Rebuilt battery and now its worse?

[pishta]

Silverfish style 48v13Ah drop in battery had a range of about 4 miles. Just low capacity? Took every cell out and tested it on a Liitokala M4 charger that has a capacity test, I'm assuming a timed discharge into a 5ohm 1w load down to ? I reused all the cells that tested higher than 2200mah and had a internal.resistance under 30mOhm just as 70% of the cells showed. The other 30% were shot: 0V, under 2000mah or >30mohm internal resistance. I replaced these with 2500mah <25mOhm cells and welded them all back into the frame with new nickle strips made for this staggered frame. Reused original BMS. Battery shows full on meter and on display. Now the bike goes 2 blocks and shuts down. It will start up again but the controller shows full battery but the battery meter shows 1 bar. I literally tested every cell in this 65 cell battery over the course of getting the strips mailed to me and the welds are good enough to leave nickle on the battery if you try and peel off. What gives? I'm wondering what the cutoff voltage is on the capacity charge cycle. I can use another tester that is more manual and will set the cutoff voltage to 3.4 v and see if that capacity matches the liitokala charger. Only thing I can theorized that the liitokala tests down to some ridiculous 3v or something before stopping the "clock" that shows the capacity? Bummed that this didn't work as it took me 2 weeks to get the staggered nickle strips.

 

[E-HP]

Bummed that this didn't work as it took me 2 weeks to get the staggered nickle strips.

Not the worst outcome. Nothing blew up or burned down, so not all bad .

 

[lnanek]

Sounds like the BMS is tripping. That's usually A) due to hitting max discharge limit, which can be set lower in the controller or B) voltage sag seen in a weak battery when putting out more amps than it can handle and hitting the BMS low voltage cut off or C) an individual cell group reaching minimum voltage cutoff.

C can happen due to loose BMS cabling. It should have a connection to every p-group. It can also happen due to the battery being out of balance. If the BMS supports balancing, typically you can leave it on the charger for a really long time and it will drain the p-groups that are at max while it fills up the lower ones. Some cheap UPP BMS don't have the balance resistors populated to accomplish this, however. You could check the individual voltage of each p-group after it powers off, I suppose.

Voltage sag is tougher to catch. Might be able to measure voltage while dumping the battery into a large dummy load like a resistive heater. If a certain p-group can't handle the amps, might be possible to replace just that one. Cells can be damaged by heat, so maybe some were damaged during the repair.

 

[slaphappygamer]

It can also happen due to the battery being out of balance.

With the original cells testing at 2200mah and the new ones are 2500mah, would the pack ever balance properly? I mean, assuming the BMS supports it.

 

[docw009]

Now the bike goes 2 blocks and shuts down. It will start up again but the controller shows full battery but the battery meter shows 1 bar. I literally tested every cell in this 65 cell battery over the course of getting the strips mailed to me and the welds are good enough to leave nickle on the battery if you try and peel off.

What gives? Either the BMS is doing its job which means a low voltage cell group, or the BMS itself has gone bad. Check the 13 groups for voltage and balance. One of them might be down to 3V. You need to get this data to find out what's really happening,

We could speculate and guess what is wrong, but need to see your voltage results to assess the health of the battery.

 

[amberwolf]

With the original cells testing at 2200mah and the new ones are 2500mah, would the pack ever balance properly? I mean, assuming the BMS supports it.

If the cells are different internal resistances, then it will never *remain* balanced, even if they were all the same capacity (which they aren't either).

But as long as it is left on the charger long enough for hte charger to stop cycling on and off (which might take hours, days, or more), it will still "balance properly" each time it is charged.

 

[pishta]

Ill have to charge the battery again, run it around the block twice until it cuts out and remove the cell frame and test each P group for a delta in voltage. I hope I find a dead P group again as I found an entire group of 5 way out of spec when I took the old pack apart. Perhaps a faulty BMS, I got plenty of those. Will update. Im just stumped its WORSE now...?

 

[lnanek]

Some BMS will cut out based on reaching a certain voltage delta between p-groups as well. E.g. your battery could work at 4.2V for all, 4.2V for some and 4.1V for others, and then stop immediately if some are 4.2V while some are 4V due to having a .2V difference. If the p-groups really have different total capacity now, it might be quite quick to reach that point.

Maybe you should invest in a Bluetooth BMS. They can often show you the voltage for each p-group in an app on your phone and have programmable settings. Some even have a switch you can use to force active balancing immediately.

 

[pishta]

Some BMS will cut out based on reaching a certain voltage delta between p-groups as well. E.g. your battery could work at 4.2V for all, 4.2V for some and 4.1V for others, and then stop immediately if some are 4.2V while some are 4V due to having a .2V difference. If the p-groups really have different total capacity now, it might be quite quick to reach that point.

Maybe you should invest in a Bluetooth BMS. They can often show you the voltage for each p-group in an app on your phone and have programmable settings. Some even have a switch you can use to force active balancing immediately.

good point. I do have a 350W Xiaomi scooter pack with a BT BMS but I think it needs its little dashboard to activate BT. Ill check it out.

 

[docw009]

Some BMS will cut out based on reaching a certain voltage delta between p-groups as well. E.g. your battery could work at 4.2V for all, 4.2V for some and 4.1V for others, and then stop immediately if some are 4.2V while some are 4V due to having a .2V difference. If the p-groups really have different total capacity now, it might be quite quick to reach that point.

I have read that comment often here, but I've never seen that. I guess the cheap ones I use are not that discriminating. The BMS in my packs only shut off when a group hits LVC. On unbalanced packs, that differential has been as high as .6V.

I have a pair of JBD bluetooths I will eventually try out. Perhaps the max delta can be set on them? I wouldn't do it though,

 

[lnanek]

Hmm. I've definitely read about that protection state in spec sheets before. And have had plenty of BMS act dead until I manually balanced all the p-groups using a volt meter and 1s cell phone charger on the low groups.

Isn't it this VDIFF error code in the JBD app screenshots?


I admit, I never had a BMS let me change or disable that setting, though. Just had to manually balance a lot of batteries and off they go.

 

[docw009]

Hmm. I've definitely read about that protection state in spec sheets before. And have had plenty of BMS act dead until I manually balanced all the p-groups using a volt meter and 1s cell phone charger on the low groups.

Isn't it this VDIFF error code in the JBD app screenshots?


I admit, I never had a BMS let me change or disable that setting, though. Just had to manually balance a lot of batteries and off they go.

I've only put the dumb BMS into play,

 

[amberwolf]

The reason for the voltage delta shutdown to be used is*** that any pack that has a significant difference between cell voltages has a significant problem developing, because the different group(s) either have failing cells, disconnected cells, etc., and so the pack could be becoming unsafe to use. Since it's better to err on the side of caution than fire the failsafe is to shutdown when this kind of difference occurs.

a 0.1v difference is generally quite a bit of variance on a capacity curve / SoC chart, so that is a common voltage-delta-shutdown factory setting.


***assuming a well-built pack of well-matched cells to start with; in cheap packs and generic packs of any price-point, that's not a good starting assumption but in better-quality packs, OEM-built packs for big-brand-name bikes, etc., it should be.

Cheap packs often enough don't even have a balance function in their BMS, so it's not surprising when they don't have any "extra" protections beyond a simple cell-level LVC (and hopefully HVC).

 

[amberwolf]

Isn't it this VDIFF error code in the JBD app screenshots?

Yes. Different apps / companies use different codes or symbols for it.

Most of the non-BT BMS don't have a way to tell the user what's wrong, because they're not intended to help the user do anything other than not start a fire from a potentially damaged or failing battery.

I admit, I never had a BMS let me change or disable that setting, though. Just had to manually balance a lot of batteries and off they go.

Some of the BT / smart BMS have that option available, some don't. There are threads where this is mentioned and some where poeple change it or even disable it so they can continue to use their poorly-matched and/or failing cells.

 

[Zambam]

A few days ago my moped/ scooter had an unplanned shutdown 3.5 miles from home (as opposed to planned when I was range testing my battery much closer to home).

App on JBD BMS says "cell under volt" on my 22S 2P 30Ah LFP battery , cells delta 0.467V, group 9 @ 2.627V. Took the bus home, strapped my back-up (1P) battery on a cart & took bus back to moped. Swapped battery (2 min), rode moped home w. cart and dead battery.

Didn't think of it at the time but I probably could have avoided taking the bus by lowering the LVC which was set @ 2.5 V to 2 V.

I've since R&R'ed the 2 cells in group 9, which had AC IR's of 6 & 14 milli ohms, much higher than the 2 milli ohms when the pack was built 900 miles ago. Dead pack is back in operation.

 

[j bjork]

With the original cells testing at 2200mah and the new ones are 2500mah, would the pack ever balance properly? I mean, assuming the BMS supports it.

It is not so much a question of the difference in the new and old cells as in how they are grouped together.
If every group has 5 old and 3 new cells for example, it could probably work fine.
If they are randomly mixed, not so much I would guess

 

[docw009]

It is not so much a question of the difference in the new and old cells as in how they are grouped together.
If every group has 5 old and 3 new cells for example, it could probably work fine.
If they are randomly mixed, not so much I would guess

In my opinion, mixed cells work well in low current applications, like powerwalls, where by design, all the cells are worked lightly and probably rarely drop under 90% SOC. . In an ebike, where 50% SOC is common, mixing cells in a parallel group should be avoided.

If the IR's are like 50% apart, it can cause enough power to be wasted in parallel currents (commonly modelled as zero) that significant capacity is lost in the group. The five old, 3 new is likely to work fine, but it might work better with 8 old, 8 new, etc,

Anyway, just writing from personal experience with mixed cells in a p-group giving me problems. Mixing an entire group has worked for me.. Yes, it might discharge differently, but balance BMS compensates.

 

[docw009]

With the original cells testing at 2200mah and the new ones are 2500mah, would the pack ever balance properly? I mean, assuming the BMS supports it.

The IR is another parameter to consider. I think 2200 mah and 2500 mah should work together if the IR is similar. If they are far apart, it can cause the cells in parallel to discharge unevenly, and in some cases charge each other, Wastes power.

If the BMS can balance, it will balance.

.

 

[pishta]

2 mohm? Wow is that low or standard for a good high discharge battery? The best I've found in surplus unused packs was 9 mohm. My pack is 5P and the Amp draw curve on this 48V/1500W (rated) hub motor is 91A at full torque @ 2730W and 27A at 1203W cruise. So I'm gonna need at least 20A discharge x5 to cover the +90A draw and stay safe? Or increase the P and lower the discharge rate to suit the battery spec?

>>Liitokala Lii-M4 has a fixed discharge rate of 500mah so my capacity tests were conducted at a pretty low discharge rate, possibly skewing the results? My manual PCB based tester loads the battery to 990-1000mah through a beefy ceramic resistor/heater and tests accordingly, but may give me similar results? <<

 

[pishta]

The IR is another parameter to consider. I think 2200 mah and 2500 mah should work together if the IR is similar. If they are far apart, it can cause the cells in parallel to discharge unevenly, and in some cases charge each other, Wastes power.

If the BMS can balance, it will balance.

.

I have quite a few loose BMS from scooter packs, what do I look for to determine if the BMS can balance? Someone mentioned vacant PCB pads for cheap non balancing models?

 

[Zambam]

2 mohm? Wow is that low or standard for a good high discharge battery?

Link to spec of the LFP cells in my pack with AC impedance between 1.5 to 3 milli ohms. Gotion 32135 15Ah 3.2V LFP Lithium-ion cell IFR32135-15Ah

 

[pishta]

Link to spec of the LFP cells in my pack with AC impedance between 1.5 to 3 milli ohms. Gotion 32135 15Ah 3.2V LFP Lithium-ion cell IFR32135-15Ah

Ah, I see that <3mohm is AC internal resistance. They list DC IR as <9. I wonder how you measure DC IR? The TS457 meter "AC four-wire method" should be to add a specific Hz AC voltage to the object being measured, and then use a high-precision voltmeter + ammeter to measure the data, and then calculate the resistance. This is what I used to determine my range of batteries IR values.


Dang, these batteries are 1.84 mohm....cool! 15Ah per battery? Forget stacking >70 18650s .....15 of these in series would make a 3C (45A discharge rate) 48v 15Ah battery for less than $50? "...20 Pcs Gotion 33140 15ah 15000mah 3.2v-3.65v LiFePO4 Battery Cells NEW ...$50.." Pardon Our Interruption...

 

[pishta]

Those look to be a little longer than 2 D batteries? 15 in series would be about 264x66x140mm package for a 48v E-bike battery. Interesting......

 

[amberwolf]

Dang, these batteries are 1.84 mohm....cool! 15Ah per battery? Forget stacking >70 18650s .....15 of these in series would make a 3C (45A discharge rate) 48v 15Ah battery for less than $50? "...20 Pcs Gotion 33140 15ah 15000mah 3.2v-3.65v LiFePO4 Battery Cells NEW ...$50.." Pardon Our Interruption...

How *consistent* are their IRs? That is just as important a question as what the IR actually is.

for example, if you have a 1.84mohm *average* cell model, and when you get your batch to build a pack from they range from 1.4 to 2.3mohm, that's around a 25% variance, which is huge, even though it doesn't sound like a lot by the numbers (less than half a mohm). It would result in a 25% variance in the voltage sag from a specific current, and/or a 25% variance in the ability to deliver that current at a specific point in the SoC curve.

If it's just a 1% variance (which is still significant but easier to deal with), or say about 1.82 to 1.86mohm, which would result in a 1% variance in the results....

The differences in characteristics between series cells (or groups) in a pack also affects their balance, and thus the available capacity and current delivery capability of the entire pack.

 

[pishta]

Thanks for that info! Ill be sure to match all my cell groups capacity s well as IR for my next pack.