Suitable LiFePO4 battery to replace Li-Cobalt?

[offGridDownUnder]

Like many other machines discussed here, I am using Li-ion batteries. I presume Li-Cobalt cells. The sort than can catch fire.

My current pack is from Cap Rouge in Australia, and uses name-brand cells (Samsung or Panasonic if memory serves) - the best possible approach as I understand it for these sorts of cells.

Lithium titanate appears to have the longest lifespan, but LiFePO4 seems to also have long lifespan, and is reputed to be fire/explosion safe. My ride is based on a Greenspeed Magnum Big Wheel, and has room for a larger, heavier battery. LiFePO4 appears to be about 2x weight and size of Li-Cobalt.

I'd like to hear comments on me using a battery like this:
48v LiFePO4

My current battery is 48v, 20Ah, and feeds to a Grin Baserunner and then a Grin V3 All-Axle motor. I also have a Phaserunner available to use instead.

It seems to me that the volts, amps (30 out rated), and total capacity meet my needs - what else comes into play? Can I use something like this? I have a Grin Satiator and also a programmable solar blanket for charging.

Any concerns, other issues to consider or find out values for? Is any of my information (or beliefs) mentioned above incorrect?

 

[amberwolf]

Just make sure that if you are buying a pouch-cell pack (vs cylindricals), that it is built with proper compression of the cell stack. If it isn't, it will probably have a much shorter lifespan than it could otherwise, because the cells can have gas build up between layers and cause degradation, and eventually visible swelling and failure, and even rupture of the pouch itself and leakage of electrolyte.

Same if it's built of prismatic (box) cells.

(this is also true of any other lithium chemistry cells, but it's very common to find LiFePo4 "box" style batteries built of pouch cells, and almost none of the ones I've seen the insides of have compression at all, much less correctly done. See here

Teardown: LithiumPros LithiumPowerPack 12v and 36v

I won't get to the actual teardown for at least a few days, but Cvin dropped off these failed units for experimentation. The 12v is more swollen than the 36v, but the 36v is actually broken across the bottom of the case (probably from an impact rather than the swelling, can't tell till I get...

endless-sphere.com

for a couple of examples; there's many more around the forum and elsewhere).


When looking at any particular pack, if you can't get the manufacturer / seller to provide images of the internal construction, it may be safer to pass on it. You will not find "cell compression" as a feature listed on any pack, even the ones that have it. You'll only know by looking at how it's built.

No need to tell them what you're looking for, as that may simply enable them to find images of a pack that does have that and send them to you, instead of actual images of what they are selling.

 

[jkrienert]

On the notion of LTO cells: Two of these in series would maybe work for you?

27.6V 45Ah 1242Wh LTO Toshiba SCiB 20,000 Cycles – 24x 23Ah Cells

These are excellent value genuine Toshiba SCiB LTO modules testing near full capacity. Each pack contains 24x 2.3V 23Ah cells in a 12S2P configuration (45Ah nominal). LTO chemistry can be discharged very low and performs extremely well in car audio and cold-temperature applications. The cells...

batteryhookup.com

Those would provide you 48v 40Ah up to 6C (>10Kw). A BMS would be needed, and mounting for a low CG since they'd sum to around 60 Lb.

Its a phenomenal price point on LTO prismatic. Although they are used, there is probably a lot more service life in them.

 

[offGridDownUnder]

Two of these in series would maybe work for you?

I've seen these, and I'm interested, but I don't know enough.

As you write, I'll need a BMS. I know what they are, but how do I find one that does the job with these batteries? What are the relevant characteristics to match?

I'd like to learn, but how do I learn about what is required for this application?

 

[offGridDownUnder]

that it is built with proper compression of the cell stack

Hi Amberwolf. This is most of your reply - does this mean that the other matters are probably ok?

This is substantial money, and I'm a bit in the hinterlands here. I want to be certain before I commit to fitting, buying, etc. So, I'm checking.

Thanks.

 

[amberwolf]

Hi Amberwolf. This is most of your reply - does this mean that the other matters are probably ok?

Probably.

LiFePO4 is less likely to burn for the same type of faults as LiCo, but it's still a chemical energy storage, and it *can* still burn.

Quality of cells *and* pack construction make at least as much difference as whcih chemistry as to the amount of hazard a pack creates.

The hard part is finding a pack that can do what's needed that also has good cells and construction, *and* doesn't cost a fortune.

Cheap packs are almost certainly going to lack at least one, and possibly, of the qualifications.

My personal option would be to find used EV modules to do the job, that are individually built with the necessary compression if they are not cylindrical cells, and adapt them as they are already built to do what you want. For instance, if you can't find an exact voltage match, get one of higher voltage and wire the main + connection to a cell group that gives you the voltage (relative to the main - connection on the module) that you need. That leaves you with spare cells at that end that aren't used, if you have to do a pack repair later, and for now you don't have to modify any of the module wiring or construction beyond that connection.

The used EV modules are very likely to be built of well-matched cells, so they all perform the same, and are also very likely to be true top-tier cells, so that even in used condition they are better than those found in most typical non-EV packs out there. The construction is also very likely to be much better. All of these are because warranties are expensive to honor, especially for batteries in a large EV, and not having to do so is better, even if they spend a little more upfront to do so.

Places like Greentecauto have good deals on various modules; you can also usually find a fair amount of info about them out there because plenty of people repurpose them for all sorts of things...while there is almost no data about any particular non-EV pack available anywhere except in posts around the web where people are trying to fix a broken one. :/



Additionally, in general, but especially for a longterm usage pack, I would always recommend buying a battery with a large margin of extra capacity (Ah), and a large margin of current-delivery capability (A), so that as it ages and loses capability, it still performs as required. That also happens to put less stress on it so it ages more slowly from usage, and incidentally gives you emergency range extension before it ages enough to lose that extra capacity.

 

[amberwolf]

As you write, I'll need a BMS. I know what they are, but how do I find one that does the job with these batteries? What are the relevant characteristics to match?

You need one that is either user-programmable for any chemistry you might use, or one that is factory-made for the chemistry you know you are using.

It must be able to handle *at least* as much current (A) in discharge as you will ever use, worst case.

It must be able to handle *at least* as much current (A) in charge as you will ever charge at, for the fastest charge you will ever need. (the cells ahve to handle that too).

You must decide if you want a FET based BMS or a contactor based on. FETs are smaller and simpler, already on the board, but have common failure modes that sliently prevent the BMS from protecting the pack. Contactors are larger and may require you to wire them up, or may be prewired as part of the BMS box, but they don't usually fail the same way as FETs (they can).

If you go with FET-based, you must decide if you want a common-port (charge and discharge thru same wire) or separate-port (separate wires). Each has advantages and disadvantages, if you look around for my posts on BMS you'll find breakdowns of these and other things.

There are other options, such as bluetooth, that allow you to use an app to monitor various things, sometimes evne control or program the BMS remotely...but these are dependent on phone/tablet OS and such, so unless you have a device with updating disabled, so app and OS always remain the same, *and* the app does not have a timeout, you can end up with a system that at some point simply doesn't work if it is dependent on the app to work. (because the OS might update to something the app doesn't work on, and the app might not get new updates, or the app might update and require a newere OS feature, and the device not work with any newer OS. Some apps may also be server-dependent, so that they will only work with internet access adn only if the ocmpany keeps that server online, etc.. Etc.


I got distracted by JellyBeanThePerfectlyNormalSchmoo needing attenshunals, so I probably forgot some things in there.