SOK bluetooth password

[Aly_Oops]

That sounds like they may have fixed the algorithm.

I am hoping this new BMS fixing the problem. Still too early to conclude. So far it balancing back and forth between 110 and 115mv. I keep asking SOK for password, they just ignore my request. I caught one jpeg of a 100amp and when I took a close, it was an edited copy.

 

[Aly_Oops]

They are not hiding the settings, but they have made a business decision that they do not want customers to change them. I can understand that because some people will get it all screwed up and then expect SOK to help them fix it.... or blame SOK for problems. However, the decision comes with the consequence that they will drive away some people in the DIY space.

No issue with making changes. They can allow one to view it but not to edit it. What are they hiding. I suspect they set the parameters for max life and not max capacity.

 

[Alkaline]

The latest BMSs in the 12v batteries use a 33 ohm resistors, so they dissipate right around a tenth of an amp (100ma). After about 6-10 cycles, it's more than enough to bring things into line. Most BMSs only have a 20-40ma balance current.

The 48v batteries use a 43 ohm, but function in the same fashion.

JBD have 150-200 ma

 

[Aly_Oops]

If interested in getting SOK, may be best to wait until the new BMS is widely avaiable. If for some reason it doesn't work out, I may replace the BMS with third party BMS. Look like I can fit another brand. Getting dimension first.

 

[Aly_Oops]

That's not true at all. Here are all the default parameters - they are extremely typical for LiFePO4 batteries:
View attachment 133936View attachment 133937

Prove it, give us the password. If not then your statement isn't credible.

 

[Aly_Oops]

That sounds like they may have fixed the algorithm.

I believe they make a hardware change, resistor change, voltage increase, otherwise it seem the same logic they are using.

 

[robstrom]

20010 isn't the password. What is?

 

[FilterGuy]

I believe they make a hardware change, resistor change, voltage increase, otherwise it seem the same logic they are using.

They have been tweeking the FW as well..... and this can change the behaviour significantly.

 

[robstrom]

I'm not questioning the behavior.
I just want to see the numbers

 

[FilterGuy]

I'm not questioning the behavior.
I just want to see the numbers

Sorry, I was replying to @Aly_Oops post #26

 

[robstrom]

Copy that

 

[Aly_Oops]

20010 isn't the password. What is?

that the password for Basic Setting, doesn't work for Parameter Settings. I wonder if I can get Shortcut script to cycle through all the numbers, to hack the password. It a 6 digit password.

 

[Swagon]

With A grade cells? I disagree - based on the fact that the with this amount of current the batteries in the field always balance out after less than a dozen cycles. The balancers turn on above 3.4v, during charge, if cell deviation is 20mv or above. If you take a brand new battery and do an extended absorb for about 6 hours, the BMS will have balancing dialed in perfectly.

The latest BMSs in the 12v batteries use a 33 ohm resistors, so they dissipate right around a tenth of an amp (100ma). After about 6-10 cycles, it's more than enough to bring things into line. Most BMSs only have a 20-40ma balance current.

The 48v batteries use a 43 ohm, but function in the same fashion.

I bought my SOK battery from you, BTW. So far, I cycled my 12.8/206Ah battery 2 times, and it reads 210Ah. Great! HOWEVER, cell 1 is the weak point and hits 3.7V BMS protection while all the other cells are only 3.335V or so. (Cell 1 was also the first to hit cutoff when deep discharging the battery. Cell 1 has lowest capacity and seems to have slightly higher resistance. But it has more than the rated capacity.) Anyway, I'm seeing a top imbalance of 365mV. So I've been trying to top balance. I can see balancing kick in while changing when cell 1 reaches over 3.375V. It is obvious when it kicks in because the cell voltage drops by 40mV. As charging continues, the balancing toggles on and off because the cell voltage is going above 3.375V when balancing is off and below 3.375V when balancing is on. Eventually it stops toggling as the cell eventually charges enough to stay above 3.375V with balancing on. Also of immense importance, balancing only works if I am changing at 1A or more. If I turn my bench supply down to 0.9A, it doesn't do any balancing. When I go from 1.1A to 0.9A, I right away see the 40mV jump in Vcell1, indicating the balancing is off. Clearly, 1A minimum charging is needed to balance. So I've been cycling my battery from 99% to 100% at 1.1A charging current. I've done it about 10 times already. It has started to help, just barely. One of the other cells (cell 3) is finally starting to round the corner at 3.386V before charging stops when cell 1 hits 3.7V. The other two cells are still a long way off at about 3.355V. I estimate it will take another 30 or more 99%-100% cycles at 1.1A charging to complete the top balance. I've already done it 10 times and I'm not even close. I'm charging at only 1.1A to maximize the balancing time. If I charge a 15A, there is so little time for balancing, I don't think it would ever get balanced in my real charging application. I think it is a flaw in the settings for balancing. The settings are probably OK once the cells are top balanced or nearly top balanced. But with the imbalance I'm seeing, the settings are not good. (BTW, I am an electrical engineer an specialize in power electronics. I define semiconductor solutions that my company then makes. I have defined Li Ion charger ICs, for example. After this experience, I'm considering to define an active cell balancer.)

 

[Swagon]

With A grade cells? I disagree - based on the fact that the with this amount of current the batteries in the field always balance out after less than a dozen cycles. The balancers turn on above 3.4v, during charge, if cell deviation is 20mv or above. If you take a brand new battery and do an extended absorb for about 6 hours, the BMS will have balancing dialed in perfectly.

I tried an extended absorb at 13.8V, and it didn't balance at all. The reason is that balancing turns off when charge current falls below 1A. This is why I'm cycling at 99%-100% at 1.1A charge. Since the balance is only 100mA, the net charge to the cell is 1.1A charge - 100mA balance = 1A net charge to cell 1. It's going to take soooo many cycles. The other cells are getting the full 1.1A. Let's say there is a 2% difference is cell SOC. 2% of 206Ah = 4.12Ah. At a 100mA delta, it can resolve balance in 4.12Ah/100mA = 41.2 hours. Since all the balancing is in the final 30-40 minutes of the charging at 1.1A, it might even take 100 of these 99-100% cycles. If charging at 15A, it would take thousands of cycles to balance. It's not good. Discharging to 98% makes it take longer to even start balancing while charging, so that is not an answer. For discharge, I run a 900W load for less than a minute. Then start charging, and that was even too much discharge, so it takes a few minutes to cell 1 to get to 3.375V and start balancing. I've fine-tuned the 900W discharge to just 50 seconds now, at then the balancing commences more quickly. In the end, I may need to disassemble the battery and balance the individual cells inside. But I don't want to do that if it violates the warranty. I'd love to get into the BMS parameters to see if I can improve the balance settings if there are any. But it doubt there are, otherwise, you would be able to help us.

 

[Swagon]

With A grade cells? I disagree - based on the fact that the with this amount of current the batteries in the field always balance out after less than a dozen cycles. The balancers turn on above 3.4v, during charge, if cell deviation is 20mv or above. If you take a brand new battery and do an extended absorb for about 6 hours, the BMS will have balancing dialed in perfectly.

One more thing. I see a 40mV drop in Cell 1 when 100mA balancing turns on. This seems too high, right? 40mV/100mA = 400mohm. Is my cell 1 damaged? Should I seek warranty replacement? Cell 1 still achieved 210Ahr, per the BMS full charge cycle.

 

[HighTechLab]

One more thing. I see a 40mV drop in Cell 1 when 100mA balancing turns on. This seems too high, right? 40mV/100mA = 400mohm. Is my cell 1 damaged? Should I seek warranty replacement? Cell 1 still achieved 210Ahr, per the BMS full charge cycle.

You identified the resistance of the balancing wiring harness, not the cell. It’s pulling full capacity, meeting advertised specifications, so no you shouldn’t seek warranty replacement, at least not yet…you are covered for 10 years so if it ever does turn into a bigger issue, we will take care of you.

I will get SOK engineers feedback on your other comments.

 

[Nobodybusiness]

You identified the resistance of the balancing wiring harness, not the cell. It’s pulling full capacity, meeting advertised specifications, so no you shouldn’t seek warranty replacement, at least not yet…you are covered for 10 years so if it ever does turn into a bigger issue, we will take care of you.

I will get SOK engineers feedback on your other comments.

There is something to say for that 10 year warranty.

Gives me quite the piece of mind knowing I won’t have to worry about my batteries until 2033.

 

[Swagon]

One more thing. I see a 40mV drop in Cell 1 when 100mA balancing turns on. This seems too high, right? 40mV/100mA = 400mohm. Is my cell 1 damaged? Should I seek warranty replacement? Cell 1 still achieved 210Ahr, per the BMS full charg

You identified the resistance of the balancing wiring harness, not the cell. It’s pulling full capacity, meeting advertised specifications, so no you shouldn’t seek warranty replacement, at least not yet…you are covered for 10 years so if it ever does turn into a bigger issue, we will take care of you.

I will get SOK engineers feedback on your other comments.

Thanks, I also realized this last night after I sent the message. It's because the voltage measurement and the 33mohm balance share the same wires to the cell. Ideally, the voltage should be Kelvin sensed with its own wires that don't ever carry current. Why? Two obvious reasons: 1) My Cell 1 is getting overcharged to 3.74V. Because the battery is balancing, it is measuring only 3.700V when the protection MOS opens, but then the balance stops and the cell is actually at 3.74V. (See attached photo.) This is aging Cell 1 in my pack at an accelerated rate. I suppose once the cells are balanced (eventually), this problem may go away. 2) As the cells try to balance, the balance will not work as hard because it will oscillate on/off. When trying to balance, it turns on the 33ohm, then the next voltage measurement is 40mV lower, so it thinks it doesn't need balancing, so it turns the 33ohm off, then the next voltage measurement shows it back 40mV higher and still needs balancing, so it turns the 33ohm back on. IMHO, the 33ohm balance is already pretty weak for a 206Ah battery, but now it will be ON only half of the time and only at the very end of charging. Effectively, I think it is just a 100mA*50% = 50mA effective balancer, which is only 50mA/206A = 0.00024C. And it can only work in the last 1% of charging due to the 1A charging limit and the 3.375V minimum cell voltage. So, 0.00024C * 1% = 0.0000024C effective correction. Is it enough to keep up with aging mismatch? (rhetorical question) My suggestion to SOK is to use two separate wires for balancing and for voltage measurement (true Kelvin sensing). Or as a cheaper solution, keep the existing single wires, but add 40mV in software to the measured voltage whenever the balancing is turned on. (They should actually characterize the 40mV themselves and for each wire in the harness.) That will effectively eliminate the on/off toggling. Then, try to make the balancing run for a longer time. For example, use the same criteria (1A minimum charging, 3.375V minimum cell voltage, ???Vdelta minimum mismatch) to turn ON the balancing, but use different criteria to turn it OFF. For example, reduce the 1A minimum charging threshold as far as possible, even maybe 0A. Also reduce the Vdelta mismatch. Can they add a software timer to extend the balancing beyond the end of charging? That would help tremendously. Obviously, active balancing would be better, but it is more complicated and more expensive. I think with a few changes in the software, they can improve the passive balancing.

 

[HighTechLab]

What BMS has kelvin sensed balance leads? Even some of the big boys with active balancers (REC) don't even have this...and certainly not at the pricepoint of the SOK. Don't get me wrong, from an engineering standpoint, yes this is the absolutely most accurate way...but consider if it's practical.

SOK has released their V8 BMS (note, they only use even numbers so there have not been 8 versions) with a lot of improvements across the board...more than I can list but I know charge detection (so things can happen when a charger is applied regardless of if the charge is flowing through the FETs) has been added and that solves the balancing issues and improves the heating pad operation functionality.

I am not sure which version you are on, but it's likely a v6.

If you want to force it to balance faster, enter the basic settings, click on idle calibration, password 200010, then type in 2A. This will trick the battery to run the balancers at idle current. Once you are satisfied with the balance, disconnect everything from the terminals of the battery and re-do the idle calibration at 0a, then do a full drain to 0v and recharge to 14.6 and it will bring the capacity reading and thus SOC% in the app back in line.

Yes, this process will take a while to run through all the stpes, but on the flipside you can do nothing and put the battery into service and it will correct itself over time, on its own.

 

[Swagon]

Thanks for these inputs! I like your trick and will us it. Would have saved me well over a week. With the default settings, I spent 2 weeks slowly moving the cells towards top-balance. I didn't finish. I had to use my van for a trip, so I reinstalled the battery. But at least now, the charger terminates the charging instead of the BMS protection terminating charge. So I'm comfortable with that.

Regarding the Kelvin wires, it turns out that the normal wires for Cells 2 through 4 have wires with much lower resistance, and are fine. When they finally started to balance, the 100mA balance current only caused a small voltage change of maybe 2 to 4mV. But for Cell 1, the balance causes a change of 40mV consistently, which causes oscillation of the balancing on and off. This 40mV is odd. 40mV/100mA is 400mohm. That's pretty high for a short wire, isn't it? Is the wire damaged? Or is it connected in a weird way to the BMS?

And, yes, you are correct that Kelvin wires should not be needed for voltage sensing with a measly 100mA passive balance. I was just confused by my particular battery because I was tired and didn't think about how 40mV (400mohm) shouldn't be normal. Also, for my job, I started working on an Active Balance chip with competitive advantages, and I plan to use the wire impedance as a benefit. My company doesn't yet make any BMS or Balancer chips, but we are always looking for good opportunities. We already sell some other chips into Chinese battery packs and power stations.