DIY battery building questions

[jonpad01]

Hey guys,

A few months ago a I built my first 20s7p battery using molicel p42a cells. I used the copper nickel sandwhich method using a malectrics spot welder and a 12v car battery. After reading about the differences between the two, I now wish I had bought the kweld. I had some trouble getting good welds as the car battery would drain and had to recharge it a few times.

On my BMS I have a cell group that is about 0.02v lower than the rest of the groups at 90% charge. The rest of the groups are all within 0.005v of each other. I noticed when riding hard last night I managed to fully drain the battery to the point where the BMS shut off. I looked at the voltages and that group was at about 2.8v while the rest were at about 3.2v. That group has returned to the usual 0.02v offset after recharging the battery.

Please note that I dont know very much about lithium batteries.

So I have a few questions, do I need to worry about this cell group?

Lets say one of my welds was really bad and broke loose, what would happen? The battery is secure but I have already dropped this bike down a hill.

This last question is about discharge current. On the datasheet it says these batteries are rated for 45 amps, but Ive seen real world testing that puts them closer to 30 amps before overheating. What would happen if I were to draw 50 amps from them for a very short period of time, about 5 seconds, is it just the heat that damages them? Would they even get hot enoguh to worry about? Currently my controller is set to about 300 line amps, which is about 42 amps per cell. Im not even sure I can draw that much because the bike pulls a wheely and I let go of the throttle.

 

[harrisonpatm]

On my BMS I have a cell group that is about 0.02v lower than the rest of the groups at 90% charge. The rest of the groups are all within 0.005v of each other. I noticed when riding hard last night I managed to fully drain the battery to the point where the BMS shut off. I looked at the voltages and that group was at about 2.8v while the rest were at about 3.2v. That group has returned to the usual 0.02v offset after recharging the battery.

Please note that I dont know very much about lithium batteries.

This is not terribly abnormal. It's not ideal, it means that you have a parallel group that has less capacity than the rest. But if you google the discharge curve for lithium ion, you'll see that there's not much of a difference in capacity between 3.2v and 2.8v; the voltage drops off hard once you get lower than 3.3-3.2v.

So I have a few questions, do I need to worry about this cell group?

"Worry" is a strong word, I'd say you'd want to keep an eye on it and make sure it doesn't continue to age faster than the rest. The fact that it's coming back up to a 0.02v deviation at a full charge is good, maybe pay attention the next time you run the battery down to 0% and see if the deviation remains same or similar.

Lets say one of my welds was really bad and broke loose, what would happen? The battery is secure but I have already dropped this bike down a hill.

It's entirely possible that a weak weld came loose and is causing that parallel group to show less capacity than it should. The only way to know for sure is to visually inspect the welds. You may decide it's worth it to check (incidentally, this is why I chose to use a transparent case for my 840-cell build). Alternatively, it's possible that that group of cells just had 1 or 2 lower performing, lower capacity cells to begin with. Where did you get your cells, are they new? Did you individually test all 140 cells to make sure they're as identical as possible? Even small differences in age and performance can add up. And like I said, even though a 0.4v deviation sounds bad, and certainly isn't ideal, if it continues to only happen at the fully discharged end of the curve, it might just be a symptom of the health of your cells. Only way to know for sure is to do a capacity test on each parallel group individually. You may decide it's not worth the work.

This last question is about discharge current. On the datasheet it says these batteries are rated for 45 amps, but Ive seen real world testing that puts them closer to 30 amps before overheating. What would happen if I were to draw 50 amps from them for a very short period of time, about 5 seconds, is it just the heat that damages them? Would they even get hot enoguh to worry about?

We've discussed in multiple threads on the forum the case of discharge ratings for cells. Just because a cell can technically output 45 amps, doesn't mean it should, doesn't mean it's good for the cell. And it doesn't mean the connection is necessarily good for 45 amps continuous.

Test of Molicel INR21700-P42A 4200mAh (Gray)

Have you seen independent tests like this yet? You'll see at 30 amps, the cell is reaching 50-60 degrees C. In my opinion, that's hot enough to worry about. And you don't have 1 cell pulling 30 amps, you have 140. Is there enough airflow in the center of your pack to dissipate some of that heat?

Currently my controller is set to about 300 line amps, which is about 42 amps per cell. Im not even sure I can draw that much because the bike pulls a wheely and I let go of the throttle.

That's a good question. If the bike is already wheelie-ing, why are you considering pulling more power?

 

[jonpad01]

This is not terribly abnormal. It's not ideal, it means that you have a parallel group that has less capacity than the rest. But if you google the discharge curve for lithium ion, you'll see that there's not much of a difference in capacity between 3.2v and 2.8v; the voltage drops off hard once you get lower than 3.3-3.2v.


"Worry" is a strong word, I'd say you'd want to keep an eye on it and make sure it doesn't continue to age faster than the rest. The fact that it's coming back up to a 0.02v deviation at a full charge is good, maybe pay attention the next time you run the battery down to 0% and see if the deviation remains same or similar.

It's entirely possible that a weak weld came loose and is causing that parallel group to show less capacity than it should. The only way to know for sure is to visually inspect the welds. You may decide it's worth it to check (incidentally, this is why I chose to use a transparent case for my 840-cell build). Alternatively, it's possible that that group of cells just had 1 or 2 lower performing, lower capacity cells to begin with. Where did you get your cells, are they new? Did you individually test all 140 cells to make sure they're as identical as possible? Even small differences in age and performance can add up. And like I said, even though a 0.4v deviation sounds bad, and certainly isn't ideal, if it continues to only happen at the fully discharged end of the curve, it might just be a symptom of the health of your cells. Only way to know for sure is to do a capacity test on each parallel group individually. You may decide it's not worth the work.

We've discussed in multiple threads on the forum the case of discharge ratings for cells. Just because a cell can technically output 45 amps, doesn't mean it should, doesn't mean it's good for the cell. And it doesn't mean the connection is necessarily good for 45 amps continuous.

Test of Molicel INR21700-P42A 4200mAh (Gray)

Have you seen independent tests like this yet? You'll see at 30 amps, the cell is reaching 50-60 degrees C. In my opinion, that's hot enough to worry about. And you don't have 1 cell pulling 30 amps, you have 140. Is there enough airflow in the center of your pack to dissipate some of that heat?


That's a good question. If the bike is already wheelie-ing, why are you considering pulling more power?

Thanks for all the helpful info, after charging the bike it looks like the group I'm worried about (group 19) has been slightly overcharged. I will keep an eye on it.

Where did you get your cells, are they new?

I while ago batteryhookup had some unused robotic battery packs, thats where I salvaged these batteries from. About the only thing I did was test the individual cell voltages to make sure none of them were way off. I believe though these are basically brand new batteries, aside from the damage I may have caused building this battery pack heh.

That's a good question. If the bike is already wheelie-ing, why are you considering pulling more power?

I was thinking about it in the opposite direction. Currently I like the power I'm getting, but if pushing 300 amps through this pack for a few seconds during hard acceleration is going to damage the battery then I think I should turn it down.

 

[harrisonpatm]

Thanks for all the helpful info, after charging the bike it looks like the group I'm worried about (group 19) has been slightly overcharged. I will keep an eye on it.

So group 19 was the one that dropped to 2.8v before the others? And now it's a bit higher than the others? Sounds like it's just a weak group. Could be because of poor welds, could be because you just have 1-2 bad cells. Only way to know for sure is too physically inspect them and/or capacity test each parallel group individually.

was thinking about it in the opposite direction. Currently I like the power I'm getting, but if pushing 300 amps through this pack for a few seconds during hard acceleration is going to damage the battery then I think I should turn it down.

Excellent plan

 

[jonpad01]

So group 19 was the one that dropped to 2.8v before the others? And now it's a bit higher than the others? Sounds like it's just a weak group. Could be because of poor welds, could be because you just have 1-2 bad cells. Only way to know for sure is too physically inspect them and/or capacity test each parallel group individually.

Here is the snapshot i took about 2 hours after charging the battery last night, note that my charger is set to stop at 4.05v. I also disabled the balancing feature to see what that group would do when sitting over night.

Total battery voltage reported as 81.39v



Here is the snapshot about 10 hours later. Total battery voltage reported as 81.34v.



Most of my groups are within 1-2mv, basically the BMS's maximum resolution. Group 19 was overcharged by about 8mv, and then dropped a total of 17mv which puts it as undercharged by about 8mv. Is this anything to worry about? Should I let the battery sit longer and see if that group keeps dropping in voltage?

My balance setting triggers when group voltages are greater than 0.010v in difference, maybe I should change it to something like 0.030v?

 

[harrisonpatm]

My balance setting triggers when group voltages are greater than 0.010v in difference, maybe I should change it to something like 0.030v?

That's not a terrible idea. I stand by what I said earlier, that it looks like you just have one weak group, which isn't the worst thing in the world. See if you can get repeat data for the bottom of the charge curve.

 

[jonpad01]

Thanks for your help with this, I am on vacation next week so I should have plenty of time to ride it until it dies

 

[jonpad01]

Heres another shot from tonights ride, battery wasnt quite dead, maybe 10% left?

 

[harrisonpatm]

Heres another shot from tonights ride, battery wasnt quite dead, maybe 10% left?

View attachment 357930

That's still tracking with what you've been saying. So it's still up to you if you want to open it up and test.

 

[jonpad01]

When I was building the battery, I was welding the positive side of a cell and I missed the center button, which blew molten copper and nickel down into the crevice under the positive plate. I think i had to pry a piece out before continuing. I'm pretty sure it was on a group that was 2nd from the end. I probably should have stopped and replaced that cell, oh well.

I'm hoping it won't get worse until the fall/winter when I stop riding.