน
I always thought that the charge termination curves engineered into the different BMS brands would certainly have an effect on the tail end of the charging.. I have two identical 17 cell 280a EVE batteries with identical (almost) JK 2A24S20P BMS controllers with the only difference being the few months difference in manufacture date and a slight difference in firmware versions.. One of the two banks always seem to lag the other when charging and even at idle where one bank should trickle into the other and eventually equalize but they don't.. So my solution was to slightly raise the full charge cut off voltage point of the trailing battery a bit higher than the leading battery and this seemed to bring them much closer in line with each other.. I charge at 3.51vdc have the full set to 3.5vdc and the float at 3.45vdc per cell..
There's a calibration feature for both voltage and amperage. Wouldn't that make difference if the calibration was off? For instance, voltage from one BMS is 56 for pack x, but the other BMS would show 54 from the same pack. If that happened, then the behavior would be different... Maybe?
Hi Andy, could you or one of your subscribers please help, I'm a newby to battery's and solar but wanted to give it a go...I have 4x 12V 120ah lithium battery's and with my 10a lithium charger its take's them up to 14.6 volts when charging but after awhile they settle at 13.3v...is the 14.6 doing any damage to them...also what voltage should I not go past when discharging them...thank you so much for any advice....Steve from Perth WA
Hello Steve🙋♂️ 14.6V is pretty much the maximum these batteries can be charged to. As there are 4 cells in each battery, it means, each cell will be charged to 3.65V which is again the maximum these cells can handle. The battery management system (BMS) in these batteries will turn off charging if one of the cells goes over this 3.65V. So, it is almost impossible to charge to exactly 14.6V because the cells are never 100% balanced. Recommended is a more conservative charge voltage of 13.8V-14.2V for these batteries. This will still fully charge the batteries but is more gentle to the cells. That the batteries are settling after you take the charger away is perfectly normal. They are still 100% charged.
It looks like its a combination of the higher resistance (so harder to charge) and the active balancer (which is constantly bleeding off some current 24 hours a day) means the middle pack is effectively getting less into it. Because they are in parallel, you can't really "see this" on the voltage as the voltage will always automatically equalise between all 3 packs - but it appears it manifests as strange but repeatable differences in current? Very interesting and to be honest it doesnt really matter as long as all packs do charge to 100% and all cells are kept equal in each pack? Also I don't think you can see internal resistance during discharge as the cells can effectively discharge instantly (if you short them)?
mines still working very good saving us power bills , after 13 months ,27v - 8 eve 280ah cells . Victron 1200va and 100/50 mppt in 24v mode ,battery is very efficient , only had that one issue with BMS crapping itself ,, love it free power free aircon
I have 4 PCS 3.2v 280AH bank 24v all though charging I set the charger to 5A for better absorption , “ the issue was I charged the bank to 14.6v = 3.65v so post 2 be but I have a 2 batteries 1 in 2 different banks each 1 of those batteries only charged to 3.37v and 1 battery in 2 different banks charged up to 4.0v and the 3.37v batteries stayed there ! What’s you idea “ Please “ ?
I think your guess is right. The middle battery get's fully charged last and the top and bottom absorb earlier. But it could also be the cells or the BMS voltage drop 🥴
I am going to suggest that changing the BMS on the middle shelf is not going to change the charging behaviour. Andy confirmed that there is very little difference in the battery resistance which just means that the batteries are of a similar condition. It is possible that the each BMS will also have slightly different, but I think that will have very little impact on the charge times. Is it possible that the actual internal chemistry of each of the battery type on each shelf is slightly different even though the batteries are charging with similar charge current, the time it will take to charge the differnet battery types will be different. Using the same argument, each battery type will also discharge differently which will then impact the charge time required to fully charge the batteries again.
I don't think that this impedance measurement can be used to explain anything. As you mentioned: it uses a 1 kHz signal to evaluate the impedance. Your BMSs will have some capacitors across the rails and therefore disturb the measurement. Finally, it is the resistance at DC, which steers the current, not an impedance at 1 kHz.
I'm almost sure that the problem arose because of BMS Because every BMS measures current based on the voltage drop on both ends of a resistor (that is, within the internal design of the BMS, resistors of a small value are placed to measure the current). It seems that JK BMS uses resistors to measure current that have a lower value than other companies A while ago, I had a similar problem, and the reason was the internal resistance of the multimeter (for measuring current). You can confirm that this is actually the cause of the problem by placing the JK BMS in the lower or upper battery pack and observing the results
Hi Andy, To me, the high frequency resistance is simply not a good indicator of the battery charging rates. There may be low/mid frequency resistances within the cells that you cannot assess this way.
That is the industry standard. All battery manufacturers are using the 1kHz signal to determine the Ri of their cells. The only thing could be capacitors inside the BMSes which mess up the result...
I agree that the BMS can play a part in the result; and as you said, manufacturers use the 1 kHz signal to determine the Ri of (single) cells, so the test would be more meaningful if it were carried out on individual cells (no BMS, no - possibly heterogeneous - cells in series... ). What you could also do (without taking everything apart!) is measure the Ri at several stages of charge/discharge. The results could be surprising ;). I'd be interested to know.@@OffGridGarageAustralia
Hey Andy, Happy New Year. Would you believe it, but I've just ordered SIX Seplos Mason 15kWh batteries for an electric catamaran I'm completing! I'm thinking of powering it with four EPropulsion Navy 6.0 motors. It looked like your boat maybe had an EPropulsion motor too
Hi i would like to make a battery pack for my house and i need your help! I have 169piece li-ion(3.3-4.0v) ~40ah battery's and i would like to make them as 13s12p system! 1 modul would be 13series ~48v I will have an MPP Solar PIP 58048 MKX inverter for charging! 1. Which bms should i use? (i need 12 parallel strings and later more and more) 2. How to make the communication with 12 bms? (or do i need to communicate between the inverter and the 12bms) Thank's László!
Very hard to tell. As this is a new system you need to make your own design. Think about the maximum load you will have on your system now and in the future. How much solar, how big is the inverter? Do you want communication? So a lot of data needs to go into your system design first. If you have 12 parallel batteries with 12 cells in series, you need 12 BMS. You can never go wrong with the JK-BMS in my opinion.
Thanks i thinks the Same. 5kw off grid inverter and 5-6kw solar panel. Max load is 32a one phase, but never will max at the same time i think. Jk 0.6a or the new 2a? Thanks
The jk battery was unbalanced, so only 1 cell was full, raising the voltage. This happened to me when put toguether a battery with 15 cells from one batch and 1 from another that was quite a lot more charged..
I think the problem comes from the chemistry of the cells, they are not the same brand, therefore different manufacturing, the chemistry must vary. It would be necessary to make a charge/discharge curve of one of the cells of each pack and check if the absorption duration varies between the cells.
Andy look to your Mosfets - Not all Mosfets are equal - BMS with Mosfets have different forward resistances and/or different numbers of Mosfets in parallel will have different resistances - older cheaper Mosfets tend to have higher forward resistant which get worse as the heat up - a crude test would be to check which BMS get hot under similar high loads - Hugs from UK
I agree with Trevor. Also, the on resistance of Mosfets is non-linear over current. In other words, you will get different resistances at different currents.
@@marklockwood7592 That's so right- there have been a lot of advances in Mosfet' design and material purity in recent years - I imagine not all manufactures have caught up - some e-bikers have pepped up their bikes by the simple expedient of changing the Mosfets in their controllers.
@@marklockwood7592 Intersting. Maybe use three identical BMSs as proposed by @stefanvanzyl9090 to see how things change.
I would think that he could swap bms to see if the condition migrates with it. That might shed some light on the issue.
@@wg6215 Sounds like a plan
I've got to say, these JK Inverter BMS are nice, I've got 3 running in parallel right now, soon to be 5. Andy, you have helped me a lot in understanding these BMS and how to connect them and get them working right, thank you from sunny Nevada USA.
I’m doing something similar but 7s 24v 280AH my BMS is a Seeplos 3.0 7s 24v 150A I have a stand - up battery case ! You are 1 of the top people on earth in knowledge about this !
First, please notice the charge voltage and the voltage at the battery
At 8:09 they match for each battery (approx), - the 55.2V to 54.8V offset is due to the bus-bar-to-battery drop at 80A?
V1 V2 V3
Batt V 54.84 54.82 54.8
I think it is due to the balancer, the middle shelf will have a more even cell capacity and so they all reach 100% at about the same time, meaning it can absorb more energy, the other two shelves will be more imbalanced so once ONE cell is at 100% the overall current will drop because of them all being in series, even though the other cells may only be at 95% or so.
Good to see you again Andy!
Glad you had a nice break with the family 😊
Andy, I believe that the middle battery pack is better balanced than the other two, which causes it to have a slightly higher resting voltage than the other two banks. When you slowly discharge them, the middle pack then discharges more as it has a SLIGHTLY higher voltage overall for the first few minutes/ Ah discharged
When you do a higher discharge rate, the overall voltages of each pack are dragged down into the flatter portions of the curve, which will then make the battery packs all supply similar current, hiding the higher resting voltage.
Since you are working with 4 sets of variables here: cell size (304 vs 280), cell busbars (CU vs AL), bmses and the effects of longer cabling for the lower packs, I propose that you fit 3x identical BMSes for 1-2 weeks - preferably connecting them to your raspberry Pi / VRM to allow for logging and totalizing - This is "easy" since you already have all these new JKbmses which were bought at the same time and can run the same firmware.
My hypothesis is that the higher trailing charge rate for the middle pack will disappear after 3-5 cycles
Agreed. Eliminate some variables, and my guess would be the same as yours.
Agree with this statement
The voltage difference at charging is the key give away..
Top and bottom pack have a higher pack voltage when charging, there for are less balance.
Furthermore, as you paralleled 3 packs, once the difference between the packs becomes more than 2 mv, the batteries will try to balance out ( from my own exp.)
@@houseofancients It looks like the top and bottom batteries have a lower delta voltage in the cells. at 9:58 the top and bottom battery have a delta of 25mV and the middle battery has a delta of 62mV. Although that might be because of the higher current in that battery. I wonder if the battery cells were balanced if this issue would be solved.
Aaaannnnnd. Balancing is wasting some energy right? 😊 So little bit here and little bit there. My opinion is that with such count of variables this "problem" not worth another try/guess. It may still end up with a wrong conclusion. Thanks Andy for all the good work!
Have a look at cell 11 in the middle shelf. At 87% SoC it is already above 3.45V. I believe this single cell causes the internal resistance of the whole battery bank to be higher than the other banks, resulting in this "problem".
I have two Seplos banks (with NEEEEEEEY active balancers) where I observe similiar behaviour, when charging with 100-140A (50-70A per bank on average) cell 9 in bank 1 goes to much higher voltage than the others. When discharging at moderate Ampeages the voltages are fairly equal, but when discharging with >100A the same cell has the lowest voltage.
Highest voltage when charing and lowest when discharging means the weakest cell. I also observe that the bank with the weak cell charges slower than the other bank.
That’s my issues , great observation !
Happy New Year Andy!!! ...and from us in the Caribbean.... Happy Three Wise Men Day!!!
Happy New Year to the Caribbean!
Hi Andy Really enjoy your video info and style and may have partial answere to above glitch (not seen in comments). Following things I have learnedfrom LIPO ebike builds. First, assume things you can't change are valid/the same (cell chemisty, internal resistance, etc.). Next, make similar the controllable variables such as: cell temps, BMS model, BMS voltage calibration, lead length/size, cell SOC, etc. This would mean manual "balance" to actual full capacity or SOC, replacing vertical bus bars with equal length cables (maybe only for test) or at least connecting to bottom of one of them), measuring internal resistance at actual usage high current dc level vice ac impedance, replacing aluminum bars with same copper, and NO balancing during test. This should point to most of cause if ONE variable at a time changed. Or, since that would take a long time, just ignore it but keep in mind for next time 3 or more banks need to be in parallel. Can't find it right now but correct way to parallel 3 batteries/banks really looks weird (SLA but LFP also due to flat charge/discharge curve). Keep up the good work please and cold one sent., Dick
Thanks a lot, Dick!
I have some Jakiper JK48V100PRO batteries. With one of the packs, I observe some of the same as you do. That is the oldest pack and it has slightly different cells inside. Different model cells and possibly from a different manufacturer. Possibly with a slightly different charging curve? But as you said, it is not a problem.
Interesting Topic, ... as I have three 24v 280 Ah Eve battery banks (purchased one or two years apart), that all have the same 3 BMSs; with one battery bank charging or discharging at lower amps. I have suspected difference in individual battery cell chemistry and internal resistance; not sure. Was going to switch battery bank positions and keep bms positions while cleaning all bus bars and bms wire loom connects (someday), .. as a test for learning more all the time ... for later. Not immediate concern as my setup is working good. Look forward to your next discovery on this topic. 🐾🐾🎶🌠 15:26
OK, so, if your banks are from different age, they may charge a bit different. That will be quite normal. Calendar age takes its toll. But yeah, swapping the positions can mitigate that.
That is why i use different solar and charger array for each battery bank, and combine the battery banks through diodes so no power is transmitted from one battery bank to the other. Will be making some videos about my system soon. Amazing work Andy. I love your channel and your expertise. Keep it up 👍
The balance current from one bank to another is not a problem with LiFePO4. It is very small or even non existent. Depending on your system size, you can install separate chargers and arrays for each battery as you did.
With my system it is better with the diodes because the BMS's control solar charging for each battery bank but all BMS's control the inverter, so if one cell would fail without the diodes all power would go to the failing cell and fire/explosion could erupt.
@@StarJackerSRL That will not happen. that's what the BMS is for. If one cell fails, the BMS will disconnect this bank from the system. The diode will not help.
@@OffGridGarageAustralia 123SmartBMS does not do that on it's own.
My BMS only controls the inverter and incoming charging from solar panels, that is why i use Victron diode battery combiner.
I have a similar situation between my 4 parallel batteries. Two of them are CALB329Ah and two of them are envision 306Ah and 308Ah. The Envision batteries seem to get full first and under certain conditions they can even backfeed energy into the CALBs. I assume it has to do with slightly different voltage curves And as you said it's not an issue in practice.
Thanks!
Thank you very much! Great support.
The explanation might be easier to explain using the charge curves of the batteries.
When we build these batteries we found that matching charge curves for the 16 cells is more important than the internal resistance of the cells (importance of using same manufactured batch).
What you have is a slightly different charge curves beween the batteries at the 3.35V-3.48V cell voltage range. Thus at the "knee" of the curve the middle battery is just a bit slower to get to the bend. It actually means that the cells used in that battery is better matched. The benefit is not huge but if you do the calculations the capacity would be 0.25-0.8% better than the bottom battery.
Yes, that is a very good point. If the charge curves are different. the banks will charge/discharge differently.
I have noticed the very same thing with my system.
In particular my 1 bank of 3P x 16s Identical cells (EVE 280's), identical BMS x 3 packs wired similar to yours.
All batteries were from a single batch, BMS's from same production run. Active balancers also same production run. All were installed and commissioned within days of each other.
In my case (and yours) cell capacity matching of the banks is the only difference. The best "capacity matched" bank, takes the longest to discharge and recharge as compared with the others. In your situation, you could also add the small differences within your different bms's.
Thanks for sharing. I have also the aluminium bus bars, so more variables to play with. I really don't think it's the BMS...
Sou do Brasil tenho um sistema off grid e adoro seus vídeos parabéns !
Glad you had a great family Christmas Andy, you deserved it. 👍
Happy 2024, and have great year.
Interesting… I’m still watching the video but my speculation: maybe different BMSs are not such a great idea??
I think is the BMS not the batteries, I think the JK will get more current when you are almost fully charged (say 55.0). But we will see that in your next video. Thank you for doing all the hard work for us !
Thank you Marco!
Measure the voltage drop across the positive connections and then the negative connections (input to output) of each bms with the same current flowing through each to see if the BMS is causing the issue (maybe the FETs have a higher forward bias)
Happy new year Andy and community. Can you also measure the battery bank itself, this way you have your bank and BMS resistance. There could be a difference. Maybe also validate with the same clampmeter what the real current and voltage is, set the BMSes accordingly. Keep up the good work this year! Thanks for all your efforts in debunking all myths and issues related to batteries, chargers etc🎉
Thanks Edward and Happy New Year to you as well.
Yes, some have mentioned that here that the BMS could interfere with the measurement. I'll do more testing.
The one symptom I noticed was the middle shelf having a slightly lower voltage than the other two when charging above 3.4V, which explains why it charges for longer.
But as to why it has a lower state of charge, I guess you'd need to look at each variable in isolation, such as swapping out the cell busbars or the BMS with one of the other batteries etc.
There could be a small difference in measuring the voltage across the three different BMS. Some here have seen a slightly higher voltage, you see a smaller one...
I have more insight in one of the next videos...
Just to the side of the discussion. I use the YR1035 to quickly test the internal resistance of LiFePo4 cells. I then tested the same cells in a combination for battery banks via the EBC-A20 PC software and finally the same application with the EBC-A40L. All these tests were performed on dozens of Winston 60Ah cells and 16x EVE 304 Ah cells. All the Ohm values were completely different from each other even for single cell. Often even orders of magnitude different. You perform this task with the YR1035 for all three banks, so the implication of the instrument will most likely be very similar for all measurements. However, please remember that the real internal resistances will MOSTLY be quite far away from the results you define.👋Anyway Thank you.
Yeah, if you use different devices to measure the resistance, you will most likely get different results.
How long until the middle bank catches up? I can't imagine it being more than a minute or so? Do you see the same if you charge to 3.5V? If you do, how long does it take to equal out now?
Thank you for sunny Australia.
Always sunny down here... almost!
Hi Andy! I have been figuring out the daisy chaining batteries and the different voltage losses in those parallel connecting cables. If you think about battery voltage difference with almost empty and almost full battery bank. That voltage difference is much more than the voltage difference caused by the non equal battery cables. So the small voltage difference caused by the cables is not so big deal. The batteries will become almost empty and almost full :)
How long does the middle takes longer? How much energy more or less or equal? Complete curve of middle and bottom in direct compatison?
I would say around 10min until it catches up with the other two banks. At an average charge current of ~25A that could be 5Ah. Seems excessive to me...
I don't really have a device which can measure the charge curve of a whole bank... wait.... maybe...💡
I think the different charging current is as a result of the availability of internal active balancer in the JK BMS.
All the cells in the JK BMS bank are balanced and takes the all the energy they can store.
I suggest you verify the overall Watts hour stored in each bank and compare with the rated capacity.
Thank you for all the good work you are doing.
Sweet setup! Your garage is the electricity and free solar energy Valhalla. So I bought a BMS active balancer that doesn't have any visual or bluetooth notifications. I connected it for a few days and think it worked bc the cell voltage difference is minimal. Do you suggest leaving it on? Doesn't it interfere with mppt or inverters/buck converters..?
Hi Andy, try to measure the voltage drop across the battery bank BMS during discharging and charging. Maybe the BMS internal resistance changes/different during discharging and charging.🙂
also measure the voltage drop across each bms at high current level and see if it has high deviation berween 3 bms (at same current)
Andy regarding alu or copper the thickness is important but also the contact area and material of the bolts count. I have a similar situation, need to find time to investigate since it stress more one of the batteries.
Well, the contact area is not too important if the torque is sufficient. Whe have tested this here and could push over 230A through the tiny contact area of terminal and bus bar: th-cam.com/video/VPQvs-2YYzA/w-d-xo.html
And I'm still running this exact setup in the top and bottom battery.
Hey Andy, Happy New Year! Just wanted to ask about the 3D printer update and if you share any of your STL files you designed or find useful?
Always enjoy your videos and congratz on the boat!
Yes, I will share all the STL files. I haven't done any printing since but will be designing some stuff in the near future.
@@OffGridGarageAustralia Thank You!! cant wait :)
Is the middle shelf the 280K and the bottom 280 not-K and of course the 304's are also non-K type. Just a subtle difference in electrolyte and therefore a slight change to the charge curve?
Like you say, its not causing problems, and other than swapping round BMS's or busbars eliminate those variables its the only difference other than capacity - unless the middle shelf were B-grade, or a different factory or batch is maybe all it takes to shift the charge curve a little.
I love this Battery bank Self !
Hi Andi, das JK misst eine niedrigere Voltzahl, während die anderen beiden bereits die 55,2 Volt erreicht haben. Man kann doch das JK von Hand kalibrieren. Wenn man das JK bei beispielsweise 54,7 Volt auf 55 Volt kalibriert, dann wird es das erste BMS sein, welches die 55,2 Volt erreicht und auch als erstes runterregeln. Sind es diese kleinen Zusatzboxen, die die BMS mit dem Cerbo verbinden, welche für das Herunterregeln verantwortlich sind? Normalerweise sind doch die Solar-Charge-Controller für das Herunterregeln zuständig und nicht das BMS? Das hat sich doch erst mit den kleinen Boxen geändert?.
The BMS does not control the current flow. It's on or off. The voltage is what it is, independent of what is displayed.
So, what is the title of the video that answers the question presented in this video? The next videos appear to discuss the "reset" issue with the JK Inverter BMS.
Merry Christmas and a happy hangover.
Did you zero the low ohm meter?
How repeatable is the zero measurement?
How repeatable are the resistance measurements?
How matched are all the cells?
I think the difference are caused by the measuring accuracy of the BMS. The BMS measures the voltage and decides if the battery is full or not. You have determined that this happens at 55.2 volts. But each BMS has its own accuracy. We are talking about small differences, as you note yoursel, so a measurement accuracy of 0,1% could mean that one BMS trips in reality at 55.25Volt while the other trips at 55.14Volt.
The BMS only measures the voltage but does not control or regulate the charging. Even if one BMS shows only 50V, the battery itself will charge exactly the same. (Obviously, the OVP of the BMS will not work correctly ins such a case)
First see if it follows the BMS when swapped. If not, I would set up a high current flow and measure for voltage drop in the milliamp range on both the positive and negative cable paths of the middle battery. That is how I found a weak connection on my solar system.
Hi Andy, what if the problem was the real capacity of each battery?
maybe as the years go by they have different abilities that bring you this problem.
I don't think the problem is the BMS.
Hi,
the different impedeance would cancel out during an discharge/charge cycle, but you have not taken into account:
a) if discahrged, (and then leaving idle during the night) the banks will equalice by voltage and so charge the bank witch was discahrged most du to lower resistance.
b) your top bank has higher capacity, and so lower resistance provinding more power
Could it be that because of the lower total resistance, bms included, of the center bank more ah were removed and therefor mire AH were required to charge the cell? The voltage was lower during charge, after all.
More testing needed. When I setup the Peter boards, I found something interesting...
Can't wait to hear about what you have found.
My daly banks and the JK banks give similar results
Seems like that middle shelf voltage is always a .1-2v lower. Need to check voltage drop of bms and circuit breakers under load.
I guess the Battery Shelf 2022 might deserve its own video about the performance and cycles in the past 2 years I guess.
And then it might also deserve a protection well known as DENKMALSCHUTZ cause it is so iconic cause everyone can recognice the long and winding road back then.
And it is clean and nice - compared to the collection of batteries that appeared seconds later.
After a long and winding road with EVE LF160 cells grade B we finally got a set of 16 x EVE LF280K grade A from a german electrician shop in Hannover.
Grade A means in this case: all produced on tthe same day same production site and a row of serial numbers with a 305 Ah result for each cell over 3 single discharging test with EBC A40L in a heated room with 25°C according to EVE 's test procedure described in their 20 page long description.
305 Ah means they are also running quite close from top to bottom with the Seplos BMS balancing close to nothing cause max deviation is about 0,002 V but only 2 are better cause 13 are 0,001 .
And that is most likely not a cell difference but caused by the position in the battery cause when I mixed it up the results were not the same and the worst became the best. When I put it back in the previous order I got the same result as before, this time again the worst cell in the battery, whicch means from best to worst it is just a question where the cell sits in the battery cell chain. And the top balancing is a rather short procedure till its done.
This is a complete different result and world compared to the LF160 Grade B cells (have been lasered and produced within about 7 months).
Did you calibrate the Voltage measurement of the JK BMS? There is an option in the settings to calibrate it. It could be your just thinking it is not fully charged?
Yes, it's calibrated.
The difference may be in the bmses themselves. It would be worth checking the impedence across the bms alone, then the battery banks alone.
good afternoon andy andy question, I'm a little confused about the communication via can bus between the cerbo gx and the new jk bms, you have to modify the internet cable or you can use an original one without modifying it, that's my question, I'm confused and thank you in advance, andy
Yes, you can either make you own cable or buy the overpriced Victron Type A cable. Both will work.
Andy, good afternoon, thank you for the response, but do you have a diagram on how to make the cable? I would appreciate it a lot.@@OffGridGarageAustralia
Different BMS, different circuitry. I have 4 batteries and made them with identical cable lengths, cells and BMS, I do have some variation between the batteries during use, but only minimal - up to 5% of the charged status maximum.
Sorry that sounded a bit rude, I'll try again....
Hi Andy,
I think it could be Calibration of the different BMS's If you check voltage at each battery bank voltage will be same. but JK BMS is showing lower hence will keep charging as it thinks it's at a lower voltage.
I have a 5 bank that did the same thing.
And keep up the great work Best channel on here.
Hi there,
but the actual voltage on all BMSs is the same - and non of the BMSs is limiting the current or stops charging the battery. The measurements of the BMS does not matter at all.
What do you think about that?
Greetings
What about the individual battery isolators’ contact resistances? You might want to test them also.
Hallo Andy, ich weiß nicht ob ich es abschließend richtig verstanden hab, aber den Stromfluß zu deinen Batteriebänken bestimmt grundsätzlich mal der Spannungsunterschied. Die Zellen haben beim Laden eine bestimmte Spannung erreicht (z.B.55,0V). Der Unterschied oder die Differenz zur eingstellten Ladespannung (55,2V) beträgt nur noch 0,2V. Wären jetzt alle Innnenwiederstände gleich, würde überall der gleiche Strom fließen. ABER, die Spannungen der Zellen (Bänke) sind unterschiedlich, verrechnet mit dem jeweiligen Innenwiederstand bekommt man einen anderen Stromfluß bis zur vollen Aufladung. Am Ende der Aufladung haben alle Bänke die gleiche Spannung. Aber bis dahin ist das Verhältnis aus unterschiedlichem Innenwiderstand und Spannungsunterschied der Bänke mit der Ladespannung entscheidend. Nur ein Spannungsunterschied treibt den Strom durch einen Widerstand. I=U/R also "Spannungsunterschied, geteilt durch den Innenwiderstand = der aktuelle Strom" oder 0,2V/0,011Ohm=18A, was deinen Beobachtungen nahe kommt. Kannst gern auf deine Batteriebänke umrechnen. VG Jan
When you swap out the bmses, please redo the impedance measurement to check for the influence of the bmses!
Yes, absolutely!
I don’t know if one of the battery cables has one more strand of wire or a thickness and size of the aluminum bars. But I will check out a cold beer 🍺 and get back with you 😊
Hi Andy, Switching bms could be a good test to see what is happening.
My 2 cent is simply cell. You don't have 3 batteries but 48. Over 48 cells, they are suffering differently over the time.
As the charging voltage applied to the shelf is the same for the three banks, for more current to flow through the middle bank its resistance must be lower. Using ohms law actually calculate it rather than dicking around trying to measure it. Use your amp meter to confirm readings and Swap the shunts around after checking all connections to confirm that there is no variation between them..
Dear Andy, what I found in my shelf, that when charging with Jk weeker cell goes higher voltage, then active balancer takes out energy, but when current go's down at 55.2 this weeker cell starts to use more ampers and, and active balancer put back energy to weeke for absorption. While pasive balancer was not able to make so big trouble for weeker cell.
So, my prediction is that please change active balance current from 2A to 0.1A and look for a few days. You will be surprised that weeker current is enough.
I have the same but i am using 3 identical jk bms. But my cells have a different manufaktorer. My CATL cells still take Power even if the eve and the rept stopped.
My take is that the middle battery has more losses for some reason. Hence, when discharging it is discharging a bit more than the other two, and it has to recover more energy on its charge cycle.
I’m a bit puzzled: By not measuring the impedance from each busbar across the cables into each MCB, through each MCB, and across each battery, you still have 3 unknowns (at least, unmeasureds) in the system-the Positive busbar to MCB positive output and negative MCB output to negative busbar for each of the 3 batteries. This might explain higher resistance and lower voltage of the middle shelf battery.
A second source of variation in resistance could be your busbar connection configuration. I’d have to go back and see the video(s), but I think you did use a diagonal config and kept the MCB to busbar positive cables equal length (to each other) and negative cables equal length (to each other). In theory, each of the unmeasured cable/busbar cumulative resistance should be the same if you used diagonal wiring and had same distance between each battery cable.
Unmeasured Sources of variance could be torques, washers, crimps, different wire chemistries, corrosion, and-outside the wiring-the MCBs when closed.
If, after you have done the swapping of the BMS’s (I’m assuming YT production times and you’ve already done the swapping), and the slower charge/load rate doesn’t follow the middle BMS, you might look at these things which are currently assumed.
Happy NewYears!
Hey Andy,
when you switch the BMSs, you should check the Resistance of the Breakers too (measured after the Main Breaker).
It would be also interessting if the effect can be observed if you charge to a higher voltage. I think it could also be a slight difference in chemistry of the cells.
Looking forward to see what you can figure out there!
🐸
Yes, breaker could play a role but I doubt that would make such a huge impact.
Charging to a higher voltage shows the same with the middle battery still charging longer.
In my case the 100Ah set says "battery fully charged" at 55V and disconnects charging. EVEs of 300+Ah stabilize at 55V and don't absorb anymore, while 300Ah+ CATLs get to 55.2. Maybe something to do with cable lengths also? Cable lengths are pretty random between ~25cm and 2m. I'm thinking of a small V-drop across them.
The higher the current the higher the battery temperature and lower the internal resistance. Lower the resistance, current gets even higher. So one battery always wins the charging race. Adding series resistance (thinner cables) will help to balance the charging rates.
Why is there only 91k subscribers? He is super informative!
We seem to have a lack of solar and battery nerds that enjoy the Australian/German flair of dear Andy.. I saw a channel with a female who is quite a bumbler when it comes to putting batteries together as she has had sparks flying and was giving poor advice but yet had a huge number of subscribers.. Maybe Andy needs to get his beautiful wife more in the videos as this would seem to boost subscriber counts?? His advice is backed up with first person testing and he has grown with this knowledge over the past several years IMHO to be an authority/expert from doing..
hi, would you be able to do an experiment with a 15 cell battery (like a pylontech or home made) and a 16 cell with a pacebms. I have 80kW with the Pace (8 batteries) and 37kW with pylontech (11 x us2500C); what happens for me is: when both are connected, only the pace discharges unless the load is extreme (like 10kW, then the pylontech will take a very small % of the load, maybe 5%), If i turn the Pace off and allow the Pylon to discharge as its the only bank, when i turn the pace back on, it dumps whatever it has into the pylon to take that back to 100%. the batteries are connected as follows: 6 pylon, 4 pace to a lynx power in, 5 pylong 4 pace to a second lynx power in, then a lynx shunt , then a lynx distributor which has 3 inverters (voltronic 11kW) the shunt is the aggregator for the battery stats, with the pace connected via can to a cerbo and the pylontech connected to solar assistant for data, i also use SA to monitor the inverters. any insight appreciated.
You cannot parallel 15s and 16s batteries. That will never work! They have to stay completely separately.
Hi Andy , hace meses te sugerí usar una misma marca de BMS , solo estás haciendo que algunas baterías trabajen más duro que otras 😅
Hi Andy, I have the same question as you! I connected two battery packs, each with 16s 280Ah, using the same BMS (which don't communicate with each other), cables of the same length, an equipotential bar, and the same brand of cells. The only difference is the production date of the cells. I also tested the resistance by measuring voltage drop. But nothing! I just can't figure out why the cells are behaving so strangely. Sometimes they seem perfectly synchronized, other times one discharges much more than the other. I think there are many factors at play, including cell temperature. However, it seems more like a theoretical question because, in practice, the system works. Overall, the cells seem to work similarly over time, so I assume aging will be similar. Thanks and greetings from Italy!
I have EVE LF280k from different production lines and only a few days apart. They have in average around 4Ah difference in capacity, but they have diffrent charge curves. Seems there are big tolerances in materials and manufactoring.
The charge curve could be what is different here.
Even having a 100Ah and 280Ah in parallel did not show such a result as the 280Ah bank always charges/discharges faster and they meet again at the end.
@@OffGridGarageAustralia In reality, the behavior seems completely random, as batteries often reverse roles. Is it not related to the quality of the cells? It appears to be linked to random factors.
Thanks Andy
Hi Andy i have a Question to the JK BMS. is it possible to Update the Firmware also on the B1A8S10P?
Nope, not possible.
Hello Andy, and happy new year....
Just a down side to things......here in UK ....our days are getting longer and yours are getting shorter......ha ha
(Don't pay it back in 6mths)
I take it all your BMS have same settings, but as you said in last minute of video.....
You have different BMSs and all 3 are working on 3 different banks....
Ok top back is 305ah
Bottom 2 are same but with different bus bars and different BMSs.......for get the top one.
Now the lower 2..... different BMSs, different bus bars , different temperatures (due to hight day night thing)
Different age.
I'm thinking it be something simple.... forget data thing......also middle one is first of the 280 ah cells connected to dicarcarge point???
Is it as simple as this bank is next low ah bank and so discharge a little more everyday than bottom bank????
I know you have built a fantastic battery bank and its well perfectly well laid out.....
You do have a Mish mash of cells BMS and bus bars.....
If it was all the same.....then I would ask the question as you have.
I'm a fitter on plant equipment......you can have two machines.....but both are not alike...
Never over look the basics.....forget data.
I don't think it will help, but thats how end up fixing faults.
Sorry for going on Andy.
Look fwd to out come.
Hello Andy! Happy New year ! Can you install a Victron Battery Balancer? Would that help?
That would not help as the batteries are in parallel all the time anyway,
doesnt the middle one have active balancer? could it be using more energy ?
Yes, it has an active balancer. It uses a lot less energy as it 'recycles' the energy rather than burning it off as in a passive balancer which the other two BMS have.
It is because of the active balancer, it is lowering yhe voltage when balancing this causing more amps to flow.
The passive balancer will do the same. And even worse than the active balancer.
I think manufacturers have varying charging curves so your results don't seem too odd. Would be more odd if you were using the same BMS brand across all three batteries with mirrored settings. Are all of your BMSs set to the correct chemistry?
please test us some NA-ion 220Ah batteries! PLEEASEEE
happy new year😎
To you as well my friend!
My goodness.
Battery current diverging during charging as each battery comes up to fully charged at a different time due to variations in manufacture, wiring etc etc, leaving the one with a lower relative SOC to soak up potentially undesirable levels of current in a system that does not monitor individual battery current? That can't be right as it's been stated a few times in the past by the channel operator that he's checked and his batteries don't have this problem. I do remember some guy banging on about this in the past and that seemingly cluey guy also said that it happens to pretty much every battery system out there which is why the commercial setups monitor this sort of thing. We're not quite there though as apparently it's not really a problem, but give it a couple of years and we'll see how the story evolves.
It is true what they say. You can lead a horse to water but you can't make it drink. You just have to leave the horse there and hope it eventually works it out for itself.
Yeah swap bms would tell if bms is the cause.
But i guess it is because of the active bms method.
Maybe it will waste more energy than passive and take longer to fully charge. 😅
passive balancing discharges higher voltage cells via resistors - throwing away energy as heat.
Active balancing discharges a higher voltage cell into a capacitor, and then uses the stored energy to charge up a lower cell - less heat overall
@@stefanvanzyl9090in theory, but maybe in practice the JK pulls more power over time?
@@LibertyDIY BMS power consumption by itself is in the order of a few milliamps. These packs are cycled daily and the behaviour is highly repeatable one day to the next -even a 20mA consumption/wastage difference over 24 hours is only a 0.48Ah difference, which would be replenished in less than a minute at this 40A charging rate difference
The only thing I can think of is there is a slightly higher self discharge on the middle battery. Is there a difference if the time between full charges is long or short?
Since the start of summer, these batteries see full charge almost every day
Smart Wife. More on the Boat please.
It's on the other channel. Video will be coming out Monday 5am AEST 😉
Andy can you notice or measure any degradation to the batteries for the 2 years of use? Maybe determine true life expectancy?
I made a video about it here: th-cam.com/video/GFuY8MrDAfA/w-d-xo.html
My heltec bms screen shows like your jk-bms. I wonder why ? Referring to toe video appearance ! It identified my heltec bms as JK_BD6A17S8P . Is Heltec using jk-bms software ?
No, not at all. The Heltec uses its own software.
Voltage Current Calibration of the BMS! I use 3 Pace BMS after calibration oft the BMS on 3,65 Voltage everything is spot on! Also The Neey Balancer can be calibrated… I tested 10 Neeys on 1 Battery at 58,4V they are all slightly off 😉 after calibration everything is good… balance Leeds calibrated is also possible!
But this would not explain, why the middle batterie charges longer if the BMS is slightly uncalibrated.
Would measuring and recording the voltages across each individual bank with an accurate DMM while simultaneously recording the amperages, either charging or discharging, allow you to use ohms law to determine the overall resistance of each bank? I imagine each bank uses different sensors to evaluate the amperage going in and out of the battery banks? There might be some significant variance between them as well, particularly at high or low charging rates. Their response may not be as linear as one would like. It would also be interesting to check the voltage differences at different locations on the bus bar - at high amperages, even small resistances could cause significant, and therefore measurable, voltage drops.
Did you ever get an answer to this? I was thinking the "Profile" that each BMS has in the firmware.
I always thought that the charge termination curves engineered into the different BMS brands would certainly have an effect on the tail end of the charging.. I have two identical 17 cell 280a EVE batteries with identical (almost) JK 2A24S20P BMS controllers with the only difference being the few months difference in manufacture date and a slight difference in firmware versions.. One of the two banks always seem to lag the other when charging and even at idle where one bank should trickle into the other and eventually equalize but they don't.. So my solution was to slightly raise the full charge cut off voltage point of the trailing battery a bit higher than the leading battery and this seemed to bring them much closer in line with each other.. I charge at 3.51vdc have the full set to 3.5vdc and the float at 3.45vdc per cell..
There's a calibration feature for both voltage and amperage. Wouldn't that make difference if the calibration was off? For instance, voltage from one BMS is 56 for pack x, but the other BMS would show 54 from the same pack. If that happened, then the behavior would be different... Maybe?
Hi Andy, could you or one of your subscribers please help, I'm a newby to battery's and solar but wanted to give it a go...I have 4x 12V 120ah lithium battery's and with my 10a lithium charger its take's them up to 14.6 volts when charging but after awhile they settle at 13.3v...is the 14.6 doing any damage to them...also what voltage should I not go past when discharging them...thank you so much for any advice....Steve from Perth WA
Hello Steve🙋♂️
14.6V is pretty much the maximum these batteries can be charged to. As there are 4 cells in each battery, it means, each cell will be charged to 3.65V which is again the maximum these cells can handle. The battery management system (BMS) in these batteries will turn off charging if one of the cells goes over this 3.65V. So, it is almost impossible to charge to exactly 14.6V because the cells are never 100% balanced.
Recommended is a more conservative charge voltage of 13.8V-14.2V for these batteries. This will still fully charge the batteries but is more gentle to the cells.
That the batteries are settling after you take the charger away is perfectly normal. They are still 100% charged.
@@OffGridGarageAustralia Thank you so much Andy
It looks like its a combination of the higher resistance (so harder to charge) and the active balancer (which is constantly bleeding off some current 24 hours a day) means the middle pack is effectively getting less into it. Because they are in parallel, you can't really "see this" on the voltage as the voltage will always automatically equalise between all 3 packs - but it appears it manifests as strange but repeatable differences in current? Very interesting and to be honest it doesnt really matter as long as all packs do charge to 100% and all cells are kept equal in each pack? Also I don't think you can see internal resistance during discharge as the cells can effectively discharge instantly (if you short them)?
The active balancer is only working if one cell hits 3.45V. It does not work 24/7 as most other balancer will.
mines still working very good saving us power bills , after 13 months ,27v - 8 eve 280ah cells . Victron 1200va and 100/50 mppt in 24v mode ,battery is very efficient , only had that one issue with BMS crapping itself ,, love it free power free aircon
I have 4 PCS 3.2v 280AH bank 24v all though charging I set the charger to 5A for better absorption , “ the issue was I charged the bank to 14.6v = 3.65v so post 2 be but I have a 2 batteries 1 in 2 different banks each 1 of those batteries only charged to 3.37v and 1 battery in 2 different banks charged up to 4.0v and the 3.37v batteries stayed there ! What’s you idea “ Please “ ?
I think your guess is right. The middle battery get's fully charged last and the top and bottom absorb earlier. But it could also be the cells or the BMS voltage drop 🥴
I am going to suggest that changing the BMS on the middle shelf is not going to change the charging behaviour.
Andy confirmed that there is very little difference in the battery resistance which just means that the batteries are of a similar condition. It is possible that the each BMS will also have slightly different, but I think that will have very little impact on the charge times.
Is it possible that the actual internal chemistry of each of the battery type on each shelf is slightly different even though the batteries are charging with similar charge current, the time it will take to charge the differnet battery types will be different.
Using the same argument, each battery type will also discharge differently which will then impact the charge time required to fully charge the batteries again.
I don't think it's the BMS either. My bet is on the bus bars and connection to the aluminium terminals of the cells...
I don't think that this impedance measurement can be used to explain anything. As you mentioned: it uses a 1 kHz signal to evaluate the impedance. Your BMSs will have some capacitors across the rails and therefore disturb the measurement. Finally, it is the resistance at DC, which steers the current, not an impedance at 1 kHz.
Good point. Thanks for sharing.
I'm almost sure that the problem arose because of BMS
Because every BMS measures current based on the voltage drop on both ends of a resistor (that is, within the internal design of the BMS, resistors of a small value are placed to measure the current).
It seems that JK BMS uses resistors to measure current that have a lower value than other companies
A while ago, I had a similar problem, and the reason was the internal resistance of the multimeter (for measuring current).
You can confirm that this is actually the cause of the problem by placing the JK BMS in the lower or upper battery pack and observing the results
It's because the three bms's are fighting each other. For power. ❤❤
Hi Andy,
To me, the high frequency resistance is simply not a good indicator of the battery charging rates. There may be low/mid frequency resistances within the cells that you cannot assess this way.
That is the industry standard. All battery manufacturers are using the 1kHz signal to determine the Ri of their cells.
The only thing could be capacitors inside the BMSes which mess up the result...
I agree that the BMS can play a part in the result; and as you said, manufacturers use the 1 kHz signal to determine the Ri of (single) cells, so the test would be more meaningful if it were carried out on individual cells (no BMS, no - possibly heterogeneous - cells in series... ). What you could also do (without taking everything apart!) is measure the Ri at several stages of charge/discharge. The results could be surprising ;).
I'd be interested to know.@@OffGridGarageAustralia
Hey Andy, Happy New Year. Would you believe it, but I've just ordered SIX Seplos Mason 15kWh batteries for an electric catamaran I'm completing! I'm thinking of powering it with four EPropulsion Navy 6.0 motors. It looked like your boat maybe had an EPropulsion motor too
Hi i would like to make a battery pack for my house and i need your help!
I have 169piece li-ion(3.3-4.0v) ~40ah battery's and i would like to make them as 13s12p system!
1 modul would be 13series ~48v
I will have an MPP Solar PIP 58048 MKX inverter for charging!
1. Which bms should i use? (i need 12 parallel strings and later more and more)
2. How to make the communication with 12 bms? (or do i need to communicate between the inverter and the 12bms)
Thank's László!
Very hard to tell. As this is a new system you need to make your own design. Think about the maximum load you will have on your system now and in the future. How much solar, how big is the inverter? Do you want communication? So a lot of data needs to go into your system design first.
If you have 12 parallel batteries with 12 cells in series, you need 12 BMS. You can never go wrong with the JK-BMS in my opinion.
Thanks i thinks the Same. 5kw off grid inverter and 5-6kw solar panel. Max load is 32a one phase, but never will max at the same time i think. Jk 0.6a or the new 2a?
Thanks
The jk battery was unbalanced, so only 1 cell was full, raising the voltage. This happened to me when put toguether a battery with 15 cells from one batch and 1 from another that was quite a lot more charged..