Mixing 18650 and 21700 cells in same parallel pack in battery pack ok?

[YoshiMoshi]

TL;DR Summary

Mix 21700 and 18650 cells in parallel?

I'm doing a 3S3P pack. My box can hold 9 18650s. It can also hold 3 18650s and 6 21700s. Is it ok to mix 18650s and 21700s in the same p pack? I would use the same 21700s and the same 18650s in each p pack. I would arrange them to maximize capacity.

So say 18650 cells are Model A cells (making up model number her for discussion purposes). All 3 18650 cells in my battery be model A.
The six 21700 cells are Model B cells.

One P pack would contain 2 model B 21700 cells and 1 model A 18650 cell.
I make three of these identical P packs described above. I would then put these P packs in series to create my battery pack.

Good idea? I don't see any issues. Each P pack would have roughly the same capacity and max current draw as each other (each cell slightly different even from same model) and would be identical.

 

[berkeman]

I have no idea what you are asking, but hopefully somebody else who is familiar with those battery designations will be able to answer. I just know that hooking batteries in parallel is problematic unless they are identical with identical past histories (age, charging histories, etc.).

 

[Baluncore]

The difference is form factor and capacity.
18650 = 18 mm diameter x 65 mm long.
21700 = 21 mm diameter x 70 mm long.

When using different capacity cells, or cells from different manufacturers in parallel, you will have reduced capacity before damage occurs.
When using cells in series, a battery management controller should be used.

Rather than two parallel batteries of different cells, use three series strings in parallel. That will require three series battery management controllers.

 

[Flyboy]

Yeah, definitely don’t mix different capacity cells in the same string. If anything, try to match your capacities as much as possible in each string.

 

[Baluncore]

Battery management.
https://en.wikipedia.org/wiki/Battery_management_system

 

[YoshiMoshi]

Thanks for the help!

Yea, they have different dimensions, and that 21700s have a larger energy density than 18650s. Some 21700 also have higher maximum continuous discharge current than 18650s

Just making things up here:
Model A 18650 cell has capacity of 3,000 mAh, maximum continuous discharge current of 25 A
Model B 21700 cell has capacity of 3,500 mAh, maximum continuous discharge current of 30 A

Say my container has physical dimensions that allows for the following configurations of 9 cells:
1) 9x18650s
2) 6x21700s + 3x18650s
3) You can expand 2 to get rid of any number of 21700s and replace them with 18650s, such that no more than 9 cells exist in the container

So physically, due to higher energy capacity, I figured option 2 would be the most optimum solution, and then I asked myself if combining cells of different sizes into a parallel pack would be a good idea or not. I would do a 3S3P configuration. Meaning I would have something like this:

1) If I had in parallel A-A-A my overall capacity would be 9 Ah, maximum continuous discharge current of 75 A
2) If I had in parallel A-B-B (shown in the picture above) my overall capacity would be 10 Ah, maximum continuous discharge current of 85 A
3) My container will not fit this If I had in parallel B-B-B my overall capacity would be 10.5 Ah, maximum discharge current of 90 A

I do suffer .5 Ah capacity and 5 A maximum discharge current, by not using all the same 21700 cell from the same manufacturer (Option 3), nine of them. But if that's all I suffer, is capacity and maximum continuous discharge current, and I'm willing to take those losses, than is that ok? Because in the end I still have a battery pack going with Option 2 over Option 1 (and my container won't physically fit Option 3)?

I would need a BMS in there (not shown) to have charge balancing. Each parallel pack would be the same configuration A-B-B. So each parallel pack/string would be the same.

I know that I shouldn't have in parallel batteries with drastically different voltages for long periods of time. That's a big "no no". Although people do try this to revive dead batteries, probably not to safe, creates a lot of heat I believe. I see that the nominal voltage of 18650s and 21700s is the same, so as long as I fully charge them than I think I should be ok on that front.

I know the capacities and maximum discharge current would be different, but I think I would be ok if I arranged them in the configuration shown in the picture above, each parallel pack of same configuration? Each instance of say the B cell would be slightly different. One could be 3502 mAh while another could be 3498 mAh. I would obviously do capacity matching, such that the overall capacity of each parallel pack was as high as possible, getting the optimal capacity.

Thanks for help in trying to understand this!

 

[Vanadium 50]

I think this is a terrible idea.

When you start out, and everything is putting our 4.2 V, no problem. Then as the batteries discharge, the two different kinds of battery have two different voltages. And you have them in parallel.

 

[Dullard]

The only way to 'safely' do this is to de-rate the cells (by a lot). That would seem to negate any 'capacity advantage' that might be had.

 

[Vanadium 50]

@YoshiMoshi I am curious as to your background. You've said that you are an electrical engineer in the past, but this design seems to have far, far more disadvantages than advantages. Including safety issues.

I would expect a EE to stay far away from this design.

 

[YoshiMoshi]

I see what you guys are saying. It does seem unsafe to me. I see lots of people on YouTube though, that "salvage" 18650s from old laptop batteries, and add them in parallel to their power wall. I thought that was a bit interesting. I see people claiming it's no problem adding 18650s in parallel with other 18650 cells, different manufacturers and models, different brands. Is this wrong? Which I don't know if that is necessarily true or not. That's when I then asked the question, under the presumption that is true and safe to do, then what's wrong with mix matching 18650s and 21700s in the same battery pack?

I think ideally cells in parallel is like forming one giant battery, were the currents add up and capacities. Maybe that's not a good way of looking at it "one giant battery"

 

[DaveE]

Maybe that's not a good way of looking at it "one giant battery"

A WELL DESIGNED battery pack can be viewed by the user as "one giant battery". To the EE that designed it, it's a network of small batteries. The only easy way is to match cells as exactly as you can, including testing and discarding outliers of the same type. There is a lot of information on the web about this standard application. It sounds to me like your in over your head here.

I see lots of people on YouTube though...

Seriously? Did you see their failure analysis, worst case and type testing? Or is it possible that they just threw something together and made a video. You can get away with a lot of bad designs for a while, until something bad happens. Did they start by saying hold my beer?

You are getting a pretty consistent reply from several people here. Why not accept the advice you've asked for?

 

[DaveE]

Since you said you were an EE, I'm sure you understand how important and common research is in the early phases of the design process. Here's one source (chosen sort of at random) that I think you should read and understand before you proceed. I would also suggest that you follow some of their links concerning battery bank charging issues.

https://batteryuniversity.com/article/bu-803a-cell-matching-and-balancing#:~:text=The capacity tolerance between cells,between cell balance and longevity.

 

[Vanadium 50]

Let me see if I get this straight. You're willing to toss away years of EE training in favor of what some random YouTuber posted?

Does this seem like a good idea?

 

[Richard_F14]

I think this is a bad idea because DIY battery packs first need to test the capacity, internal resistance, and other parameters of batteries of the same model, and then reassemble batteries with similar parameters to avoid possible malfunctions during charging or use.

In view of this, my suggestion is to build two different types of battery packs separately (such as the 3S battery pack built with 18650 batteries); Secondly, when using multiple battery packs together, add Schottky diodes to prevent current reversal; Finally, charge the battery packs separately during charging, so that both your charger and electrical devices can be shared, making it easier and more convenient.

Best wishes.

 

[Vanadium 50]

As I think about this more and more, I get more and more concerned.

If I had an engineer working for me and decided to dispense with good design practice in favor of what some YouTuber posted, I would ask for another engineer.

Similarly, if I had an engineer in over his head and he had to turn to a more senior engineer, I'd be OK with that. If he were in over his head and turned to a YouTube video, I'd be very concerned and wold again ask for another engineer.

Finally, there has been no mention of requirements. If I got a design with such obvious safety issues and they were not backed up by immutable requirements, I would again ask for another engineer.

I am wondering if leaving the thread open is the best idea.

 

[berkeman]

I am wondering if leaving the thread open is the best idea.

Nope.