Building a pack using salvage 18650 cells

[Hep!]

Hi,
First of all, I'd like to say - please hear me out. I've done some searching and have found that while some members of this forum have successfully completed such a task, it seems by and large, the overall userbase of this forum suggests against the use of li-ion 18650 cells in favor of various other chemistries, primarily LiFePO4. So I am aware that many suggest against it. That said, it looks like commercial products nearly identical to my planned pack already exist: http://www.campsolution.com/li-ionbatterymodule36v88ah288whwithpcbready40x1865010s4p.aspx

However, this is a project that I've been planning for a bit of time, and I only recently found this forum. Too bad for me.

I'm soon to be living in NYC and plan to use an E-bike as my primary mode of transport. I wanted to do this as cheaply as possible.
After looking at dozens of pre-built kits and reading reviews as well as info about different battery chemistries (it seems most commerical kits are SLA or Li-Ion), I decided that I finally had a use for the giant box of mostly working laptop batteries at work.
I traded my boss 2x2GB sticks of DDR3 memory for almost 50 li-ion laptop batteries. After looking at the ratings of the batteries and sorting them based on the cell size I expected to find inside, I found that 2200mAH cells appeared most frequently. I am waiting on my testing equipment, but I have 88 untested 2200mAH cells (more if I keep cracking open batteries). I am hoping that at least half of these will be good. Being a little hopeful, as I would like to build a pack that is 5p10s to achieve 11AH at 36-37v.

My plan was this:
1) Charge cells individually on this: http://www.batteryspace.com/smartchargerformulti37vrechargeableli-ioncellsof18650rcr123aandmorech-unli004.aspx
Cells which do not charge immediately go to the recycling bucket.
2) Once I have my cells which charge sorted out, I would like to test the capacity of each cell. I do not currently have the knowledge to test capacity, but I understand how AH are calculated and I taught myself the rest of this... so when I get there? If anyone has suggestions on this, please tell me. Thanks!
3) Wire cells in 5p groupings, using cells of same manufacturer, and if possible, cells from the same original battery. I know soldering these batteries directly is a bad idea and thermal glue gets bad reviews, so I'm still brainstorming here. Many of the cells still have tabbing attached, some do not.
4) Wire 5p groupings into 10s layout with this (or similar) PCM: http://cgi.ebay.com/PCB-PCM-36V-37V-Li-Ion-Battery-w-Equilibrium-30A-/300473778895?pt=LH_DefaultDomain_0&hash=item45f5a202cf

That's all for now. Does this make sense? I understand that li-ion is dangerous and that is why I am seeking the advice of professionals. I am merely ignorant and trying to learn, so please go easy on me.
Thank you to everyone who read and extra thanks to those who respond. Love.

 

[grifftek]

I have built a pack from these 18650 cells taken from new old stock laptop batteries. You can read the post here:
http://endless-sphere.com/forums/viewtopic.php?f=14&t=25890

We only have a few rides on the pack so far but it seems to work just fine. feel free to check out the bikes at
http://www.grifftek.com/grifftek/Other_Fun_Projects_and_Hobbies/A2B_Electric_Bike/

The batteries i used were already spot welded in 3p1s arrangement so I paralleled 4 groups of 3 in a 10s arrangement.
regards

Bill

 

[Hep!]

Dude! That's quite a nice box you fabbed up there!
I won't be going so high tech sadly... planned on just using PVC sheets cut to form a box and epoxied together.
Did you ever get the issue with your PCM resolved? If not, what are you doing now for balancing the pack during charging?

 

[dogman dan]

Any who scoff at your plan merely know there is better stuff out there, and getting better all the time. If you have money, for instance, lipo from Hobby king.

So we won't be recomending your plan if you have to pay much for the cells. Building a pack from freebie stuff is a different story. You do want to use only the good cells in the permanent pack. Lipo that has been overdischarged is dangerous lipo.

 

[stanz]

We only have a few rides on the pack so far but it seems to work just fine. feel free to check out the bikes at
http://www.grifftek.com/grifftek/Other_Fun_Projects_and_Hobbies/A2B_Electric_Bike/
Bill

NICE STUFF! I especially like the enclosure for the two cell spies

 

[grifftek]

Dude! That's quite a nice box you fabbed up there!
I won't be going so high tech sadly... planned on just using PVC sheets cut to form a box and epoxied together.
Did you ever get the issue with your PCM resolved? If not, what are you doing now for balancing the pack during charging?

thanks for the nice comments!

Did you look at the other enclosures I drew up in solidworks? they use standard aluminum tubing.

http://www.grifftek.com/grifftek/Other_Fun_Projects_and_Hobbies/a2b_9p10s_box/

http://www.grifftek.com/grifftek/Other_Fun_Projects_and_Hobbies/a2b_5p20s_2x8_enclosure/

if you look around you might even find commercial end caps to cap poff the ends of the tubes, if not they are not that hard to make. you could just use a small enclosure to hold the bms instead of the one that I designed.

I never ironed out the bms problem witht he tenergy board and hey were not interested in helping to solve the issue either so I replaced it with a BMS from hightek bikes. Its the signalab bms and it is working flawlessly. I need to update my page to show the new bms!

regards

Bill

 

[Hep!]

Fine work Bill, I'm very impressed with how OEM your packs look.
The aluminum tubing is a great suggestion! I originally had wanted to use something like that, but didn't know what to search for. I had found some extruded aluminum enclosures which I was and still am considering.
Perhaps the correct solution will be to make my own from tubing!
Thanks for documenting your whole process with so much detail and providing all those drawings, and thanks for the advice.

 

[JRitt]

I built a pack from 2 yr old lipos that were in Defib battery packs (never discharged). The pack is 10s 7p (37v~14ah) because the cells were only rated for 2c discharge. The bike has a cheap 20" 500w ebay kit and it works great. I have no idea how far it will go but he rides it daily in our very hilly (up to 10% grades) and I only charge it every 2 weeks when the meter says <20% while climbing hills.

 

[mihai_atanasiu]

I found on Ebay a lot of 18650 Li-ion cells with different capacities between 1000 mAh to 3000 mAh;

especially http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=280498128322&ssPageName=ADME:X:AAQ:US:1123

18650 Li-Ion 3.7 V 2400 mAh costs 1$ each;

I asked the seller about the discharge rate of the cell but he avoid to answer;

My plan is to build a battery pack made from 100 pieces: 10P10S, 37V 24 AH;

One cell weighs about 44g; that is 5 kg for the pack;

Have any of you more data ?

 

[Harold in CR]

Guys on the ES, pay $2.00 and up for USED cells. Those you posted about, are recommended for camera use, which is low discharge rate.

You need much better C rate. Check the for sale used classifieds on ES.

If you can afford it, Hobby King lipo packs, like Dogman stated, are the way to go.

Just learn the charge-discharge technology, so you don't have problems.

 

[texaspyro]

I found on Ebay a lot of 18650 Li-ion cells with different capacities between 1000 mAh to 3000 mAh;

Beware with more than a little suspicion of ANY 18650 advertised capacities!!!. Wildly bogus spec are the norm. And there are more counterfeit cells than real ones... The label means absolutely nothing!

 

[liveforphysics]

I found on Ebay a lot of 18650 Li-ion cells with different capacities between 1000 mAh to 3000 mAh;

especially http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=280498128322&ssPageName=ADME:X:AAQ:US:1123

18650 Li-Ion 3.7 V 2400 mAh costs 1$ each;

I asked the seller about the discharge rate of the cell but he avoid to answer;

My plan is to build a battery pack made from 100 pieces: 10P10S, 37V 24 AH;

One cell weighs about 44g; that is 5 kg for the pack;

Have any of you more data ?

LOL! Gotta love the guys that decide what to print the shrink wrap.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=280498128322&ssPageName=ADME:X:AAQ:US:1123#ht_2698wt_1139

IMR means it's a Lithium Manganese chemistry. The current world record for an IMR 18650 is aproximately 1.85Ah, and it's a lab proto.

 

[Hep!]

So... onwards with this project...
I've done more research, searching, looking at pictures, etc.
Been testing my cells. Ordered some more charging equipment since one charger isn't enough!
Feeling like I probably don't need to actually test the capacities of the cells. These aren't cheap-o knockoff cells, these are cells out of Lenovo, Sony, Dell, HP, and Toshiba OEM batteries (NOT aftermarkets).
I think as long as the cells charge properly I should be okay to use them. If this is incorrect, please inform me.

After doing lots of research, I do feel that soldering directly to cells is possible to do without harming the cells or yourself. However, considering the size of the pack I am building, destroying even one cell with heat could have very negative consequences on the end result, possibly even dangerous consequences. It's not worth the risk (or the amount of work). [http://linux-7110.sourceforge.net/howtos/netbook_new/LIbattery/LIbattery.html]
Instead, I'm going to be attaching to the cells using magnets. These magnets to be specific:
http://www.dealextreme.com/p/super-strong-rare-earth-re-magnets-8mm-100-pack-suitable-for-extending-18650-cr123a-batteries-13518
I plan on making 9 arrays of 12 magnets and 2 arrays of 6 magnets. I will be re-purposing wires from an Antec computer power supply to make the honeycomb formation. The wires are stranded steel, looks like... 18 gauge.

Here is the projected pack layout: [not to scale]

(drawn in PowerPoint because I don't have another drawing program and I'm a glutton for punishment)
Wires for my PCM/BMS will be attaching anywhere to each magnet array.

Once the pack is assembled and tested, wires will be run so they all come off one side of the pack, and the entire pack will be put inside heatshrink tubing. This will keep cells in place and prevent the magnets for shifting (which would be baaaaaaaad).

The last thing I want to do is house it in an enclosure. The pack should be about 70mm x 65mm x 325mm.
Determined these dimensions by...
Cell is ~65mm long, give some space for magnets and wires, 5mm should be sufficient
Cell is ~18mm diameter, 18*3+(.5*18)=63mm, again some wiggle room added
By solving for the height of 3 cells stacked in a pyramid formation using h=2r+sqrt((2r*2r)-(r*r))... h=18+sqrt((18*18)-(9*9))... h=18+sqrt((324)-(81))... h=18+sqrt(243)... h=18+15.5884572681199... h=33.5884572681199. Since this number is only for 3 cells (or 1 parallel pack) I need to make some changes to the work done. It's actually h=18+19(15.5884572681199) for the total size of the whole array of cells.
Which gives me 314.1806880942781mm. So my final wiggle room is added to allow space for the PCM (though... this might not be enough space!).

EDIT: Just realized my math is wrong for the pack length. Going to re-calculate that real quick, and who knows... it might even fit into the enclosure I linked. *Crosses fingers*
Should be... 18+(18*13.1769145362398)+15.5884572681199=270.7729189204363... BOOM YEAHHH! Looks like it actually will fit in the enclosure I linked.
Oh, and if this math is wrong... it's been a long time since calculus. End edit.

I was really really hoping that I would be able to find an extruded aluminum enclosure for this all, like this one here: http://www.batteryspace.com/aluminiumenclosure125x68x296mm.aspx
Dimensions sadly seem to be lacking. That leaves me with the choices of trying to design a new pack (would prefer not to, I quite like these projected dimensions), or fabbing my own box (would also prefer not to, but I will if I have to)!
Anyone know where I can find a box to better fit my pack? Or is there a cell arrangement that would fit in this box nicely that I'm blindly missing? (very possible)

When building my pack, do I want to use anything like this to limit discharge rate? If so, what should I limit the discharge rate to?
http://www.batteryspace.com/PolySwitch-730---7.3-Amp-limit-20-C.aspx

So experts, how does this look? Any thoughts, suggestions, or criticisms? (Especially if I've done... er... planned to do something dangerous).
Thanks for the all the contributions to this topic thus far.

 

[Hillhater]

I'm going to be attaching to the cells using magnets. These magnets to be specific:
http://www.dealextreme.com/p/super-stro ... ries-13518
I plan on making 9 arrays of 12 magnets and 2 arrays of 6 magnets. I will be re-purposing wires from an Antec computer power supply to make the honeycomb formation. The wires are stranded steel, looks like... 18 gauge.

i would do some more research on the likely resistance (power loss) of using all those "dry" magnetic connections, and steel conductor wires.
I dont have any data, but i just suspect that is not a good set up for a low loss , high current battery pack !

 

[Hep!]

Thanks for that feedback. I figured the wire should be okay considering the layout of the wire I planned, the wire will be soldered to the magnets. Doesn't mean I am right though.
The point though about the contact from the magnet to the cell, I cannot speak for. I will have to test the resistance of several magnets, I suppose. Dang, I was hoping that I had this all solved.
More feedback please!

 

[amberwolf]

I expect that when you heat the magnets enough for solder to stick (if you even can) they're not going to be magnets anymore.


Magnets do work great for sticking the wires to a cell while testing, or even single-cell charging stuff wired into a pack, but I wouldn't recommend soldering wires to magnets and then sticking the magnets to the cells.

However, if you used nickel strips across the cells, with strong enough magnets on top of that, to force the strips against the cell caps, then pressure-wrapped the entire setup with something non-conductive, so that cells can't shift around, it might well work.

One problem with this dry connection method might be that as it gets any moisture in there, over time it will probably corrode the cell-to-strip faces of each, causing higher resistance whereever this occurs.

 

[Hep!]

I like your style man, I can see you're a lot like me!
Nothing is junk... it's just a component of something good...

Sounds like my magnet idea is going to be bunk. Too bad I already ordered 200 of them, d'oh. :lol:
So, what would your suggestion to me be?
I can build one of those CD welders and put on my own tabs. But I seriously don't wanna shell out $100+ for parts, then build a tool, just so I can build another part of another project.
I could solder directly to the cells. I did want to avoid this.
I am pretty quick with soldering, I think using low temp solder and appropriate flux would be okay (which is the same thing I planned to do with the magnets).

Is there an alternative I'm missing?

Oh, and I have removed all of the tabbing that was on the cells I have. Some of it was solid, some of it pretty much fell off on its own. Maybe I shouldn't have done this, but it's done now.



EDIT: So, after giving this some more thought, I have an idea.
I will take four pieces of 1/4" PVC sheeting.
I will sandwich two together, then drill 60 holes in both sheets in the layout I want my cells to be. I'll leave a bit of space in between the cells.
I will then lay these on top of the other sheets, and mark where the center of the holes are, and drill out these holes.
In these holes, I will install barrel bolts. http://products.kinter.com/viewitems/binder-posts-screws-2/steel-barrel-bolts-screws
I can wire them on the top, and have the bottom be contacts. All I will have to do then is ensure pressure is maintained between the top and bottom sheets of PVC, while the two middle sheets will maintain cell spacing and alignment. This will increase the amount of work required by some as well as probably make it impossible for me to use the aluminum box I posted earlier. But... it might be worth it. Thoughts?
Should I give it up and solder? Should I give it up and find a capacitive discharge welder?
EDIT3: I've been told the above idea isn't so great if for no other reason than iron isn't the best conductor.

EDIT 2:
How about a conductive tape like this, or the conductive adhesive tabs?
http://www.tedpella.com/SEMmisc_html/SEMadhes.htm
I know glues don't work so I'm kinda in between on the tapes. I feel like they might be the right answer though!

 

[Hep!]

So sad news. I spoke to Doctorbass in PM and he advised against the use of Li-Co cells for safety reasons.
I may either be scrapping the project or re-designing the pack (if possible) to be safe(r).

So, the final word:
Doctorbass has convinced me that Li-Co does not belong in high current draw applications. I appreciate his thoroughness and patience with me greatly. The more you knowwwwww.
Anyway.
I'm now researching LiFePO4. I will use these cells for other, small projects.

 

[Hep!]

I have learned a lot in the last day about these and other cells.
I have found that the cells I have are rated for 2C discharge and 1C charge.
I was previously planning on using a 36V 500W front wheel hub motor. I think I may change my approach on this, still doing lots of reading. Depends on what my frame can handle (frame is aluminum, fork is steel, passes magnet test).
500W at 37V is 13.514 amps.
The capacity of my pack in 6p10s would have been 11Ah.
This would give me a max discharge rate of 1.229C, would it not?

By the way at this point this is theoretical, I am trying to feed myself knowledge
I think I'm going to do 2p12s with 9Ah 38120S LiFePO4 cells instead.

 

[amberwolf]

A 500W motor (like the typical 9C, such as the 2807 I have on my CrazyBike2) is just rated for 500W continous power use. That's certainly not the max it can pull.

On mine, with a 12FET controller and 48V, I see 61-62A spikes, and 30-40A for up to seconds at a time in starts from stopped.

So unless your controller is actually hard-limited to that 13.5A, with very fast response time, you'll probably see some >2C or even >3C bursts pulled from the pack. This may not be harmful, but doing it a lot for long bursts will likely reduce the lifespan of the cells.

If you have or make or get a BMS for the pack that hard-limits the current with fast response, it'll just turn off your motor power entirely, probalby until you cycle power on it. That's not helpful.

Most likely, if you're riding on the flats at <20MPH, and/or assisting with pedals, you can keep the average current draw down below 5A. That'd be <0.5C for an 11Ah pack. It's just the startups, acceleration, and hills that would draw way more than that, at least for short periods.


If you do go ahead with the pack, I'd try working out one of the no-solder methods. There are a few, in a thread about battery packs here:
http://endless-sphere.com/forums/viewtopic.php?f=14&t=17912&start=0
that give the basic idea. I had planned to do something like this, with teh LiCo cells I've got here, but have yet to get it done.


As for safety, well, LiCo in cylindrical laptop-type cells probably are the most dangerous of all the Lithium chemistries, if abused. As long as you never let them run down too far (2V for sure, 3V to be safe), and you never overcharge them, and never pull too much current from them or charge them too fast, they're perfectly safe, as evidenced by the bajillion portable devices using them with relatively few spectactular failures. :lol:

But if you arent' careful with them, then yeah, they can blow up, burn, etc. So can the LiPo RC packs, or most any other Lithium chemistry, though it takes much more abuse for most of them to do that. There are a few that I doubt would ever blow up or burn, like A123, but they also weigh more for the power they contain. LiCo is the lightest for the power they have, AFAIK. But more volatile under the wrong conditions, which are not that hard to create, unfortunately.


If you go with soldering to the cells, it's easy enough with a really hot iron and a really big broad tip. I have a Weller that is like 80-90W, and a tip the size of my pinky with a broad flat end, and it's great for quickly heating, soldering, and letting go, like one second ro maybe two on some things. I fixed a bunch of issues in my cylindrical Vpower pack with it that way, soldering down the nickel strips where welds didn't look very good, and where I had removed strips during a cell-removal repair.

BTW, if you get more cells, leave the tabbing on, as you can then just solder to that, and it's lots easier. (and safer)

Conductive tape probably wont' carry the currents you need thru the adhesive, but I've never tried it for anything other than PCB repair (where I often soldered it at the ends where it overlapped the repaired traces).

 

[Hep!]

Thank you very much for the reply amberwolf.
Almost immediately after posting I realized that it was foolish of me to assume a 500W motor could not draw higher spikes of current in short periods of time.
The other thing I realized is, that is the current draw on a fully juiced pack. As the pack drains, the current draw is going to increase as the voltage drops. The C rate will then vary too.
While I would like to use the cells in my possession, I think Doctorbass is probably right. Unless I know exactly what I'm doing, better to stick to a safer chemistry. I might tackle using these cells for something else later on. Not to mention how much harder I made things for myself by removing the partial tabs that were on the cells. By the way, I've decided that if I were to use the cells, I'd find someone with a CD welder I could use to weld tabs on. I wouldn't try to solder this big of a pack... the only reason I give it ANY thought is because I have all the tools accessible... not really because I think it's the right way... I would like to do it the right way...
The headway cells are looking very appealing. I was originally thinking a 2p12s pack would be good, but now I'm thinking I will be fine with 1p12s. It also appears a 36.3v pack of A123s would not cost terribly too much either. I'm still researching what is right for me.
As for that pack building topic you linked, thank you for that. I have that bookmarked already as a reference, but I did not think it applied to me. It looks like the cells being used in that pack build have screw holes in them? So I don't think the technique applies to me, unless I missed something.
Fortunately all of my best cells (2.4Ah and 2.6Ah) still have their tabbing intact. I did not have enough of these cells to build an e-bike battery, but I do still have about 50 between the two capacities! I read a topic by doctorbass showing how I SHOULD have been utilizing the tabbing that was on my cells, so I'll be paying attention to that when I build smaller packs with these cells for USB boosters and LED lights.
Thanks,
Hep

 

[amberwolf]

No, the cells pictured at the beginning at least are A123 cells (28650?), which are basically structurally like the ones you ahve but fatter. No screwholes.

HEadways and some others do have screwholes, but they are much larger cells.