18650 battery holder

[leamcorp]

Hi, I've started to work on creating a battery holder for 18650 and A123 cells. I'm fairly good with molding/prototyping, but not that good with battery/electronic, I would like to ask lots of question about how to arrange the holder for maximum modularity and efficiency (in terms of space).

Reason for the holder?

Cost - 18650 cells are darn cheap. I've talked to some Chinese vendors for 120-200 Ultrafire/Trustfire cells (enough for 48v 20ah) and quoted prices were very reasonable. The price I got was anywhere from $1.70 to $3.20 each. At mid point (for 3.7v 2500mh), it would cost around $330 for 48v 20ah.

Modularity - I can't solder battery and I don't want to pay for soldering service. I like the idea of just popping in battery into a case to make a pack. I understand that there will be some soldering for wiring but I would imagine anyone could do that.

Repair/troubleshoot - The pack can be buildup to whatever volt or capacity and at the same time can easily be broken up for troubleshoting and easily replace bad cell(s).

Negative?

The size will increase as holder will add additional bulk.

 

[leamcorp]

Here's are some items I've got so far from Mouser.com

Most likely I will be using 209 and 228 to make the holder. If you know of any other, let me know. These you just pop it into the casing and it stays there. I thought 211 will works good too but its a lot of work to screw down each one of those (like 120 or 160 of them).

 

[leamcorp]

Here's my first set of questions,

1) what gauge wire can I use to interlink 18650 cells (they will be soldered on back of those battery contact). I would need to calculate amount of space needed at the end so that these wires could be tucked away.

2) What would be a ideal number of batteries for each holder. It would be great if we could move it in 6v or 12v increment. The ones I've been looking at was 3.7v 2500mh

 

[dogman dan]

My impression is that each case should be a paralell group. that way each case wires to the bms with one wire. So the size of the case would depend on the ah of the pack. Each battery holder would then be a "cell". A problem cell could then be easily removed and and a replacement put in, allowing you to fix the broken cell later. For the 2500 mh, maybe 4 per group for 10 ah pack?

 

[pwbset]

My impression is that each case should be a paralell group.

Yes! Absolutely make them parallel enough for ebike use and for BMS simplicity etc. etc. Even 4p is too small in my opinion because of realistic ebike loads and I would highly doubt those tenergy/ultrafires etc could handle higher C rates for long term... if you design a really easy modular/snap together structure then you can parallel as many as you want I suppose. Is there a realistic design you can think of that would allow endless 1s1p snap together possibilities? I know for the konions I'm working on 1s10p is going to be the minimum realistic parallel that I'd be looking for personally. Please let us know how we can help!! This is a very exciting project you're doing! 18650s really need an option like this... looong overdue.

 

[pwbset]

Pretty different direction really but maybe these will give you some layout ideas for your molds... like brick vs. circular etc. etc. I plan to re-use the Makita cell dividers/holders and fabricate some sort of compressive tops/bottoms that can tighten with two screws or whatever and have subpacks in 1s10p and make 14 of those wired in series in a triangle enclosure. Anyway... around 15ah at 52v nominal and less than 20lbs is the goal. 1s10p subpacks would seem an ideal size for ebikes... 3 for 12v, 6 for 24v, 10 for 36v, 12 for 48v etc. etc. depending on the chemistry... LFP would be a little lower obviously.

View attachment dscn5131.jpg

View attachment dscn5133.jpg

 

[leamcorp]

My impression is that each case should be a paralell group.

Yes! Absolutely make them parallel enough for ebike use and for BMS simplicity etc. etc. Even 4p is too small in my opinion because of realistic ebike loads and I would highly doubt those tenergy/ultrafires etc could handle higher C rates for long term... if you design a really easy modular/snap together structure then you can parallel as many as you want I suppose. Is there a realistic design you can think of that would allow endless 1s1p snap together possibilities? I know for the konions I'm working on 1s10p is going to be the minimum realistic parallel that I'd be looking for personally. Please let us know how we can help!! This is a very exciting project you're doing! 18650s really need an option like this... looong overdue.

This is where my limited knowledge is showing.

The battery I'm looking at is 3.7v and 2500mah (I've also seen 1200 and 1400mah but I like to keep the cell # lower). So this means that to make 10ah, I would need 4 cells, correct? To make 20ah, 8 cells, etc.

Then make the serial connection - to make 36v, I would need 9 cell? and 48v would be 13 cell? What is the number to arrive at 48v designation (knowing that my Mr.Lau battery charges up to 55v and typically run at 52v). So confusing. Making the holder to hold 4 at a time seems not so efficient for ebike setup as bulk will add up. i would think 10-12 cell holder will be better and wiring can be done in variety of ways.

As for not have enough C, its more than enough for a typical bike setup and at very reasonable price. I'm typically doing 30mph with WE, 24 with Bafang and hopefully 35mph with new BMC 1000w motor. All this with very crappy and fixed several times Mr.Lau's 48v.

But I'm also making A123 holder too.

 

[leamcorp]

Pretty different direction really but maybe these will give you some layout ideas for your molds... like brick vs. circular etc. etc. I plan to re-use the Makita cell dividers/holders and fabricate some sort of compressive tops/bottoms that can tighten with two screws or whatever and have subpacks in 1s10p and make 14 of those wired in series in a triangle enclosure. Anyway... around 15ah at 52v nominal and less than 20lbs is the goal. 1s10p subpacks would seem an ideal size for ebikes... 3 for 12v, 6 for 24v, 10 for 36v, 12 for 48v etc. etc. depending on the chemistry... LFP would be a little lower obviously.

Same idea but yes different direction. I want a easier to handle holder - so that you don't have to glue, hold it with a rubber band, etc. I just want to pop in a cell once holder is all setup.

By the way, where did you get the holder thing there. I could use something like that to make the mold also.

 

[GGoodrum]

If you could do a snap-together block, like the ones shown below, but make it for a123 cells, I think you'd sell a crap-load. I know I'd personally buy hundreds of them. This one below is for the large format 40138 PSI LiFePO4 cells, which already have screw terminals on the ends of the cells, What you would need for an a123 or 18650 version is either one of those 211 springs, with a screw that holds it to the block, or with the other tab-type ones if you can leave room to solder to the tabs.

In any case, this would allow a ton of flexibility in how packs are put together, and they can be wired in either parallel and/or series configurations.

-- Gary

 

[leamcorp]

If you could do a snap-together block, like the ones shown below, but make it for a123 cells, I think you'd sell a crap-load. I know I'd personally buy hundreds of them. This one below is for the large format 40138 PSI LiFePO4 cells, which already have screw terminals on the ends of the cells, What you would need for an a123 or 18650 version is either one of those 211 springs, with a screw that holds it to the block, or with the other tab-type ones if you can leave room to solder to the tabs.

Where can I get one of those block? It gave me a good idea to incorporate the fastening mechanism.

I don't know about going commercial though, I'm making some for myself and some if anyone here is interested. It will be a low volume - probably spin casted parts using fire proof urethane plastic vs injection molding.

 

[pwbset]

By the way, where did you get the holder thing there. I could use something like that to make the mold also.

The white plastic cell spacer? That's what's inside the 18v3ah Makita reject packs that give us all these yummy re-useable konion cells for so cheap!

If you could fabricate some sort of subpack "box" with 209/211 maybe that went around 10 18650 cells like that you could have very diverse pack building options... of course the singular "lego blocks" would be even more ideal.

 

[GGoodrum]

Where can I get one of those block? It gave me a good idea to incorporate the fastening mechanism.

I don't know about going commercial though, I'm making some for myself and some if anyone here is interested. It will be a low volume - probably spin casted parts using fire proof urethane plastic vs injection molding.

AndyH sells them here: RechargableLithiumPower.

I'd still buy as many as you'd make. Finding the perfect solution for easy a123 pack building has been somewhat of a "holy grail" for me. Here is my latest attempt, and while much better than anything I've done to-date, it is still pretty labor-intensive:

It does allow me to do a very robust pack, without having to solder or spotweld anything to the cells, but it takes a coon's age to assemble all those copper strips with the 211 battery springs. Once those are finished, however, the actual pack goes together pretty quick. I don't have pics of the completed pack yet, but it is a 16s5p 48V/11.5Ah configuration, with a 16-channel LVC board on the front.

-- Gary

 

[leamcorp]

GGoodrum, I'm sure we could do some horse trade.

Anyway, I've just did a quick prototyping to do a proof of concept. Please understand that I spent less than 1/2 hour making this - so, its horrible looking prototype. The final model will need to be super thin, much shorter, etc.

I'm envisioning that there will be two different pieces - An end piece that will be used to fasten the entire pack and middle pieces for actual cells. I still think clip on tabs will be easier to work with but we'll see how that works. I do have a source that could make cooper strips with correct holes.

 

[pwbset]

I'll take 280 of them.

:wink:

 

[GGoodrum]

Looks good. Can you do one for the a123 cell as well?

 

[Gregory]

I've been thinking the plastic lego blocks would be great for the Headway cells. Do one for them!

 

[dogman dan]

Damn you guys are good! I like the combination of Goodrums springs and PSI's blocks. Being a carpenter, I can see how you could make multicell blocks out of a tight grained wood like poplar pretty fast with a table saw and a drill press. Recycled plastic deck lumber would work too, or polycarbonate stock, whatever. Still some work to put all the springs in, but this idea rocks! To fit in the triangles or whatever, 4 cell, or 8 cell, or whatever blocks could be made in all the variety of shapes by cutting up a larger piece into the shapes you need, squares, rectanges, l shape, etc. See where I'm heading? The thing could be made in one big piece, and then the customer cuts it up to the funny shapes he needs, or not, his choice. All you need to make is a big plank with the holes drilled, and the notches cut, and let the customer assemble the dang springs.

 

[MitchJi]

Hi,

Anyway, I've just did a quick prototyping to do a proof of concept. Please understand that I spent less than 1/2 hour making this - so, its horrible looking prototype. The final model will need to be super thin, much shorter, etc.

I'm envisioning that there will be two different pieces - An end piece that will be used to fasten the entire pack and middle pieces for actual cells. I still think clip on tabs will be easier to work with but we'll see how that works. I do have a source that could make copper strips with correct holes.

Clip on tabs might not work as well. The springs have a much wider range where good contact is made.

How do you plan to hold the end pieces the correct distance apart? You could have holes for threaded rods but if that required rods for every cell it wouldn't be a good solution. Maybe U shaped channels would work better? Cut them to length for the number of cells you want, slide in the cells and use a couple of threaded rods to pull them into the correct tension?

Holders for the end caps or end U-shaped channels that kept them at the correct distance would be slick but more difficult to design and build and it would be harder to make them as flexible in terms of the number of cells.

 

[leamcorp]

GGoodrum, thanks for sending me some spent cells. Will use them to mold the piece for 123block.

dogman, you should make one with the wood. Might be nice.

Mitchji, I'll be playing with various method to attach the spring or tabs - we'll see which one works better. Until its tested on physical materials, hard to predict what will work best for a particular design. The first prototype is just that, a first. It could look completely different by time its done, just nature of the game. But this is why its good to brainstorm, the original version I was thinking of went in different direction (like typical AA holders).

 

[leamcorp]

Question, will 14 or 16 gauge copper wiring work for individual A123 cell?

 

[leamcorp]

Tab version. Turns out that tabs have as much, if not more, force/coverage as the spring. The good news is that its easy to put the tab clips. The bad news is that you'll either have to use a copper wiring or do soldering on the back. Also, it maybe difficult to make since the back section will have to be fairly thin, which make it difficult for all but injection molding.

Ive just ordered a spring for C/D size (211-D) and cell blocks. Hopefully those will be little better.

 

[MitchJi]

Hi,

Tab version. Turns out that tabs have as much, if not more, force/coverage as the spring. The good news is that its easy to put the tab clips. The bad news is that you'll either have to use a copper wiring or do soldering on the back. Also, it maybe difficult to make since the back section will have to be fairly thin, which make it difficult for all but injection molding.

Could you run the Copper Strap under the clips on the outside of the holder? It would probably be better if there is something securing the tabs (screws?).

 

[dogman dan]

Tab idea looks real good, but might have to be too thin for wood, or even recycled plastic fake wood. But they could still be made in blocks that had a row of two cells, as long as you wanted. I was sorta thinking a sheet of something could be used to fasten two blocks together, completing a box, and setting the distance between them. Sheet metal, panel, acrylic sheet, I'll make a prototype of some kind this weekend and show ya'll what I'm thinking of. It will be miniature, to fit some AA cells.

Ok here's a very quick and dirty mock up, using c cells. Wrong size bit made holes too big, and too much space between the cells, but as close as I dared to drill the holes in a pine 2x4. With good hardwood, or plastic that can be cut with woodworking tools, the blocks should be the easy part. In this mock up, the cells don't lie straight, mostly due to error marking the holes to be drilled. For production a template would be used to get em perfect. So this would be a 6 cell paralell pack, a basic building block to connect to one bms wire, and series with the rest of the pack.View attachment 6p pack, 1.JPGView attachment 6p pack 2.JPGView attachment 1View attachment 6p pack 4 opened.JPG

I just cut two grooves in the back of the cell holder for copper strip or soldered wire to go in. In this prototype the bolts and springs would be used, but I think with other materials, it could be cut thinner and the spring clips would work. Mabye clips could just sit in the bottom of the hole, and just a soldered wire comes out the hole in the back. The distance could be set pretty permanently with some metal or plastic sides screwed on, as in the pic, or threaded rod could be adjusted just right, and the side piece used mainly for attachment and protective purposes. Anything that makes A123's easy to build a pack with would be great, clearly there will be more and more sources soon for the good cells. I have just hated the idea of a bad cell in the middle of a huge pack made of spot welds, like the Jimmy wu style.

 

[GGoodrum]

I'll try and get the "dead" a123 cells out to you today, but it will probably be tomorrow.

I've used the larger springs, which are easier to work with, but the problem is they don't compress as much, which could be a problem. I really like the small springs, but as I've see you have learned, you need to clip off part of one end in order to fit a screw through it. It makes the whole process very labor intensive. I really like the push-on clips/tabs, if you can make those work. It takes far less time to solder wires to these than working with the springs and screws. Not only will the latter be labor-intensive, but then you have to do wires and lugs, which takes even more time.

For holding both ends of a pack together, the best way, in my opinion, is to use threaded rods. Figuring out how to keep these as building blocks but having holes for threaded rods (2 for 10-12 cells is fine, I think...) might be problematic.

-- Gary

 

[GGoodrum]

The size of the wires between cells depends on the number of cells in parallel in the pack. You need to first look at the peak loads the pack needs to supply, divide that by the number of cells in parallel,which gives you how much each cell needs to supply. For example, if your setup can pull 50A and the pack you ae building will be a 4p/9.2Ah configuration, the wires need to be able to handle 12.5A. If it is a 5p/13.8Ah pack configuration, the per-cell current drops to 10A, for a 50A max load. I would think 14 or 16-gauge wire would be fine between cells, for either case. If, on the otherhand, you are only building a 2p/4.6Ah pack, each cell is going to be supplying 25A, so you really need 10-12 gauge wires between cells.

-- Gary