Yixiang Batterie Consistency Test and how to identify problem cells. Wait! Have we just found some?

[@OffGridGarageAustralia]

Btw, the (B) showing in the QR code scanner means LiFePo4 chemistry. Developer confirmed.

[@JAN_DB]

I watched this episode with great interest as I have similar issue in a 16 cell pack. One cell is consistently raising quicker than the other 15. The 15 “normal” cells raise all within about 5mV difference till 3.45V but the other one raises quicker. The difference start to get really noticeable from about 3.35V. The main issue is that this one cell reaches the 3.45V before the others are fully charged. From there on the cells start active balancing but very often this process is not fully completed by the setting of the sun. I tried to “compensate” the drift of this one cell by starting the balancing earlier at about 3.36V instead of 3.45V (as per Andy’s advice). The question that raises is what energy loss (if any) is created by balancing this early. Looking forward to any return information !

[@OffGridGarageAustralia]

@@JAN_DB If it is only one cell, check all the busbar and balance lead connections. If OK, swap the battery with another one in the pack and monitor if the problem is with the cell or location. 3.36V is far too early to start balancing. There is no real difference in cell voltage at this stage as we are still in the flat part of the charging curve.

[@JAN_DB]

Hi Andy. Many thanks for your reply ! I know/understand that 3.36V is way too early to start balancing (in the ideal world), but let me explain further why I programmed the BMS this way: this one cell spikes shortly after reaching 3.38V and runs towards 3.5V and even higher way before the other cells even moved a bit closer to 3.37V. This happens within minutes. So my idea was to put a “little break” on this one cell from the very first start it’s voltage starts rising. Unfortunately, even this early “tempering” doesn’t help a lot as the BMS shuts down the charging of the complete battery as this one cell charges very quickly to 3.45V (cut off voltage) and my other 15 cells are still at 3.6-3.7V. But I’ll take your advice and will switch this one cell to an other location and keep you informed. Edit: another question raises now… when changing cells I need to “decompress” the battery. They are mounted all 16 in a row with double compression bars (10mm screw threads on either side). Does it make a difference if I loosen and re-clamp them when fully charged or rather in low SOC ? Thanks again !

[@OffGridGarageAustralia]

@@JAN_DB Yeah, I understood why you want to start balancing early to catch this one cell. If you have a runner like this, swap it with another cell. Compression needs to be accordingly with the manufacturer specs at 20-40%SOC, never at full or empty. M10 is far too much compression if you ask me. 6x M6 rods only need 0.5Nm each to achieve that required 300kgf.

[@rcinfla9017]

Manufacturing screw up in matching the graphite anode and LFP cathode electrode thicknesses for proper electrode capacity ratio. There is normally some electrode ink printing thickness variation onto the conductive foil which is measured and matched up with an associated opposite electrode to yield proper charge/discharge capacity match between electrodes. Also, number of jelly roll wraps is adjusted to keep cell capacity near same for thin/thick electrodes tolerance variation. Usually, graphite neg anode is given a little extra spare capacity over LFP pos electrode. Quick voltage rise at top of charge is positive LFP electrode reaching full state of charge which happens abruptly. Normally the positive LFP electrode thickness determines the capacity of cell. The slow voltage rise happens when negative graphite is reaching full state of charge before the pos LFP electrode. Graphite has a slower transition when approaching full state of charge. It says cell is short on neg graphite electrode capacity. The graphite electrode potential vs. SoC also causes the characteristic voltage bumps at about 65% and 25% SoC for open circuit rested cell voltage which are more visible during charging compared to discharging. Graphite electrode potential is 0.5 to >1v at full discharge and gets close to zero volts at full charge. LFP positive electrode potential is almost perfectly flat until it gets close to fully charged or discharged (3.43vdc). Cell voltage is pos LFP electrode potential minus the neg graphite electrode potential. They are not 'A' grade cells. They are manufacturer rejects.

[@patrickvanneck]

Great testing again Andy ! The most important message to me from this is, that because of these differences, one should charge a LiFePO4 battery bank only to 3.40v/cell and discharge only to 3.10v/cell (I do 3.0 because I have more than 40A discharge current, which is lowering the voltage of the discharge curve for the same SOC). Beyond those limits, the cell voltages may start to differ too much within the bank and may create a false balancing pattern on the top end, and, an over discharge of certain cells at the bottom end This is also the main path for creating longevity of the battery: we do not want to use those risky areas on both ends of the curve. The little extra capacity that one could get out of that is just not worth the risk. Thank you again Andy ! All your testing has given me great insights and has helped me building a 32-cell battery that I am using for 2 years already now uninterrupted .... This battery cools down my wife all night every night. 😃❄

[@upnorthandpersonal]

I've seen this before with brand new cells before they have a single cycle on them, and has to do with the SEI (solid electrolyte interface) not having established (or not fully/correctly). Give them another couple of cycles and see if things improve.

[@korishan]

Agreed. Cell characteristics can change over the duration of the first few cycles as the electrolyte gets fully saturated into the substrates. Would be interesting seeing a revisit of these "problem" cells to see if multiple cycles change the end results. Also, as pointed in another comment, temperature plays a role as well. Considering it's winter in Aussie atm, how much temperature swing between tests were there? Could the odd tests been on colder hours compared to the others? 7 hour test windows is a long time and temps can change easily in an un-insulated room

[@OffGridGarageAustralia]

Yeah, OK, I'll build the battery and compress them. We will see if they behave after a few cycles.

[@hanskroeger6239]

​@@OffGridGarageAustralia Hi Andy, If you assemble a battery with those cells showing different charge curves an active balancer Like the JK BMS would get totally screwed up. This JK balancer requires nearly identical charge curves of the cells in series. I have multiple Tests showing, that the variation of the charge curve due to aging can get a battery more unbalanced than turning Off the Balancer!

[@john_in_phoenix]

I know the results can vary depending on the temperature and the tail current. As noted, you need 5 to 10 full cycles to get consistent results as well.

[@teardowndan5364]

Thought the same thing when he overlaid his re-test on top of his first run, end of charge curve already got quite a bit sharper.

[@launacorp]

I would compress the cell, charge/discharge zhem ~10 times and measure the internal resistance charged and cischarged to see if the cell has potentially to much air in it (possible first cell after changing the isolator roll or some other materials from the production). It could be possible that the cell changes after a few cycles under compression.

[@MMMM2MMMM2MMMM]

And charging and discharging will remove air out? What by charging an discharging would get the bubble(s) out? Temperature? If so, Andy could push a cell next to some cold SPATs and put it out in the sun after. If it is air, putting a cell on a vibrating plate would get the air out faster. An air pocket within the sheets? I have been reading somebody told the forum LiFePo cells need to be calibrated by multiple charge and discharge cycles. I have serious doubts about it. If so... putting cells on its side would potentially be more harmful to the cells, because these air pockets can not escape easily, right?

[@launacorp]

@@MMMM2MMMM2MMMM its the electons hat move to the cathode wich increases the size in atomic dimensions which will push the air out. Couse no vacuum on earth can be perfect. So the vac while producing the cells isn't too. And yes, operating the cells on the side is not the favorite way for the first couple of charges. There are two 1,5h vids from a guy named zerobrain who had a Li specialist for an interview in his studio. Watch it, then u know.

[@OffGridGarageAustralia]

Yes, we will see.. If these are A-grade cells, they should come clear after a few cycles.

[@geekmystique]

I have never seen proof here on youtube of bad cells getting better by charging cycles and/or compression. I feel that Ali scammers just ask you to do it since your purchase protection will run out - what kind of person has the equipment to do multiple cycles on a 280Ah cell in 3 days? (Well.. Andy.....)

[@clintoncoker6]

As someone who uses LFP batteries in New Zealand, where it's colder than sunny hot Australia, this effect on the voltage/SOC curve is due to temperature. As soon as you get down to 15°C or below, the batteries act as if they have more internal resistance i.e. the voltage will be higher (at lower temperatures) with a given charge current at a given SOC (and lower while discharging). If you repeat this test again at 25°C, I bet you will see a difference.

[@bu5hm4nn]

This!

[@moestrei]

Interesting and makes sense.

[@korishan]

This is pretty interesting. Wonder what the temperature variation of the garage was from day to day during the tests, considering that each test took about 7hours. That's a lot of time for temps to change in a non-insulated room 🤔🤔

[@john_in_phoenix]

This. The cells like the same temperature you do. Even above freezing, if you are cold, they are too. Don't try a 1C charge rate below 10 degrees C.

[@clintoncoker6]

@@john_in_phoenix My cells are in a low demand application, I charge them at 0.025C below 10°C! Hehe

[@9111logic]

Excellent Andy, thanks 🙏 for this demonstration. I am currently testing a full set of 314A cells so as soon as I'm done I shall also compare the curves. I had done that before but from the spreadsheet, I didn't realise you could overimpose the curves directly from the software.

[@ekolekol4389]

I would be interested to see the 16s pack charge, discharge, and balance etc with those curves

[@OffGridGarageAustralia]

Yeah, we will find out shortly!

[@danfitzpatrick4112]

Fascinating, and a great find! Certainly interesting.... It's our analytical minds that keep us young 😆. Always striving to learn and find answers is therapeutic my friend.

[@OffGridGarageAustralia]

Thanks a lot Dan!

[@john_in_phoenix]

If you want reliable and consistent test results, you need consistent test conditions. This is especially true for lower temperatures. First thing I noticed years ago testing in my garage during the winter, results change with the temperature.

[@nigelcharles511]

As nyloc nuts have resistance when tightening it will reduce the clamping force on the busbars for a given torque setting. Their smaller diameter provides less coverage across the busbar slot so a washer might be advisable.

[@NaamNatuurlijkniet]

Wouldn’t the same torq setting on a smaller surface increase pressure ? Also as the contact surface of the nut is smaller and so there is less friction between the nut and the bussbar ? How is the thread of those nuts in comparison to flangenuts ? I guess some current will flow from the thread of the studs, thru the nuts to the surface of the bussbarr, so less contact with the nut due to no flange and a perhaps smaller thread may increase resistance and thus create more heat at higher amps. I think that is why Seplos (used to) deliver extra washers for inbetween the studs and the bussbar, to compensate for the loss of contact surface in unround/oval holes. Flex bussbars may have round holes (I hope), so that would also be better.

[@OffGridGarageAustralia]

I don't think there is much current going through the stud itself and then into the nut and finally to the bus bar it holds down. 99.99% will go directly from the terminal to the busbar.

[@abid583ind]

​@@OffGridGarageAustraliacurrent takes all parts according to their resistance, no conductor is without current when it's between two different potentials

[@jogi_54]

The first question, I asked, as you started with the inconsistency was ??? did you measure as well the Tenperatur of the cells/air while the individuell tests ?? It is only an idea cares jogi

[@OffGridGarageAustralia]

All cells were roughly at the same temperature. All cells were sitting in the garage at the same temperature The Gobelpower cells were maybe slightly warmer but it should not affect it that much.

[@jogi_54]

@@OffGridGarageAustralia I mean, you didn't perform the tests at the same time in parallel - I think, day after day. In cologne, the average temperatur of the last 5 days had bee between 25°C and 18°C. DT of 7°C is not nothing...

[@Niftynexuss]

Just received 16 of these cells from the exact same batch as yours Andy , hope to see some more testing with these soon.

[@sjdtmv]

This is why having a good active balancer is critical, I have 16 batteries coming from Gobelpower they are MB31, just sailing down the east coast of Australia as I write, so hopefully see them in a couple of weeks

[@BenBrand]

How would an active balancer make all of the batteries have the same charge curve? What other advantage does an active balancer have over a standard one in this situation? If the capacities are the same, there's no reason to do extra balancing since by the end of the charge all the voltages are the same?

[@sjdtmv]

@@BenBrand I never said that an active balancer makes the battery the same charge curve. but it allows a longer charge time before the BMS stops the charge due to over voltage.

[@TimothyDovel-w1s]

Thanks!

[@OffGridGarageAustralia]

Thank you very much for your support 🙏

[@glencooke494]

Hey Andy, At 06:50 the scan of the first battery had a couple of values in brackets but did not come up on the export except in the QR code EVE Power (04Q) LiFePO4 (B) LF280K (76). Do you have (B) grade cells?

[@OffGridGarageAustralia]

They are supposed to be A grade cells. I'll ask the developer if the scanner picks that up. (B) could mean anything...

[@DivyangM]

Nothing to say, enjoyed the video. This is for the KZbin algorithm.

[@OffGridGarageAustralia]

Thanks a lot, appreciate it!

[@seymourpro6097]

Do I really want Nyloc (ish) nuts when they have to be torqued down and over torque could mean stripped threads??

[@natesdiysolar]

Hi Andy. Of the 2 i tested, they were exactly the same capacity... 301ah on the nose.. Matched or just luck.. who knows. Thanks for the bit about cut-off value.

[@OffGridGarageAustralia]

Thank you. Yeah, capacity is one thing but the charge/discharge behaviour is another...

[@urbanbeeman5402]

I would have to wonder how this would effect balancing at the top of the curve.

[@BenBrand]

Seems like it would depend on what your balancing start voltage was set to. If it was set to 3.6 then the cells got there at the same time. But if you set it to something like 3.5 or 3.45 then the cells with the gradual rising curve would hit that much sooner and start possibly unnecessarily balancing early. It does seem like even if the cells are fine operating that way you would want all of them to be that way. That way they would all start balancing early since a lot of times you don't want to have to charge them to 3.65 every time

[@PowerPaulAu]

I have seen this myself with a small handful of cells. I deemed them NQR so they did not go into regular PowerPaul batteries. I would say either you've been deceived about the grade, or there's been an error made with the grading. As well as the slow voltage increase, you will probably notice a very drawn out current decline if you charge to a low cut-off voltage instead of 14A. And you will almost certainly see an increase in temperature compared to other cells. This is probably best observed when you build it into a battery pack. The end result is a battery that needs to top balance pretty much every cycle, and that top balancing is slow and drawn out, so the battery with these cells should not be used for mission critical tasks.

[@OffGridGarageAustralia]

Thanks a lot, Paul. I appreciate your insight. I will build this battery soon and see what I can find out. It will certainly be interesting.

[@darrenorange2982]

Andy how do the new MB31 cells look in comparison they are the newest chemistry.

[@OffGridGarageAustralia]

They have a far higher capacity though, almost 330Ah. I have the curves on my website.

[@darrenorange2982]

@@OffGridGarageAustralia referring to the curve.

[@williamos1985]

what you are seeing are imperfections and deviations in all the used materials, hence the different stages of charging and discharging give you those curves. for example, 1 cell has 99% good materials, the other maybe 97%. which can cause 0.02 difference in the average voltages. but in the end follows the 0% and 100% settings/values of your charger

[@williamos1985]

thats why when almost fully charged, its good to balance. if you dont do that, that why cells different like these ones, go out of balance when used together. because the voltages strat drifting apart on different stages of charging/discharging, but in the end you need them at the same values in order to keep working together

[@OffGridGarageAustralia]

Well, we have never seen this before with any other cells... if these are true A-grade cells, it makes no sense to me.

[@AirIswhenibreathe]

I think, that two cells with the different curve will give the built in balancer of the jk an good exercise everytime the cells reaches 3.45v. And my OCD will give me a thought that the cells are not balanced everytime those cells reaches 3.45.

[@OffGridGarageAustralia]

Or maybe just once to top balance all cells and from there it will be OK? We will find out!

[@Vekselbass]

Hi Andy! It seems to me that this may be due to the end of charge current, perhaps it makes sense to check the charge to a 1A cut off current. Most likely the issue is the scale of the graph display on the time scale. When charging up to 1A, the time of the last charging stage will be longer, and accordingly the graph will increase more steeply.

[@OffGridGarageAustralia]

I've checked that as well. There is no difference, and the scale is exactly the same. Their voltage just lifts off a lot earlier than with the other cells.

[@Vekselbass]

@@OffGridGarageAustralia thank you!

[@NaamNatuurlijkniet]

Hi, Andy. That is something new indeed. What is strange is that even cells with almost the exact moment and place of production have these differences. As indicated below they may just be virgin new, but why then mix them with batteries already cycled a few times. When sold as a pack the difference in capacity would make for a lower capacity pack, as with any capacity differences within a single pack. Not sure what happens when balancing though. It will balance out eventually and during discharge the BMS may comeup with an early low cell voltage shutdown, while the pack voltage might still be in the clear. Also the balancer may try to balance 2 cells with a different chargecurve at some point but with the same capacity. Testing them all might give insight in what other surprises are in the shipment. When cycling them does not fix the issue the pack may only be used at a % of its rated of potential capacity. The question is why this is happening. Did you notice in the past a difference in the curves when charging at 1C for example? How to measure resistance under load ? Perhaps the voltage rises prematurely due to a higher resistance during charging (internal heat buildup) at some point. Or the electrolyte is contaminated which would suggest a fault during production. If these are original Eve cells, maybe Eve can comment on these curves and of why they do what they do. All in all, if all curves were the same you might just have a somewhat different chemistry in the cells. But they don’t, so what they delivered for building a pack is just bad. Finally you could check the tester with one of your other cells. The looks somewhat like a combined CP chargecurve . Anyway, this is new and this is where we can learn. Keep up the good work.

[@OffGridGarageAustralia]

Thank you. It could be that the reseller is doing only random capacity tests on a few cells to get an average reading of capacity of a batch. We will find out once the battery has been build.

[@chrisw1462]

Do you have a chart that shows how the internal resistance is affected by underwear color? Please link or post! 😀

[@740GLE]

That’s why I go commando when testing my cells!

[@urbanbeeman5402]

You would hope there would be alot of internal resistance LOL

[@OffGridGarageAustralia]

You need to find me on onlyfriends for that🤭😆

[@chrisw1462]

@@OffGridGarageAustralia ROFL.. umm.. that's okay.. I'll pass!!!

[@peterle987]

check internal resistance at end of charging (or in the middle where the spread occurs). It may change during load process because of chemistry issues. As you said yourself: those measurements are just a snapshot - and you onl did at 30% SOC as they arrive.

[@OffGridGarageAustralia]

Yeah, I will do that once the battery is build.

[@TimothyDovel-w1s]

I am looking for the lifePo QR battery code reader.

[@jackoneil3933]

Excellent tutorial Andy, thanks!

[@OffGridGarageAustralia]

Thanks for the feedback.

[@BenBrand]

I know it's a lot of time and work, but it seems like it would be worth it to test a few more of the cells

[@OffGridGarageAustralia]

I will build the battery first and see how they go.

[@tullgutten]

And also the whole graph looks like it's shifted to the left for cell 6 so i think they are just at different charge state. Might also just need compression. Don't think there is anything wrong with it but yes it's a little suspicious and could be fun to track how it goes over a long time. But please, parallel all then fully float charge for an extended time then test again:)

[@OffGridGarageAustralia]

The graphs start at the same voltage as explained in the video. No parallel balancing needed any more for LiFePO4 if you have a MS with active balancer.

[@NoelBarlau]

In case anyone is curious, I just finished an initial parallel top balance on 16x 306ah EVE cells at 10 amps constant current and it took from July 1 until July 19 and used 35kWh on the Kill-a-watt meter due to the inefficiency of this method. Starting voltage was 3.227 and ending voltage was 3.650.

[@BluBioBike]

Ouch. That's not nice for the cells. Low current, high voltage for a long time. You will probably have already damaged the new cells.

[@NoelBarlau]

@@BluBioBike Just like beekeeping, in any group of three individuals you'll get four different opinions. I've personally never heard of that possible quirk of LFP chemistry. Where did you get this information from?

[@BluBioBike]

@@NoelBarlau nordkyndesign: "charging-marine-lithium-battery-banks" there is the topic is pretty good eplained. Also Andi from off-grid garage made serveral videos about that, f.e. "LiFePO4 charging and discharging curve explained. How far to go?"

[@NoelBarlau]

@@BluBioBike Thanks for the references. I'll check it out.

[@OffGridGarageAustralia]

Yeah, don't have them in parallel for that long and trickle charge them. If your BMS has an active balancer, just build the 16s battery and charge them to 3.45V/cell. The slowly increase the voltage and let the balancer do the job. If all cells are at 3.6V, call it. That's all you need for an initial top balance.

[@ShauneAlcott]

Hi Andy I received 48 EVE MB30 cells form EEL unfortunately 3 x were damaged, would a internal resistance test and a capacity test determine the damage was only cosmetic ???

[@LaVale_bios]

Hi! have you ever had the problem of double hole shiny-coated connectors ruining the conductivity of the cells? Usually it is the vendor who welds the connections on the cells. This vendor measures the internal resistance only at the soldering points of the aluminum piece, but not on the surface that should contact the busbars. We have had good cells but the connectors appear to be powder coated, like shiny paint that ruins the conductivity. It' s impossible in that point also measure the internal resistance of the cells. On the surface connecting with the busbars the conductivity is so lousy that the tester does not detect good resistance anywhere. This is dangerous and could lead to overheating problems and ruining the cells. The seller won't listen to reason and says that internal resistance has to be measured only at the soldering points -and not on the surface that should make connection with the busbar.(!) -. 🙁What do you think? Thank U in advance.

[@beebop9808]

Hmm that's interesting. Not sure how that would effect performance in the end? I guess right up front I would question how that's going to effect the end of the charging cycle? May be giving the balancer a real workout or cut the pack charge short if you don't have a balancer? Reckon you'll have the answer to that soon enough...

[@MrDingaling007]

Would it be worth logging Temperature as well? Ambient and cell temperature. Not sure just throwing ideas out there,

[@OffGridGarageAustralia]

Well, all cells were in the same room and at the same temperature. There was no major change during that week of testing. There would need to be a temperature change of 20° or more for one cell to have an effect on the charge behaviour.

[@1982757]

Great video. Could this behaviour explain why the SOC of my two paralleled 5KwH batteries drift apart during heavy amp draw? One could show 50% and the other 65%. Both are connected to busbars with the same length and thickness wires. Thank You.

[@OffGridGarageAustralia]

50% and 65% is not too bad of a drift, considering you are in the absolutely flat part of the discharge curve. Do they come together again once you fully charge them?

[@1982757]

@@OffGridGarageAustralia Yes, sir. They synchronise at the end of charging.

[@marioprzepiora]

Question: why use 280 cells when 304 and 314s are available? Is it backward compatibility with other banks, price per Ah, density, or some other factors? Just trying to understand the reasoning behind it.

[@moestrei]

I doubt that they do any testing before shipping. But the production has achieved such a high level of quality and consistency that they can make that claim.

[@leumel900]

Where can we buy this incredible red T-Shirt with the "Weather 3.5Amps" print on it?

[@OffGridGarageAustralia]

@Averagejoe made it for me!

[@clarencewiles963]

91k subscribers and moving to ??? Now you’re getting there!

[@Knightliner69]

I guess Yixiang won’t come up with a explanation about that charge curves. They are selling those cells only and the cells are within the specs (capacity and internal resistance wise). Only EVE as the original manufacturer and with their experience can give you an explanation maybe me thinks.

[@MMMM2MMMM2MMMM]

This is exactly why i buy directly from EVE myself.

[@adon8672]

​@@MMMM2MMMM2MMMMis it cheaper to buy directly from EVE? What is their minimum order quantity?

[@mfgxl]

@@adon8672 No, usually more expensive. For what reason? Maybe the keep the best of the best for their direct sales?

[@MMMM2MMMM2MMMM]

​@@adon8672 You may choose to buy from China or from a "local" wharehouse. When you buy from China, it will be cheaper, but.... shipment will take a long time. Buying from EVE directly is a fraction more expensive. But never had a weird performance with any cell. I bought 64 MB31 cells lately. EVE does test the cells before cells are being sold to resellers. You will get the best cells from EVE itself. I will ask my contact if i may share her contact information.

[@OffGridGarageAustralia]

The question is, does the difference in charge curve really matter? I will build this battery and see how it behaves.

[@user-ey1mo5oe5x]

How do the internal resistances compare to the graphs?

[@cubeamspro1003]

We are here for their boxes, not their batteries... :-) These are not good batteries for sure. Andy thank you for taking the time to graph them.

[@OffGridGarageAustralia]

I wouldn't mind another reseller with good quality batteries.

[@DavidBTB]

As for using too much energy, couldn't you use a regenerative discharge by draining the batteries through a boost converter to dump the energy into another battery? Use a shunt meter or coulomb counter to measure the energy?

[@OffGridGarageAustralia]

Yes, I could, but how do I control all this and also have a constant current draw through the entire process and voltage range from 2.5V to 3.65V? It's easy said but very hard to build. The ZKE tester does it perfectly but also wastes all the energy.

[@utubeuser1024]

Interesting about the rest voltage - if you absorb the battery down to 0.000A (will take ages), then I suspect the voltage will "float down" when disconnected much much slower (if at all)....

[@OffGridGarageAustralia]

That depends on the absorption voltage though. 0A means usually overcharging at 3.65V.

[@blackeyes18ro]

how the bms will behave when the curve is in the highest voltage difference?

[@andreasw5925]

According to the config. Balance when the balance start voltage is reached or cut power when OVP is triggered

[@OffGridGarageAustralia]

We will test this!

[@mikybinns9343]

Maybe you will get a voltage spread from the cells at about 3.4 so they wont balance so well higher than that...i have a similar issue with some cells. The solution is to lower the top voltage and reduce overall capacity a bit. Im experimenting with 17s which allows a lower max voltage so all the cells have a similar value at that point Interesting video

[@mfgxl]

I did the same with lowering battery voltage and my cell issues went away.

[@thomasschmucker14]

I have seen this in calt 280 cells balance turn on at 3.4v 3 cells hit balance sooner than the rest of the cells active balance takes out charge out of these 3 cells then when the rest of the cells actually reach 100 percent 3,45 these same 3 cells fall behind it's the same issue you have with the charge curve. Solution moved balance voltage up to 3.45 so the cells wouldn't start balancing until full have not had a issue since

[@tonydickerson999]

Looking at the results of your second test the charge curve has changed and I suspect it will continue changing and will eventually be in line with the other cells

[@OffGridGarageAustralia]

Interesting interpretation. We will find out!

[@tullgutten]

I wonder if paralell connecting them all and float charge to 10 A total then retest if they match better then

[@OffGridGarageAustralia]

We will let the JK BMS do that job.

[@VitpilenJpn]

Very interesting. I have two 24V batteries with 105Ah cells. One is homogenous with cells purchased from Alibaba, well balanced, the other is a plastic case battery purchased from Aliexpress that I took apart and replaced some bad cells with. I'm trying to figure out how to use this jumble of batteries. The JKBMS 1A balancer is not helping, one cell is hitting OVP/UVP and not getting the full capacity. If I put 8 cells in parallel and do a few charge/discharge cycles from 2.5V to 3.45V, will the variation decrease? In the future I'd like to put them in series and change it to a 48V system, and add a 280Ah 16S battery, but it's a tough road ahead.

[@clintoncoker6]

Put the (bad) cells in parallel and charge them all to as high as you dare (I would go to 3.55 maybe) and leave them there for a couple of days. This will ensure a good top balance. Cycling them together won't achieve anything. Be aware that mismatched cells will never retain their balance under load, so only worry if they are actually going outside of spec (i.e. under 2.5V). Only worry about their balance when the battery is unloaded.

[@clintoncoker6]

Although if one cell is consistently hitting UVP & OVP, it sounds like that cell is either under capacity compared to the others, or has a high internal resistance.

[@VitpilenJpn]

@@clintoncoker6 Thanks for the advice. I have failed to balance the top of the cells in parallel several times. Maybe because I set the voltage too low (3.5V) or because I didn't charge them long enough, but they looked balanced, but the charge levels were not the same. A few days ago, I tried to balance them in series. After charging them with a 5A charger, I connected a constant voltage power supply set to 28V and charged them slowly at about 2A so that OVP would not occur. At one point, I reconnected the constant voltage power supply to the cells that were not increasing in voltage from around 3.36V and supported them up to 3.4V. As a result, all the cells stabilized at about 3.48V, and the balancer did not work, so I left them like that for more than 24 hours. I connected them to an inverter, and since the sun was not fully charged or fully discharged after that, it looked like they were balanced.

[@madsskov_dk]

I’m a bit concerned too. The active balancers will work overtime on a battery with the mixed cells you have. Like a Frankenstein battery… What balance start voltage to set, hmm? I wish it was practically possible to test all 16 cells like 3 times, to see how they behave. With Home Assistant connected to a JK BMS, maybe via Victron, cell data could be collected and graphed together, when working as a battery. Would require lots of energy, but give lots of results, faster :)

[@OffGridGarageAustralia]

Maybe it will be OK once top balanced and the balancer has gotten them all inline at 3.6V. We will see shortly.

[@Francesco105590]

ciao, per 2 pacchi batterie 15Kwh di 16 celle ciascuno come si collegano col bms jk? occorre uno specifico?

[@davidrenn6897]

Andy, have you seen the Yoshino solid state batteries? How long do you think it will take for the technology to be available in our environment?

[@OffGridGarageAustralia]

It will not come to the home storage market soon. LiFePO4 is here to stay for a very long time. What would be the advantages of SS batteries in solar storage systems in your opinion?

[@davidrenn6897]

@@OffGridGarageAustralia there are claims for greater energy density and battery life but that is yet to be proven.

[@OffGridGarageAustralia]

@@davidrenn6897 solid state seems still a long time away. All these new chemistries are good for certain applications but lfp is still unbeaten for home solar storage.

[@boringsoftware2093]

can it be explained by differnt mixture of the (black) lithium paste? (more lithium less lithium used?)

[@OffGridGarageAustralia]

We will never know, but unlikely.

[@mausball]

That looks like there is something going on with the construction or the electrolyte, likely the latter.

[@clarencewiles963]

This was my thought as well the electrolytes were not balanced. But I would like to see a new check after six months of use. He could check all of the batteries 😂

[@Franky6767]

@@clarencewiles963 I'm not sure if Andy will use these cells at all, he certainly looked very unhappy

[@OffGridGarageAustralia]

Yes, I will use them and build a 48V battery. Then we will see if it is an issue.

[@Kwunt-l3q]

Hi andy, im in Oz and just found your Chanel and subscribed. I would like to build a simple 12v portable system to use when going camping using a 2000w inverter. Where and what would you recommend i buy?

[@bstaab2590]

Hello Andy, I'm just looking through your playlist of the battery build. I'm missing the part where you tension the cells, or did you not do that and if not why not? Greetings from Germany

[@OffGridGarageAustralia]

I use all my cells in the shelf uncompressed. There are a few videos on my channel where I explained why.

[@bstaab2590]

@@OffGridGarageAustralia Do you have a link for the corresponding video? I've already seen a few, maybe I've overlooked or missed it. Unfortunately, my English isn't that great either.

[@GeoffHou]

Would be interesting to see internal resistance when the voltage starts creeping up.

[@OffGridGarageAustralia]

We will find out.

[@Dutch_off_grid_homesteading]

Heya, I don't think there is a standard.does curves look more like leadasset cell's lol (olmost) but yes that is strange like there is some differant materials used insite. can't we get curves from the JK BMS's, that would be nice so we could get more graphs in the future for comparisons. got my 2st batch of 16 @320Ah cell received from gobel power hope to start testing soon.🤣🤣🤣😂🤣

[@TimothyDovel-w1s]

When you say, i will post it down below, how do I go down below?

[@aniketos7336]

so did eve go back to single stud cells? you mentioned it was latest design. wonder why they went away from the double stud cells

[@dlya_svoih]

As I understand, eve can ship with double terminals and without any terminals in bulk packaging. And resellers are free to weld any kind. Some sellers offers my studs as standard and double at extra couple of usd.

[@awbean5254]

1. ANDY, HAVE YOU EVER PARALLEL CONNECTED DIFFERENT MANUFACTURERS BATTERIES 🔋 🔋 TO ONE INVERTER? 2. I WAS INFORMED THAT DIFFERENT MANUFACTURERS BATTERIES HAVE TO BE CONNECTED TO SEPARATE INVERTERS. 3. WHAT HAVE YOU EXPERIENCED? 4. WOULD ‘CANBUS’ INTEGRATION BE THE REASON FOR SEPARATE INVERTER CONNECTION?

[@bsod5608]

Why are u screaming?

[@OffGridGarageAustralia]

Yes, we have parallel and series connected cells and batteries of different manufacturers all to one inverter. No problem. All the videos and findings are on this channel.

[@tomberger6484]

Indeed this is very interesting. But in practice this shouldn't play a big role because the BMS intervenes here. IMO this is a strong plea for active balancing.

[@boringsoftware2093]

of course the big danger in differnt charge curves is: overcharging of some cells (with explosion?) can JKBMS remedy this problem? (dude used MPPSolar 5kW on 15S without a bms for half a year... and it worked... but one cell was only 270Ah)

[@OffGridGarageAustralia]

LiFePO4 don't explode. That exactly what the BMS is for to catch such cells and disconnect the battery if something goes out of specs.

[@boringsoftware2093]

@@OffGridGarageAustralia let's put it like this: if one cell is bad, the whole pack can not be charged to it's full capacity? (because one cell going up in voltage faster than all others? any idea how to remedy this?)

[@PeppePer-e2r]

Hi Andy. I'm pretty sure it change after a few cycles of charging/discharging... Peppe from Napoli

[@OffGridGarageAustralia]

Yeah, that could be true. We will find out.

[@TheDefpom]

When comparing v and milliohm going to 3 decimal places might give slightly better data to compare relative health

[@OffGridGarageAustralia]

I don't think my devices are precise enough in that space.

[@TheDefpom]

@@OffGridGarageAustralia it’s not so much about precision, but relative measurements, which from what I could see in the video were perfectly usable measurements for relative comparison between each cell

[@TheDowe]

Now you should test all 16 to se how many have this voltage deviasjon.

[@OffGridGarageAustralia]

I know, i just don't have enough energy atm due to being in winter down here. And the time... So I will probably build the battery and let Home Assistant catch the cell voltages during cycling.

[@wallacegrommet9343]

You are funny as hell! Which I greatly appreciate in such a serious technical topic

[@OffGridGarageAustralia]

Thanks you.

[@pteepan2708]

I would think encouraging the flexible bus bars would be the way to go.

[@Paulisherebrewing]

Hi everyone, I am looking for a bit of help, I want to upgrade my inverter a soltaro ESS hyper3680 to a hyper5000. but I am finding it difficult to get one here in Ireland.(supplier no longer selling Soltaro or just the inverter) it looks like the soltaro batteries can only communicate with the soltaro inverter brand, can any one confirm this? or is there a way around it. I can get solis inverters through work, would this inverter communicate with my 2 soltaro lfp batteries. thanks Paul.

[@933NIls]

Hi Andy Do you think that daily overcharging of the cells is very harmful or do you have any information on this? I am referring to the issue of allowing the charging current to saturate to 0A with a voltage per cell of 3.45V, as you have been doing for a long time according to VRM. According to nordkyndesign, this is supposed to be dangerous. However, every BMS I have seen so far does this. Nice Video 🙂

[@OffGridGarageAustralia]

Charging to 3.45V and let the current taper off to 0A is fine and will not overcharge the cells. It is very different at higher voltages. I know the Nordkyndesign website and graph and I believe this is just one person's opinion without any real scientific background or testing consistency. It also says this on the website itself if you read though the text.

[@blablubbb0815]

Danke für das Video. Ich bezweifle dass der Innenwiderstand von der Farbe Deiner Unterhose abhängt, aber er wird sicherlich vom Ladungszustand abhängen. Meine Vermutung ist, dass die abnormalen Zellen bei hoher Ladung einen höheren Innenwiderstand haben. Das könnte diese Kurve erklären. Ich würde aber auf keinen Fall versuchen bei 100% zu messen, eher 80%, also wo die Abnormalität ist und bevor die Spannung hochgeht.

[@Scat.original]

Could you please add a direct link to the Yixiang Batteries?, can find them on your site Thanks

[@OffGridGarageAustralia]

Sure. It's there now in the description or here off-grid-garage.com/batteries/

[@bigthamal]

Well, it means these batteries charged faster than other batteries. Andy what is the charging time difference between other vs these batteries?

[@OffGridGarageAustralia]

Cell6 charged for 7h10min (297.8Ah charging capacity) Cell12 charged for 7h15min (293.9Ah charging capacity) Time can really be different depending on many factors.

[@ЁлыйПалыч]

I think you have grade B batteries. To verify this, you need to peel off the black plastic insulation that is on top of the cell and carefully examine the quarcode. It is possible that the factory markings have been sanded off and a new code has been applied. The fact is that at the factory, cells that do not qualify for testing for grade A are marked with the letter B on top of the quarcode. Small wholesale trading companies buy these cells from the factory at a reduced price, then grind off the quarcode with the B mark, apply exactly the same quarcode without the letter B, weld threaded rods to the terminals and sell these batteries as grade A.

[@dlya_svoih]

Yes, some rubbing with alcohol to remove glue can reveal a "halo" of sanding or polishing around qr code area. Sometimes they even don't bother to match qr code to writtenserial number ;)

[@OffGridGarageAustralia]

If there is another barcode under the black cover. it would be fake batteries. B-grade cells usually have a B stamped onto the QR code as we have seen and tested before.

[@ЁлыйПалыч]

The fact is that on such cells the code with the letter B is polished off and in its place the same QR code is applied, but without the letter B. Last year I bought 48 of these cells on Alibaba. However, they turned out to be quite good new cells.

[@dlya_svoih]

@@OffGridGarageAustralia there is a technique how to "relase" B qr code. Without peeling off the plastic cover it is nearly impossible to determine this fact.

[@OffGridGarageAustralia]

@@dlya_svoih I know, we have shown this this here on the channel before.

[@DanBurgaud]

3:40 no corner plastic protector?

[@OffGridGarageAustralia]

The foam is no thick, I don't think it is necessary. Saves a bit on plastic.

[@boringsoftware2093]

would be great if a JKBMS does all that for the user :D

[@whatmattersmost6725]

Just found your channel, awesome work! I found it because of the Overkill IOS app! Best app for JBD BMS, I already changed my BMS to have a higher cutoff low temp charging (trigger condition) from -15c (factory) to -1.0c, hoping less cell damage when charging in below 0c in the winter? Thank you!

[@OffGridGarageAustralia]

Welcome to the channel! You cannot charge Lithium batteries below 0°C. It will damage them instantly. Set this to 2-3° and get a battery heater if possible.

[@whatmattersmost6725]

@@OffGridGarageAustralia Thanks for the help! I made a mistake and I changed the "Discharge under temp" I think it was set to a lower value, and I changed it to -1F release condition 14F and Trig Delay same 5. So I should change Charge under temp Trigger Condition is 14F Release Condition 23F. This is what the BMS was set by the battery seller. Thanks

[@adon8672]

This is not directly related to this video but I need urgent help. My JK BMS shuts off power briefly, then reconnects if a cell exceeds the overvoltage protection limit. I noticed on the app that it shorts off charge only and leaves discharge on but it still cuts off power briefly then reconnects power. This keeps happening as often as a cell exceeds the overvoltage protection limit. Please is this normal behaviour? How can I prevent this problem from happening over and over again?

[@madsskov_dk]

Sounds like normal behavior. When a cell reach over voltage protection level, for example 3.65V, then the BMS disconnects charging, until the recovery voltage is reached.

[@terrencewpalmer]

If battery charge controller max voltage is to high and 1 cell could be running away . You could lower charger voltage a bit . Then look at balance. Terry

[@terrencewpalmer]

How many cells What's charger voltage. What's total voltage when single over voltage trips

[@adon8672]

@@madsskov_dk thank you for the prompt response. I understand the part of the BMS disconnecting charging being normal. What I don't understand as normal is the part that the BMS also disconnects discharge for a few seconds and then reconnects discharge again. Is this normal as well?

[@adon8672]

@@terrencewpalmer thank you but is it even normal for the BMS to interrupt the discharge process, then reconnects? I thought when a cell hits the overvoltage protection limit, only charging should be cut off.

[@leondavibe]

note to self , only buy one colour undies higher balance start voltage, have a battery with a degraded cell that i noticed it would top the voltage sooner ie hit 3.45 balance start voltage then move power from it to other cells only to move the power back when reaching full so i just lifted the balance start voltage to 3.48 to give others time to catch up before it starts to balance can only imagine how certain cells will get out of balance with a passive balance bms ie discharge for long time unable to correct during the short CV phase causing it to degrade rinse repeat then

[@OffGridGarageAustralia]

You mean, I cannot mix black, blue and pink (sometimes)???😯

[@DanBurgaud]

29:30 could it be because these are a generation apart from gobel or certified cells? ie, slightly different chemistry? I do admit, I prefer the charge curves of gobel and certified more than yixiang's cells. I guess these yixiang when assembled as 16s, will cause the balancer a hard time.

[@OffGridGarageAustralia]

No, you have seen the specs and parameters of the cells. they are new and fresh.

[@DivyangM]

Andy, when you said, it's just a marketing thing that cells are matched, I started wondering, why not ship batteries connected parallel? Even if it's thin flexible wires, it would keep them matched right? Or it will damage the batteries in someway! (Not an expert here)

[@upnorthandpersonal]

You're mistaking 'matched' with 'balanced'. Truly matched cells are those that have their internal resistance can capacity matched over the entire curve. Balanced just means at the same state of charge (which in LFP can not be determined by voltage).

[@putteslaintxtbks5166]

I know alot of people, when they get their cell put them all in parallel to balance befor building their batteries. Shipping with them in parallel could be a bit riskier because if any were got some type of short, all in that box would also be shorted, but as you said, it could be with thin wire that would just fall apart if more than a couple amps was going though them. If you had a bad cell that self drains, it would also bring those other down to, possibly causing them to be damaged if voltage dropped to low.

[@upnorthandpersonal]

@@putteslaintxtbks5166 LFP can not be balanced by putting them in parallel. The voltage curve is too flat for that.

[@DivyangM]

Thanks to all for replies. Good points that I didn't know.

[@PhilippeCJR]

Are the gobel cells v2 and the Yiuxiiuiang cells v3?

[@OffGridGarageAustralia]

Yes, I think so. If the terminals are round, it's V1 or V2.

[@PhilippeCJR]

@@OffGridGarageAustralia v3 is stacked technology and v1 v2 is winded. This could maybe have influence.

[@binauchich]

These cells will slow down the charging voltage at the end of charging. Because only at the end when the voltage rise they will get charged too.

[@mathschut505]

Mit wieviel AWG oder mm2 schließt du deine Akkupacks an? Grüße aus der Nähe von Kassel :)

[@OffGridGarageAustralia]

Depends on your circuit breaker. I use 35mm cable for a 100A DC breaker.

[@louwrentius]

Andy, if a cell fails, what would the behavior look like?❤

[@OffGridGarageAustralia]

Depends on how it fails. Voltage sag, internal resistance, heat, overvoltage... many possibilities.

[@louwrentius]

@@OffGridGarageAustraliafriend had this: one cell wasn’t torqued 😮😮😅 nut was discolored

[@OffGridGarageAustralia]

@@louwrentius always double check the connections and mark them with a pen. If you have, also check with a IR camera for hotspots.

[@louwrentius]

@@OffGridGarageAustralia yes he has done the marking now. We saw the affected cell spike to 4.5 volt and under load sag to less than 2 volt. Multiplus / cerbo was screaming the whole time about both overvoltage and undervoltage. We were confused at first, thought the cell was bad. Fortunately it wasn’t but this is how you burn down your house 😬😬🌷🤷‍♀️

[@GregOnSummit]

It would be nice to be able to opt out of the buss bars and yellow plastic and maybe save some $$. All of the ELL boxes I've bought have flexible buss bars, yellow sheets, and foam pads ... which I add to the the yellow sheets, not the battery.

[@OffGridGarageAustralia]

Yeah, on the other hand, too many options may increase the price again because boxes need to be opened and accessories added or removed.

[@volodymyrmatsyuk8528]

The QR code scanner shows Cell Type - LiFePO4 (B). Who knows what the letter (B) means? Maybe they are Grade B cells? Andy ignored it.

[@educateorstagnate9183]

Lithiu. Batteries do not have an internal discharge chemistery. If the ells are drifting off , they are faulty or you got used batteries. Storage will not discharge cells to 0.001 volts at the time of storage. If they are discharging then thise are faulty bateries.

[@nguyenhoaihan9643]

XIN CHÀO, BẠN CÓ THỂ CHO TÔI link mua sản phẩm pin EVE 280ah không. Cảm ơn ❤❤❤

[@oilpressing]

ANDY why are there no labels on the side of your batteries???? I have the same type of battery/barcode and they stay perfectly ballanced without a ballancer

[@OffGridGarageAustralia]

What labels? Some resellers put labels/white stickers on the side.

[@oilpressing]

@@OffGridGarageAustralia Yes I never seen batterys come without stickers/labels before, bit fishy if you were to ask me

[@OffGridGarageAustralia]

@@oilpressing this is not something the manufacturer does. Some resellers put them on, some don't. It does not indicate anything. I have probably half half..

[@oilpressing]

@@OffGridGarageAustralia thanks for the info

[@alanaland1990]

Difference possibly caused by temperature and chemistry. Maybe test cells at the same temp, say 25 Deg C.

[@OffGridGarageAustralia]

all cells are in the same spot in the garage at the same temperature.

[@robbot50]

As you say it is a black box (blue box 😀) with chemicals. So the differences are guilty of that. The supplier will undoubtedly experiment with this. We will never hear what the differences are in the compound chemicals, Date, producer, production machine number, etc. are the same however the serial number is not in line with the other! The question is how reliable is the data provided? 😉

[@OffGridGarageAustralia]

Yes, we can only scan the QR code and see what information it delivers. I would assume these are not true A-grade cells and because of that variation they got sorted out as B-grade or storage grade.

[@MMMM2MMMM2MMMM]

Dear Andreas, why dont you test all cells? You have two out of 16 strange behaving cells. Maybe you have more. I would have tested all 16 myself.

[@OffGridGarageAustralia]

Not enough energy atm. We are in the middle of winter. And this is super time consuming as well. It takes weeks.

[@Dionysios_Skoularikis]

Πιστεύω πως αυτή η διαφορά οφείλεται σε διαφορετικές ρυθμίσεις των μηχανών που κατασκευάζουν την άνοδο ή την κάθοδο των κελιών.

[@andrewradford3953]

10 day delivery, much better than TWO WEEKS for a chlorinator!

[@upnorthandpersonal]

I understood that reference!

[@OffGridGarageAustralia]

Hahaha, yeah, you wonder how this works sometimes!

[@ValRigoli]

Knowing the capacity side is all good, I would be more interested in charge curves only after a few good cycles. I think it's too soon to be overly concerned.

[@OffGridGarageAustralia]

Yeah, I will build this battery and see what the cells do once top balanced.