Multiple battery chemistry in one pack

R

RobboCop

100 mW
Joined
Apr 29, 2015
Messages
40
Location
Perth
My brother in law builds quadcopters and we got talking about electric bikes.

He seemed to think it might be possible to run a battery pack that consisted of two different types of battery.

You guys know a hell of a lot more about this sort of thing than I currently do and I haven't seen this happen in other bike builds, which gives me the feeling that it may not be possible/or is a bad idea.

The concept is can I build a battery pack that is say 21S8P of 18650 cells and 21S1P LiPo (in parallel with the 18650 set)? The idea behind it is that the LiPo can give you the explosive power for acceleration and then be 'evened out' by the 18650s on the cruise. If its possible you could enjoy the range of an 18650 set with the explosive power and high C rates from LiPo. For that matter you could probably lower the series count in favour of more parallel banks given the C rates available on the LiPo side.

Again, I am really new on the EV scene, so this may be a really obvious no no.

Cheers,

-Robbo
 
Assuming you're talking about 18650 cells that charge to 4.2V and rc lipo cells that charge to 4.2V then there's no problem paralleling them. Although personally I think 18650 cells are best used as flashlight batteries and would never build a battery pack for an ebike using them.
 
Yes, I was going to use the same voltage cells.

As to your point with not liking 18650s, I guess the main reason I like them is that they are a lot safer, that said a mixed pack will be as volatile as the most volatile component. In this case that would be the LiPos. I guess it begs the question that if I am accepting the risks associated with having LiPo's in the system, why not just run LiPo by themselves?

Anyone got any thoughts?
 
wesnewell said:
Assuming you're talking about 18650 cells that charge to 4.2V and rc lipo cells that charge to 4.2V then there's no problem paralleling them. Although personally I think 18650 cells are best used as flashlight batteries and would never build a battery pack for an ebike using them.
Click to expand...

So they're good enough for Tesla cars, and most laptops.. but not for an ebike thats right in the middle of the two? :mrgreen:

:pancake:
 
Op, it would be an interesting experiment to get some more life out of a 18650 pack or something. I'm not sure if it'd be worth it or not! Wouldn't the whole pack also be limited by it's weakest link, the older or least capacity? Would they use different chargers?

Unless ^^^ I agree with you, just run all lipo if you need to run any. All or nothing lol.
You can always diy add bms to your lipo for an extra layer of protection of course.
 
One thing I have against 18650 cells is the connectivity of them in an ebike environment. It makes repair of a pack almost impossible for the average user. And most have really low discharge rates comparatively. I've thought about something like thin wall pvc tubes that one could just fill with the number of cells they want in series and then use these tubes as cell groups to make up a pack. The ends would be spring pressure connectors that could be removed for easy cell replacements much like that of flashlights. Just configure the tubes however you want and plug them together.
 
I saw one build here that did exactly what you are looking into. I will try to find back to that thread and post a link here so you can get in touch with the right person. For my ears this should be possible, and you should get the pro's you mentioned. Other then safety issues I think it should be fine. Might wanna check again about using 1p for the lipo, a little more might be needed.


Yeah it was done by roadrash, here you can see.
https://endless-sphere.com/forums/viewtopic.php?f=6&t=70148#p1061393
 
wesnewell said:
One thing I have against 18650 cells is the connectivity of them in an ebike environment. It makes repair of a pack almost impossible for the average user.
Click to expand...

And if you build your pack with high quality cells and take care of it, you probably won't ever have any issues. A pack built of new, respectable cells that are all identical, and run through the same life cycle, is unlikely to ever need individual cell replacement - if one is bad, the rest of the pack is probably also not in great shape, so replace all of them.

If you're building packs out of scrapped junk cells, sure, pack repair matters.
 
one thing to keep in mind is how much current you're going to draw from the total battery assembly.

the weaker one will eventually sag, and then a larger amount of current will be drawn from the stronger pack.... if this is a lower power bike, I wouldnt worry too much about this... but if you're pulling a good amperage off the line, you'll have to keep it in mind.
 
if the two packs are connected together in parallel at B+ and the P- of the BMS then the two packs deliver current equivalent to their respective storage capacity. so equal capacity packs deliver 50/50 to the output. something like 10Ah/5Ah is 66.6%/33.3%
 
That can't possibly be correct because it doesn't take into account pack internal resistance.

My understanding is that two packs in parallel with different internal resistances (but the same discharge characteristics) will behave as one large battery with pack IR being calculated in the same fashion as you would calculate parallel resistors.

For example

A pack assembled of NCR18650GA cells configured as 20s8p (28ah) has a pack IR of 95mOhm
20s 5ah of Nanotech A-Spec has an IR of 30.5mOhm

Connected in parallel this forms a battery of 32.5ah with an IR of 23mOhm. This should support a 100amp draw with a total pack sag of 2.3v, delivering 8150watt at HOC SOC. Contrast that with the same 100amp draw from a 31ah pack assembled just from NCR18650GA cells with a pack IR of 85mOhm which would sag ~8.5v and deliver 7550watt, presumably losing the rest to heating. To get to the same power capability a homogeneous NCR18650GA pack would need to be 20s33p and weigh ~3x the hybrid pack.


If I am wrong, please somebody tell me/prove it!
 
you are wrong. you should learn to prove your own arguments and then you would not be wrong as much. ever here of kirchoffs rule? it is important for you to understand where all of your assumptions are wrong.
 
I absolutely agree it's important to understand where I'm wrong, however I'm clearly too dense to determine which assumption I have made is incorrect.

I'm unable to find any applicable information to parallel connection of batteries of different capacities and and internal resistances, other than statements regarding current sharing among batteries of equivalent capabilities which would logically behave in a similar fashion to what you described. I am familiar with Kirchhoff's rule and can't see how what I'm proposing violates it.

Having searched extensively and still coming up dry as to concrete calculation or practical experience I would be very appreciative if you could share for the benefit of the collective knowledge of the forums. Clearly this is an area full of misconceptions.
 
when you draw a load (current) on a battery, the voltage drops. This is due to an internal resistance of the battery.
the more current you draw, the more the voltage "sags" down. You can use Ohms law to figure it out some.... but the more it sags for a load, the larger its internal resistance.

Different batteries sag at different rates. You may see a "C" rating on some packs... a 5ah lipo pack with a 20C rating is rated for 100A continuous. (5*20=100) meanwhile a 5ah lifepo4 pack may only be rated at 2C.. a 10A rating.
So lets say you have pack A and pack B... just to make an obvious example, lets make pack A the lipo above and pack B the lifepo4.
We start out with them both charged to the same voltage. Connect them in parallel for a 10ah mixed pack, and go to ride on our bike.

lets say you start pedaling first, and lightly roll the throttle on for a 20a peak draw... It splits off and draws the 10a from each.. roughly equally. a little more from the lipo pack because its resistance is lower.
now you stop, and nail it from a stop... drawing max stopped amps of whatever you're set to.. say 40a. now its hammering the lifepo4, and the lipo is doing almost all the work. your lifepo4 battery will also get hot. your lipo battery will drain faster, as also trying to raise the voltage of the lifepo4 to match it, and thats being turned into heat.

to explain that lets think of them as two batteries, not connected.. both at 4.0 volts per cell.... we draw 10a from the lipo, and it sags to 3.95v... draw 10a from the lifepo4 and it sags to 3.90.. not a huge diff.. the lipo is going to "give" a little more, and so if they're in parallel the system will appear to be at 3.93 or so.

now same thing at the higher draw. both at 4.0 volts per cell.... we draw 20a from the lipo, and it sags to 3.85v... draw 20a from the lifepo4 and it sags to 2.00.. a huge diff.. the lipo is now trying to feed the bike AND it is trying to maintain the voltage on the lifepo4 cell. The lifepo4 becomes nothing more than a big resistor sucking power away from the lipo instead of adding it to the bike motor.... this huge resistance is also going to create heat. people have burned up packs doing this.




does this make sense? I typed it out a bit fast... but is it any clearer?
 
I'll get what you are saying, but maybe the lifpo4 should be out of the equation? Maybe it is better to compare pack A as a hi qual 18650 li ion pack and then pack B as hi quality hi C rated lipoly batteries, just like raodrash did?

He already got the lipo's then he got a 20s20p 18650 pack (50Ah), and combined that with 20s2p 10Ah lipoly.

Will his pack have the same outcome? One pack working is ass off to keep the other pack on par? Or will the lipoly pack take most of the punch and deliver more current faster when accelerating?
 
macribs said:
I'll get what you are saying, but maybe the lifpo4 should be out of the equation? Maybe it is better to compare pack A as a hi qual 18650 li ion pack and then pack B as hi quality hi C rated lipoly batteries, just like raodrash did?

He already got the lipo's then he got a 20s20p 18650 pack (50Ah), and combined that with 20s2p 10Ah lipoly.

Will his pack have the same outcome? One pack working is ass off to keep the other pack on par? Or will the lipoly pack take most of the punch and deliver more current faster when accelerating?
Click to expand...

The higher C pack will be working overtime to both power the bike, AND cook the lower C pack IF the current draw exceeds the abilities of the lower pack.
If you stay within the limits of the lower C part of the pack, it will work.
If you're borderline with the amp draw near the rating, its a bit of a crapshoot... it may work, but have a slightly smaller than expected range.

overall I would not do it. I would however swap packs. unplug one, and plug the other one in.
 
Ohbse said:
I absolutely agree it's important to understand where I'm wrong, however I'm clearly too dense to determine which assumption I have made is incorrect.

I'm unable to find any applicable information to parallel connection of batteries of different capacities and and internal resistances, other than statements regarding current sharing among batteries of equivalent capabilities which would logically behave in a similar fashion to what you described. I am familiar with Kirchhoff's rule and can't see how what I'm proposing violates it.

Having searched extensively and still coming up dry as to concrete calculation or practical experience I would be very appreciative if you could share for the benefit of the collective knowledge of the forums. Clearly this is an area full of misconceptions.
Click to expand...

your primary assumption that a battery cell is a passive device that is modeled as a resistor. no idea why people use this but it just wrong.

a battery cell is an active device.

internal resistance will not control how much current is delivered by each battery in parallel. it is a derived number of no relevance in this discussion.

the output voltages remain identical by kirchoff's rule so the SOC of each battery is identical so the amount of current delivered is identical.
 
dnmun said:
your primary assumption that a battery cell is a passive device that is modeled as a resistor. no idea why people use this but it just wrong.

a battery cell is an active device.

internal resistance will not control how much current is delivered by each battery in parallel. it is a derived number of no relevance in this discussion.

the output voltages remain identical by kirchoff's rule so the SOC of each battery is identical so the amount of current delivered is identical.
Click to expand...

Internal resistance is wildly variable value depending on SOC, temperature, nature of the specific cell chemistry, physical topology of the cell etc etc however for the purposes of indicative calculation of pulse power potential I think it's very relevant.

Given this example:

parallelexample.jpg

You're saying that the laptop cells will deliver 2/3rds of the 10a load?
 
I've got a decent 100k count DMM with usb logging and a powerlab8 which can do 40amp discharge into a bank of lead plus lots of cells/lipo to choose from, I need some hands on observations because I can't get passed the behavior you're describing. I'm more than happy to be wrong, just interested in learning :) thanks
 
If you don'nt have any circuit to control it, the less resistive pack, read newer, will discharge fastter (current flows thought the less resistive path), but not probably not noticeable. Also, simply paralleling them will simply divide current thought it as they were resistors, most of the current will go thought the less resistive battery. If you want burst power, best thing to do, but not easy, is add a parallel big capacitor, but you'll also need some eletronics to control when you can use it. it might have been done before. If paralleling different chemistries always remamber that lowest voltage pack must always be the voltage limit.
 
Ohbse said:
I've got a decent 100k count DMM with usb logging and a powerlab8 which can do 40amp discharge into a bank of lead plus lots of cells/lipo to choose from, I need some hands on observations because I can't get passed the behavior you're describing. I'm more than happy to be wrong, just interested in learning :) thanks
Click to expand...

since you are approaching this with a little more intelligence than the others i will also point out that it is the capacity of each bank of cells, not the voltage or the internal resistance which i have tried and tried to explain is a derived value, that determines the split in discharge.

so you should be able to discharge both sides down to the cutoff point of the BMS and the total discharge from each side will be identical. but then the side with more capacity will start pushing current back through the drains on the BMS and it will reset the BMS and that small amount of charge will be pushed back out again and the BMS will shut off for LVC and then the pack that has not hit LVC will push current back into the side which already hit LVC, and so on ....

i still do not understand why you cannot see that a battery cell is an active device. you must know the difference between active and passive devices and their behavior.
 
So I ran the test as detailed in my shitty diagram above, except with two inline meters on each leg. The shunts on the measuring device adds some resistance and both my DMM's vary slightly, so I ran same test with meters swapped to even out the inaccuracy somewhat. It all stacked up and agreed with the discharge on the Powerlab. I also used 3x laptop cells as I already had a spot welded grouping of three.

If dnmun is correct I would expect to see 75% of the current being supplied by the group of three low C cells.

Discharging at 10a which is the limit of my meters but enough that voltage sag on the shitty cells comes into affect I saw 55% of the current being delivered by the HE4 cell and 45% from the 3x shitty LG laptop cells. This is *exactly* in line with their measured impedence/IR. Pack IR is exactly the result of 1/a + 1/b = 1/c.

I will run another test tomorrow with a 4s turnigy 20c in parallel with a bunch of shitty laptop cells and see if I can replicate the same thing. I will probably need a couple of watt meters with lower resistance shunts to do larger scale testing.

If anybody has any requests for a specific scenario I'm happy to try put it together.
 
You must log in or register to reply here.

Similar threads

B
ADVICE -Replacing/Repairing DeWalt Battery Packs.
Replies
6
Views
2,340
jas67
J
A
Lipo battery pack build question
Replies
4
Views
1,479
Atxraider
A
C
Common Tool Pack Battery Idea: Need input on Mixed cells
Replies
5
Views
1,427
CONSIDERABLE SHOUTING
C
F
Battery pack made from several small modem packs.
2
Replies
32
Views
3,561
dogman dan
dogman dan
methods
Lead Acid JumpStart + Laptop Battery
2
Replies
47
Views
3,302
nicobie
nicobie
Back
Top