DIY LiFePO4 48v pack - What am I missing?

Batteries, Chargers, and Battery Management Systems.
Post Reply
Lynx   100 µW

100 µW
Posts: 7
Joined: Apr 12 2018 8:27pm

DIY LiFePO4 48v pack - What am I missing?

Post by Lynx » Apr 12 2018 9:11pm

At first glance building a LiFePo4 battery pack seems simple. So I am almost certainly missing something critical. I swear I have done due diligence, googling for hours, then searching this forum, but I haven't found the answer I was looking for. Will somebody please tell me what I am missing in the following theory?

Get a good lot of LiFePO4 batteries:
https://www.ebay.com/itm/NEW-Box-of-80- ... SwUYNaAcNp

Put them together in five parallel rows of 15-series cells to get 48v, using copper tape for the connections and magnets to hold the copper on the contacts:
https://www.amazon.com/Tapes-Master-10f ... opper+foil

Add BMS of course:
https://www.ebay.com/itm/48V-45A-LiFePo ... SwIFtaRw8E

Wrap the whole thing in electrical tape, add foam padding where necessary, then shrink wrap the dude.

If it is that simple, why doesn't everybody do it? It would be far cheaper than getting an already-made battery pack, and it seems easy enough. But I must be missing something, or everybody would be doing it and nobody would be selling such extremely expensive battery packs. So what am I missing?

I know, I know, there are multiple threads where this is already explained and somebody will probably fuss at me for posting this thread. I can't help it. I'm tired of searching. Yell at me, just please point me to the thread I need to be reading. At least I read enough threads to know better than to ask "why does a LiFePO4 battery pack even need a BMS?" before I posted this thread. :-/

DAND214   1 GW

1 GW
Posts: 3088
Joined: Aug 10 2008 8:28pm
Location: About 1200 miles south of Chicago, in sunny/rainny Port Richey, Florida!

Re: DIY LiFePO4 48v pack - What am I missing?

Post by DAND214 » Apr 12 2018 11:13pm

Your gonna need 16s not 15. I have read that those cells aren't that great but I have no experience. Do you think the magnets will hold with all the bumps?

Dan

User avatar
amberwolf   100 GW

100 GW
Posts: 25475
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: DIY LiFePO4 48v pack - What am I missing?

Post by amberwolf » Apr 13 2018 4:36am

First, is it copper tape, with adhesive? Or is it copper strip, as bare metal both sides?

The former will have high electrical resistance, compared to the latter.



The problem with magnetically-held copper tape is first that each connection will likely have different resistance, so each cell will contribute a little less or a little more than its neighbors. That means things get unbalanced over time. How much time? Depends on the differences in contact resistance.


If the contact resistance is high, for whatever reason, then current thru the connections is limited, and heating occurs, wasting power.

A second problem can be that if you dont use insulating rings on the button end of the cell, the tape or strip can touch the can *and* the button, effectively shorting across the entire row of parallel cells.


Its been done in various ways (solderless packs, no-weld packs), but theyll generally all have higher contact resistance than something that directly connects the cells to their conductors.


Any form of humidity or dampness that gets in there will also cause corrosion, or at least oxidation, which will greatly increase the contact resistances.

I dont have a link for it, but there is a thread (or threads) about how NOT to build a battery pack, and it has a lot of good points and info, pros and cons of doing things in various ways. I wish I could remember the thread title, but its in the Battery tech section.


The cells themselves appear to be fullriver brand, and there are threads about them, dunno how good they are as I havent followed them. but you will want to check and see what their real c-rate is (vs what ebay ads say, since its safer to consider vendors often lie, or at least exaggerate), so you will know how many parallel cells it will really take to sustain the current your controller/motor will need continuously, and what it will really take to handle the peaks of current youll need on startup or hills or other high-load situations.


As noted in the previous reply, youll need 16s for 48v, rather than 15s, for LiFePO4.

Lynx   100 µW

100 µW
Posts: 7
Joined: Apr 12 2018 8:27pm

Re: DIY LiFePO4 48v pack - What am I missing?

Post by Lynx » Apr 13 2018 5:01am

Thank you both. And it WAS copper tape, not plain copper. I'll probably need to find something else. And I am glad you mentioned shorting from the button to the can... with that battery case, I was planning to shave a bit around the top to allow easier connection. :roll: Probably not gonna do that now.

As you can probably tell by now, I have never done this before. So why am I dabbling in it? Money, or lack thereof. I should just buy an already made battery pack.

Why 16s instead of 15? I'm not disagreeing, just confused. I thought 15x3.2v was just exactly what I needed. Or do you want to go for a real voltage of 51.2 to feed a 48v motor?

User avatar
amberwolf   100 GW

100 GW
Posts: 25475
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: DIY LiFePO4 48v pack - What am I missing?

Post by amberwolf » Apr 14 2018 1:38am

Whenever battery powered devices are rated, its a nominal voltage, not a maximum. so the cells you want to use are nominally 3.2v cells, but htey actually charge up to 3.4-3.6v, and discharge down to 2.8v. So the pack-level cutoffs for motor controllers generally work better with 16s (for LiFePO4; 13s for LiPo, and so on depending on chemistrys nominal voltage) so they dont cutoff early (LVC) and leave a bunch of usable charge in the battery when they call it empty.

15 * 3.6v = 54v ; 15 * 2.8v = 42v so thats 12v range from full to empty. if your LVC was setup for 16s thats 3.6 - 2.8 = 0.8v left on the table, 0.8 / 12 = up to 6% of your capacity. It doesnt really amount to that much, because discharge curves arent linear, but it gives you some idea.

Nothing wrong with building your own pack, but I highly recommend you read a lot of the other DIY threads and see what problems they had, and/or what was pointed out to them as potentially problematic, what they did to fix them (if anything), and how the pack is performing now, after however much use time.

I know it seems boring and tedious, but it could help you a lot preparing to build it, and making sure you dont have problems you didnt need to experience, up to and including fire. ;)

Lynx   100 µW

100 µW
Posts: 7
Joined: Apr 12 2018 8:27pm

Re: DIY LiFePO4 48v pack - What am I missing?

Post by Lynx » Apr 14 2018 7:48pm

Ah, that makes sense. And thank you again. I'll probably just break down and buy a battery pack... then my wallet will have its own breakdown.

I knew there had to be a good reason everybody wasn't making his own battery pack. I just didn't know what it was, and I didn't really know enough to know what questions to ask. I do want LiFePO4 though, first for many more charge cycles, second for much lower volatility. I know it will be a physically much larger battery pack for the power. I'll probably save up and get a 20-ish a/hr pack from BMS Battery or something. At least I know not to just hop on ebay and buy the first pack that looks good. >.>

Thank you again, for your answers and your patience with somebody who doesn't really know what he is doing.

Lynx   100 µW

100 µW
Posts: 7
Joined: Apr 12 2018 8:27pm

Re: DIY LiFePO4 48v pack - What am I missing?

Post by Lynx » Apr 14 2018 8:29pm

Actually the first thing I'm going to buy is the back wheel with a motor. I'll use a few 9v batteries chained to test it, make sure it actually works... THEN I'll buy the bike and battery.

User avatar
amberwolf   100 GW

100 GW
Posts: 25475
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: DIY LiFePO4 48v pack - What am I missing?

Post by amberwolf » Apr 15 2018 2:49pm

9v batteries are unlikely to work; they arent designed to supply the current to run the controller and all its accessories much less the motor. ;)

Keep in mind that every battery / cell model has its own performance limitations.

For instance, the generic LiFePO4 cells you might find in a typical ebike battery pack will be 1C cells, maybe 2C burst for a bit, without voltage sag that could be quite significant (reducing the actual wattage supplied and heating the cells from wasted power inside them at the higher current rates).

Other chemistries can also be like that, but theyre more likely to sustain higher currents with less sag, though they have other disadvantages like cycle life, etc.

There are better cells of all the chemistries, but generally the pack sellers have no idea whats actually in a pack, and even if they say it is a certain cell model or cell specification, they dont make the packs so they cant actually be certain of that, and its usually marketing hype they post on their pages, which you cant trust until you get the pack yourself and verify by stress testing it that it does what it said it would do. Even places that make the packs themselves may lie (or not actually know what they have) about what they build with.

Theres at least a few that have been known to build using recycled junk cells of completely unknown type, lifespan, performance, etc. (they might even mix chemistries without knowing it depending on the sources of the junk cells they use).


Theres threads linked in the Sticky Index in the Battery Tech section that give more info on battery specifications and how they work, but the C rate of a cell is how much current it can supply, rated by how many times its capacity that is. So a 10Ah pack made of 1C cells can supply 10A, with some voltage sag but probably not a lot. If you tried to pull 20A from it (2C) the cells would sag a lot more in voltage (total pack voltage will drop a lot), and that lost power ends up as heat inside the cells.

So lets say its a 58v 10Ah 1C pack made of really cheap cells (for a worst case example), fully charged. While riding under normal load on the flats at a normal speed, lets assume youre pulling 10A and voltage drops to 54v, so you are outputting around 540W from the pack. Then you reach a hill, or you hit a strong headwind, and your system (say its 1000w, with a 20A controller) starts pulling the max current it can to sustain the speed, and voltage drops to say 48v (really crappy cells) under that 2C load, so now its about 48 * 20 or 960W output from the battery pack. Since it wouldve been 1080w without the sag, then there could be as much as 120w of heat being generated inside the battery at that higher current draw, and generally heat inside batteries is bad, especially if its sustained.




All that said, a 20Ah 48v (58-ish fully charged) LiFePO4 pack will probably weigh around 30lbs, maybe more, depending on how its made and what kind of cells it uses (pouch will probably weigh less and possibly perform better, while cylindrical cells will probably weigh more and perhaps perform worse (as there seem to be more crappy cylindrical cell packs out there). Itll probably be about the size of an old style metal 50caliber ammocan (which is convenient if you need something to put it in and its the right shape to fit inside one).


As for volatility--its much more about cheapness of cells and packs built from them (quality of manufacturing for each) than it is about which chemistry it is.

999zip999   100 GW

100 GW
Posts: 8434
Joined: Jan 10 2010 1:40pm
Location: Dana Point So. Cal

Re: DIY LiFePO4 48v pack - What am I missing?

Post by 999zip999 » Apr 15 2018 3:43pm

Please tell more about your kit. A link . A battery most match your kit or as amberwolf said it will not last long. Quality is someti9mes cheaper then crap. Get something over your needs. Headroom

tanstaafl   10 W

10 W
Posts: 92
Joined: Apr 21 2017 2:42pm

Re: DIY LiFePO4 48v pack - What am I missing?

Post by tanstaafl » Apr 15 2018 7:32pm

I built a pack out of similar cells. They're not identical, so I can't speak to the quality of the ones you've linked, but the ones I got are pretty much just as described and seem to be able to handle a decently high discharge rate. I built a large battery and it's doing well.

In hindsight, however, I wish I had just bought commercially available pre-built batteries. I don't really need as much range as I've got (23Ah x 72v) and the weight is a minor but nagging issue. Whenever I move the bike with the battery on it, I have to handle it differently and more carefully than my other bicycle. Thanks to the longer life of the LiFePO4 cells, I'm sure I will be well tired of this battery long before it's worn out. Also, in spite of spending several hours working on it and much more money on little incidentals like wire, connectors, shrinkwrap, foam, etc, I'm far from satisfied with the quality of my work.

Post Reply