Custom E-Bike battery issues

[ebikerxx]

Hello fellow e-bikers,

I have some confusion about mixing 18650 Li-Ion cells of different mah while building a custom e-Bike battery pack.

I built a battery pack with 18650 li-ion cells with the configuration 10s2p using 3.6v, 2200 mah cells and a compatible 10s 36v BMS rated for 15amps. The pack was all fine except for the low range. Charging and Discharging cycles worked fine with the BMS cutting off at 40v and 36v.

To better the range I had to add more cells to the battery pack. I had 10x 18650 li-ion cells rated 3.6v, 2400mah. I reconfigured the pack by adding another row of cells in parallel to make it a 10s3p battery pack. So now I had 20x 2200mah and 10x 2400mah.

I charged the battery pack and there was no issues. It charged to 40v and was cut off by the BMS. The scary issue I faced when I took it for a ride is that the battery pack kept discharging below 36v and I stopped when it reached 34.8v. This wasn't the case before adding the 2400mah cells. I'm not sure if mixing different mah cells was a good idea.

I assumed this could be because of cells with charge imbalance, so to check, I once again charged the battery pack and the charging cut off at 40v. Before I take it out for a ride again to check if the discharging cuts off at 36v, I'd like to know your opinions about the configuration.

I checked online and found conflicting advice about mixing cells of different mah and brands.

 

[Sierratim]

Ideally all cells are matched for internal resistance to a fairly tight tolerance. This helps assure that they all charge and discharge with similar times guaranteeing as far as possible that they each 'carry their weight' equally. The result is max life and power out of the pack as all the cells tend to age equally.

I'm guessing that you're seeing the result of mis-matched internal resistances between the different capacity cells with the lower resistance (probably the higher capacity) cells charging and discharging faster. As soon as any of these lower resistance cells reach their max voltage under charge or min voltage under discharge the BMS cuts off the pack.

 

[harryS]

Your battery should be charging to 4.2V/cell or 42V, unless you're limiting the charger, Your BMS should allow the battery to run to 30V too, unless it's adjustable. 36 volts is just a little under 50% charge. In addition, if you are limiting the charge to 40V, do you know if all 10 groups are the same? If the pack is badly out of balance, that really hurts your range.

You didn't say whether you achieved your goal of 50% more range. All we heard is that the battery discharged to 36V as it is going to do, and then you shut it down. From your tone, it seems like you have even less range? Please elaborate,

If you did put a new 2400 ma cell in parallel with the existing new 2200 ma cells in each series group, I do not expect anything amiss. While you would prefer matching capacities, 2200-2400 is only 10% variation.

I do this on a macro scale, with a number of 36V packs that range from 4AH to 10AH, I put them in parallel when I need more range, expecting the AH numbers to add up, i.e, two 4AH packs deliver 8AH, and a 4AH and the 10AH give 14AH. And from what I see, they do.

 

[tomjasz]

I have some confusion about mixing 18650 Li-Ion cells of different mah while building a custom e-Bike battery pack.

Don't. Just don't or accept the potential cataclysm.

 

[ebikerxx]

Thanks all for your replies.

@Sierratim I ignored internal resistance of the cells since both the types were e-Bike batteries and the difference was just 200mah. Not sure if there is a standard resistance for electric vehicle cells? If I were to mix electric vehicle cells with laptop and power tool cells maybe then the resistance would be a major factor.

@harryS In my limited experience with e-bike batteries, the higher cutoff voltage and the lower cut off voltage is generally set by the BMS, and in my case the BMS lets the battery pack operate between 36v - 40v when measured with a voltmeter.

I really don't mind the operating voltage range since the sweet spot for efficient and long lasting batteries is between 30% to 80%. The trade-off in my case is lower range.

I have a voltmeter on my e-bike to measure the battery voltage. The problem I'm facing after adding the 2400 mah cells is that the lower cutoff voltage is not 36v anymore, I have got a 50% increase in range upto 36v as expected, but now the issue is that the battery is not cutting off at 36v which is a major concern.

Ideally the BMS should be switching off the battery at 36v, and I'm scared to keep riding my e-bike below 36v as I don't know what is the safe lower cutoff voltage. I can always manually switch off after discharging to 36v, but seems like the BMS has somehow lost its lower cutoff voltage of 36v.

 

[ebikerxx]

Sorry for the confusion and incomplete information. I checked with the BMS vendor who assured me that the operating voltage should be approx 30v - 41v.

I guess my original 10s2p configuration had some bad cell groups because of which the lvd was 36v. My reconfiguration of the battery pack to 10s3p seems to have fixed those bad cell groups. I can now safely ride my bike down to 30v. The range should now be off the charts

Well, it's my first time building a battery pack, and initially I was reluctant to use a BMS - to get maximum range, but the balance charger was too expensive, so I settled down for the BMS option, and for some reason didn't check the voltage specs, and made some stupid assumptions.

Thanks all for your time! Learning: Check the voltage of individual cell groups and not just the overall voltage of the battery pack.

 

[Gionnirocket]

and made some stupid assumptions.

I believe that you are continuing to do so.
Experimenting is an excellent way to learn but I do believe that you do need to do a bit more reading.
Just be very careful in the process as you are literally playing with fire.
One of many reads you should reference
Good luck

 

[tomjasz]

From endless_sphere

 

[Gionnirocket]

From endless_sphere
View attachment 74833View attachment 74834

I had a friend that smoked so much weed... he had similar finger tips

 

[ki11a]

Ive heard of green thumb before but....

 

[tomjasz]

Ive heard of green thumb before but....

Kentucky Fried Fingers!

 

[ebikerxx]

I believe that you are continuing to do so.
Experimenting is an excellent way to learn but I do believe that you do need to do a bit more reading.
Just be very careful in the process as you are literally playing with fire.
One of many reads you should reference
Good luck

Well, that is my weakness, I don't like reading, but have watched many videos of DIY battery packs. The cool thing about my battery pack is that I never used any kind of soldering or spot welding.

I was very careful with the configuration and used safety precautions like goggles and gloves. Maybe a bucket of sand next time. I followed Micah's instructions and wrote down the configuration before building the pack. I did let a few sparks fly, but had everything under control.

Here are some pics of my battery pack:

 

[Gionnirocket]

One important factor that you are ignoring is the source/specs of your cells.
All your cells should be from the same manufacturer/model/lot/date/source.
Mixing cells is just asking for problems whether now or in the near future.
Those building packs from recycled multi sourced cells may end up with a working pack... but not for much longer than it took to post the video to YouTube. It's maybe an option if you have no other.

 

[ebikerxx]

One important factor that you are ignoring is the source/specs of your cells.
All your cells should be from the same manufacturer/model/lot/date/source.
Mixing cells is just asking for problems whether now or in the near future.
Those building packs from recycled multi sourced cells may end up with a working pack... but not for much longer than it took to post the video to YouTube. It's maybe an option if you have no other.

Duly noted. Hope my battery pack lasts longer than that. I will keep updating this thread with my observations.

 

[harryS]

You also did not specify whether these were new cells or stuff that was scavenged, You also did not tell us whether you checked the individual balance voltages on all 10 groups, Based on the results (36-40V operation) I was thinking your battery was badly unbalanced, because of used cells. Now that I see your pictures, I see you're using something like a Vruzend. Probably new cells?

Some people love those Vruzends. I only see posts here from people who built crappy batteries using them. Is it user error or endemic contact issues? In any case, when you re-built it, the battery started to work normally, Says you had bad contacts.

In the future, and you cannot do it now because you wrapped it already, check those balance voltages. If you cannot get above 40V, you've still got a balance problem. That will only get worse with time.

I don't see a 200maH difference as major. Some theories predict a case where the 2.2AH cells go lower in voltage than the 2.4AH cell, and the bigger one will try to charge the two smaller ones. However, unless the parallel connection was awful, any voltage difference is millivolts, so only a minor charge effect. And what does Micah's book say about this? He's the guy that advocated connecting a booster battery to the main battery via the charging port. This bothers a lot of people. I tried it, and it works, but I don't see it working well if you are combining batteries to push more current. Then you need a parallel connection.

Anyway. keep reading. If you're going to use home built packs, you have to know a lot more. First, there's the safety aspect. Then there's longevity.

 

[ebikerxx]

You also did not specify whether these were new cells or stuff that was scavenged, You also did not tell us whether you checked the individual balance voltages on all 10 groups, Based on the results (36-40V operation) I was thinking your battery was badly unbalanced, because of used cells. Now that I see your pictures, I see you're using something like a Vruzend. Probably new cells?

Yes the first 10s2p pack had new 18650, 2200mah, 3.6v cells. The reconfigured pack has 2.5 years old 18650, 2400mah, 3.6v from the battery pack which came with the e-Bike.

My original battery pack had a failed BMS, and the cells were glued and spot welded, initially I didn't have the tools to desolder them, hence set out to build a new battery pack. After a lot of research I decided to use a solderless kit similar to vruzend as a safer option. Here are a few pics of my old battery:

 

[Gionnirocket]

Yes the first 10s2p pack had new 18650, 2200mah, 3.6v cells. The reconfigured pack has 2.5 years old 18650, 2400mah, 3.6v from the battery pack which came with the e-Bike.

My original battery pack had a failed BMS, and the cells were glued and spot welded, initially I didn't have the tools to desolder them, hence set out to build a new battery pack. After a lot of research I decided to use a solderless kit similar to vruzend as a safer option. Here are a few pics of my old battery:

Are you familiar with the story of Frankenstein?

 

[tomjasz]

I think he’s duly warned. Using Micahs’ book with attention to details like safety glasses and no jewelry are top issues. I was being super careful with a disassembled pack and didn’t notice a small piece of wire on my silicon work surface. I set the bare pack down and shorted out on the wire bridging cells. Instant POW! FLASH! Trashed a half dozen PF cells.

 

[ebikerxx]

Some people love those Vruzends. I only see posts here from people who built crappy batteries using them. Is it user error or endemic contact issues? In any case, when you re-built it, the battery started to work normally, Says you had bad contacts.

I will be more descriptive in the future, I condensed my post to get opinions about mixing cells, however now I understand that there could be a wide range of other issues not related to the cells.

I did read a bit about contact issues in vruzend kits, but I verified all the contacts were fine by checking the voltage of the pack and cell groups, but maybe I missed checking the cell groups after laying down the strips and bus bars.

In the future, and you cannot do it now because you wrapped it already, check those balance voltages. If you cannot get above 40V, you've still got a balance problem. That will only get worse with time.

The battery pack higher cutoff voltage is 40.6 volts to be precise, which is close to the 41 volts mentioned by the vendor. I'm suspecting a contacts issue in the 10s2p pack is the cause for LVD of 36 volts.

When I reconfigured the battery pack, I double checked all the contacts and hammered the caps hard to make sure there were no issues relating to contacts.

 

[Gionnirocket]

I think he’s duly warned. Using Micahs’ book with attention to details like safety glasses and no jewelry are top issues. I was being super careful with a disassembled pack and didn’t notice a small piece of wire on my silicon work surface. I set the bare pack down and shorted out on the wire bridging cells. Instant POW! FLASH! Trashed a half dozen PF cells.

Those are important considerations.
But the mixing of cells is his current problem and not one with a solution