48v lto design

[lxgoldsmith]

I currently have a 48V 21A motor/controller kit from 5 years ago, mounted to a mongoose ledge 2.1 mountain bike. I am in the process of replacing the brakes, tubes, tires, and chain.

I built my previous battery pack from recycled cells. I tossed that battery a while ago. I've since graduated college and now have the time and money to make a new battery.

I want a battery pack that can:
run a minimum of 9mi/15km between stops
charge quickly.
take abuse/inactivity with low risk.
a good size to mount above the rear wheel.

Weight is not an issue. I have to take the bike in and out of my apartment, but that's about all. The battery can go in the trunk of my car for a mountain biking trip.

For electrical tools, I have:
a 60W soldering iron
loads of braided 16awg speaker wire
loads of 22awg color-coded wire
several arduino nano that I can use for cell tests

I'm currently looking at a prismatic lto cell on alibaba -
https://www.alibaba.com/product-detail/XLD-Factory-wholesale-prismatic-LTO-battery_62200327467.html

20 cells could make a raw battery that's about 9.5 * 5 * 8.5 inches in volume and about 26.5lb in weight. That leaves room on the front for the control circuitry.

I would prefer to design a bms using some relays, a mux, and a capacitor to sample voltages. It would be easier to test bms designs with low-risk cells. The use of a microcontroller would allow me to log data to a microsd card.


I'm looking for insight on the choice of LTO and reading material for bms design.

 

[john61ct]

The energy density of LTO is so low, I really can't imagine it being used for propulsion generally, and especially not a bike, and most especially if any kind of range is needed.

 

[lxgoldsmith]

anything would be better than the first battery I made. A battery made from recycled cells took ages to build and had horrible capacity. It died way too fast, so I want to go for something that won't die and won't keep me on my toes.

I'm also looking at pouch cells for higher density, but the best I've got is a soldering iron, which may not be ideal for soldering the tabs.

 

[Syonyk]

I don't understand why you'd consider LTO for an ebike battery.

If you want a boring, safe chemistry that's more energy dense than LTO by a good margin, use LiFePO4.

LTO is useful if you care about getting tens of thousands of cycles while, by any other chemistry's metrics, abusing the battery horribly with crazy charge rates (think 10-20C - charging the battery fully in 3-6 minutes). LTO shrugs that off and doesn't care.

It's a useful chemistry for something like a bus that is quick charging at stations, all day, every day, and has 30 seconds to swallow half a route's charge. But for something like an ebike? It's just an expensive, low density chemistry. It's safe, certainly, but so is LiFePO4.

 

[amberwolf]

I currently have a 48V 21A motor/controller kit
<snip>

I'm currently looking at a prismatic lto cell on alibaba -
https://www.alibaba.com/product-detail/XLD-Factory-wholesale-prismatic-LTO-battery_62200327467.html

20 cells could make a raw battery that's about 9.5 * 5 * 8.5 inches in volume and about 26.5lb in weight.

20 cells, if all in series, is only going to be about 46v nominal, rather than 48v nominal. If your controller has a nonprogrammable LVC, it may cutoff when the pack still has significant usable power (depending on the SoC curve of the cells you end up using).

You'd probably want to use at least 21 series cells, to get closer to the nominal voltage of the pack your controller is expecting.

I don't know how many parallel cells you'll need; that depends on your range needs (since the LTo cells can probably easily handle the current even in 1p)

There are a number of threads (many by or posted in by Inwo, who tests/uses/recycles various LTO packs / cells) here on ES about LTO / Lithium Titanate / LiTi if you poke around, that show various BMS and cell options and information / discussion.

 

[john61ct]

I would think all the infrastructure around an LTO install would need to be pretty wide-range user-custom adjustable, given the huge range of C-rates tolerated, even encouraged.

At high C-rates, LVC can go as low as 2.2Vpc, while at sub-C as high as 2.4V is required.

I've seen charge termination all the way from 2.6 - 2.85V

but apparently anywhere above 2.53V isolated / resting is functionally top SoC.

None of this to be construed as my thinking such low density is appropriate for an ebike.

 

[flippy]

i dont see any reason to use LTO. its really the worst selection you can make in this case.

if you only have a size constraint as mentioned above you can fill that space with a quite manly regular 18650 battery that can take loads well into the kW range and with enough capacity to never run out so you dont have to create a problem (fast charging) when its simply not needed. you can drive all day with a proper sized pack and at best charge it with a 1kW charger to top it up a bit during a break from the car engine if desired. it would also be the cheapest solution in the long run.

a 28lbs battery made from cells used in power tools (those can take high charge rates) you can make a 1.5~1.8kWh battery. that is enough to drive more then 100 miles on a bike without much pedaling at all.
if you do mange to drain it more then you like you can use a fast charger (that can run from your car) to charge it with 1kW and get it topped up in a lunch break.
then charge it (paritally) when you get home and park it in the garage and finish charging the day before you leave and you will have a battery that will last probably longer then the bike.

 

[lxgoldsmith]

pretty sure I could make a worse 48V battery with sla, so I don't see where all the hate is coming from. I get that you all go for more power and more distance and minimizing the risk of decay or fire is "boring", but come on. The worst-case scenario for a lto pack is repurposing it.

A low energy density doesn't seem all that unacceptable.

 

[john61ct]

pretty sure I could make a worse 48V battery with sla,

Saying "LTO is not the worst chemistry for this use case"

is of course true, you can always find worse choices, right down to lemons and potatoes.

> so I don't see where all the hate is coming from

I'm sure nothing but good intentions to help you here, despite rough "straight talk" and strongly held opinions.

> A low energy density doesn't seem all that unacceptable

Do the math yourself for the Ah and Voltage you want, cubic cm volume and weight in kg, for

NMC or NCA at 260Wh/kg

vs LFP

vs LTO

Publish the results here and **then** let us know what you think.

 

[dustNbone]

If you're looking for a very safe chemistry for normalish ebike use, LiFePo4 is well proven and despite a somewhat lower energy is still quite suitable for ebikes.

The only reasons I can see for going with LTO instead are wanting to charge the living crap out of it (like 10C+ charging, 5 minute charge from empty), or the need for tens of thousands of charge cycles. 10,000 charge cycles is like 28 years of daily cycling or something.

If I were ever to use them it would be for the fast charging abilities, and it would be a small pack to allow it's ability to be fast charged to somewhat compensate for it's much higher weight. Plus the weight of the charging system I'd need to take advantage of it's charging abilities. Someone who lived and worked in a dense urban area with lots of charging stations could probably make this work really well.

No one is saying LTO isn't good, it's pretty great at what it's made to do, but I just don't see you doing any of those things.

 

[lxgoldsmith]

I've decided that I'm going to get LTO cells regardless for testing custom circuits, and running the bike will be secondary.

The consensus seems to be that LTO is bad for EV, but not how bad. By spec, it looks like a 8-30% drop in energy storage from LFP counterparts (pouch vs 'prismatic' vs cylindrical appears to make a big difference). On the other hand, being able to charge from 0v at low risk seems like a big selling point.

Dropping to 0v is exactly how I killed my last battery. If LFP could survive the same torture, I would rather use that. I haven't seen it discussed for LFP, though.

If a less space-efficient battery doesn't work well on a bike, it will still work with a small solar/ups circuit, which is also on my to-do list. In the worst case, my budget would have been the same. It will be easier to start with the chemistry that might work for both and see how it goes.

 

[999zip999]

I don't know how those cells react to a 60-watt soldering iron maybe best spot weld but I'm just guessing.

 

[john61ct]

You can get them tapped for bolts, use crimped ring terminations for wiring.