Battery config for Xebra electric car

DiggsUt

Member
Joined
Oct 30, 2019
Messages
94
I have a ZAP Xebra (chuckles...) that's been given to me. It needs batteries. I am also an avid ebiker and builder and have a good stock and supply of 18650s. The car originally came with six 100 ah 12v lead-acid batteries and I'm trying to decide best configuration for an 18650 conversion. Do I try to make a 50p20s work which is causing me problems coming up with the correct BMS or do a run five 10p20s packs in parallel each with their own BMS but may cause charging issues when all charged off of one charger? This is stretching my knowledge a bit. I've asked the question over at DIY electric cars but they don't use 18650s much there so answers haven't been forthcoming.

Thoughts?
 
DiggsUt said:
I have a ZAP Xebra (chuckles...) that's been given to me. It needs batteries. I am also an avid ebiker and builder and have a good stock and supply of 18650s. The car originally came with six 100 ah 12v lead-acid batteries and I'm trying to decide best configuration for an 18650 conversion. Do I try to make a 50p20s work which is causing me problems coming up with the correct BMS or do a run five 10p20s packs in parallel each with their own BMS but may cause charging issues when all charged off of one charger? This is stretching my knowledge a bit. I've asked the question over at DIY electric cars but they don't use 18650s much there so answers haven't been forthcoming.

Thoughts?

Sounds like a pretty good project! I assume 6 12v lead acid means it's a 6 in series so a 72volt system. So a 20S (or maybe 21S) for 18650 sounds about right. What the max amp draw is needed to match it with the correct 18650?

What's your budget for the BMS? There's plenty that will handle 20S, like batrium, or recbms, or even cheaper chinese balancers like https://www.aliexpress.com/item/32989586125.html will suffice. A 20S would mean a max of 4.2V*20=84V so as long as the built-in charger doesn't go over during the bulk charge or otherwise just use an external charger instead.

Also ever considered using other cells like the leaf modules?

Good luck!
 
I'll probably stick with 18650s just because I have most and have a free supply. The BMS budget is what ever it takes for the best setup for the project. I'm leaning towards five 10p20s (or 21s) in parallel as being easier and safer to build and maintain. The BMSs for the smaller packs are relatively inexpensive and if one pack ever fails the car still runs enough to return home. I don't know what max draw is from the motor but if I work in parallel I can always add more packs. I'll probably replace the charger with what works best for the 18650s although from what I understand the built-in charger is multi-stage but whether it matches best practices for lithium charging I don't know yet but doubt it. The charging across parallel packs is an unknown for me.

Thanks!
 
Lead acids have a range of about 10.5V - 15V.
Charged: ~90.0V
Discharged: ~60.8V

So if you went with 18650 cells, your figured would be:
22s: 70.4 - 90.2V

If you go with 11.1V - 14.5V instead, then it would be:
Charged: ~87.0V
Discharged: ~66.6V

18650's would then be:
21s: 67.2V - 86.1V

Btw, I'm going with a voltage range per cell of 3.2V - 4.1V to get the a boost in cell longevity. You can go to 3.2 - 4.0V, but then you'd need 22s (70.4 - 88.0V) but also gain even more life. However, this starts lifting the bottom voltage up to high. You "can" go down to 2.8V with most cells, but you won't get much capacity from 3.2 to 2.8 and the drop off is pretty quick.

And, the common nomenclature here for battery design is series first, then parallel. So, 21s10p, for example. A 10p21s would mean you need 21 bms units, and I doubt that's what you mean :p
I mention this because perhaps on other forums they put it the other way around. Here it's known to be series/parallel, and others reading may get confused later on (or perhaps yourself reading others threads).

21s20p is 400 cells. That's a lot for a pack. So yes, I would agree on the 20p range and not the 50p or so as a single build. They still need to be accessible and removable fairly easily if need be.
Use 3 parallel packs of 21s20p, and then use balancing connections to hook them all together. This will allow 1 bms to handle the whole lot. It's gonna be a mess of wires, though, unfortunately.
2 main wires pos/neg
22 balance wires connecting each pack together

The balance wires only need to handle a few amps, maybe <10A as balancing shouldn't draw a lot of power, especially if the cells are good and well balanced when building the packs.

Ultimately, I'd suggest to go with high power cells, like those used in power tools.
What's the kWh ratings of the electric motor(s)? If they are 1kWh running at 72V (nominal), then you're looking at close to 14A. With 50p, 14A shared across those would be 0.28A per cell. This is well within most cells capabilities. The high voltage makes this doable.

Replacing the charger would probably be a good idea, unless you can program it to function with different charge settings.
 
Korishan said:
What's the kWh ratings of the electric motor(s)? If they are 1kWh running at 72V

It's a 5 kWh motor at 72v so that presents some other issues.
 
Ok, so that's about 70A at full throttle. If you calculate a surge, full throttle while dead stop for instance), you'd need double the current of 140A. To have a safe working load on the cells of around 500mA under full surge load you'd need 280p (140 * 2).
It's possible you will never see that high of a current draw.

What I would recommend, if possible, is to fully charge your existing batteries if you can, then put a clamp amp meter with max hold and launch that car from a dead stop to full throttle. This will give you the max amps that thing can pull. Even though the motor is rated at 5kW, it's possible this design setup may never see that much usage. Maybe the vehicle will only use 4kW and the 5kW was designed in the build as safety margin. shrugs

Give it a shot
 
There's two parts to this, the first is how much current during the starting inrush, and second is continuous at full throttle.

I'd think you'd want the battery to be able to handle the starting current, which I would put more at 4x than 2x. So 70A that Korishan mentioned would mean the battery capable of handling 280A. Now most batteries are capable of handling 1C continuous. So if a battery is rated at 2500mah, it's capable of handling 2500ma continuous. Some batteries like those made for powertools and ebikes can go 2-5C continuous so they handle those surges better. The more current you extract per battery, the more the voltage will drop. So at 60P, an inrush will mean 4.6A per battery. If you're at a dead stop and at 50% capacity, by starting the car the voltage might drop to the point it may trigger the low voltage on the BMS and you find yourself stuck at the light! At 120P you're running at around 1C during the inrush which is probably fine.

As far as cruising at full throttle, it's wise to run it at 0.2C (so with the 2500mah battery it's 500ma). This will maximize your efficiency as well as reduce the heat generation of the battery. But I can see running it at 0.5C or close to 1A without issues. Any more you run into the risk of the battery voltage dropping to the point your BMS will kick in and shut down the car/ebike. So at 70A during full throttle, a 60P pack is about 1.16A. A 120P will be around .58A.
 
So I'd be looking at 2500 cells for the project. I was hoping to do it for about half that. This may be stretching my supply of 18650s more than a bit. Hmm.....
 
Many ebike, hoverboard and electric scooter batteries are 10s, 36V nominal. Why not put 2 10S packs in series for 72V. You could either use off the self packs or build your own. 10S BMS's are also common.
 
Lucan said:
Many ebike, hoverboard and electric scooter batteries are 10s, 36V nominal. Why not put 2 10S packs in series for 72V. You could either use off the self packs or build your own. 10S BMS's are also common.

Still would need the quantity of cells.

I was thinking that all packs are parallel packs and the series connect them all together. 280p (if he goes with that size) is a pretty good sized pack. Trying to make that a series in a pack would be hefty.
However, there could be 2 bms units that are connected as in pseudo series, but they would not be able to function as gate keepers to the absolute Pos/Neg connections. It'd be best to have 1 bms that can handle the full run if possible. Less headache and tracing issues down later down the road (pun partially intended ;) )
 
Back
Top