Tubular 18650 solderless pack

IdleMind

100 W
Joined
Dec 2, 2013
Messages
100
Location
SE Wisconsin
A while ago there were some threads discussing solderless 18650 pack construction. This caught my interest enough that I wanted to give it a try. I am currently using a pack made of 14 HK 4s bricks (8s7p) and it works great, so I don't have an immediate need for a whole new pack. So to start with I am making a 1s35p pack to use as a booster for my current 8s Lipo pack which will bring it up to 9s.

The outer case is made from 2 inch PVC pipe. The cells are stacked 5 high and there are 7 cells per layer all connected in parallel. At one end there are 7 springs to compress each string of cells. Between each group of 7 cells there is a flexible copper plate to make electrical contact. The copper plate is .008 inch thick.

The pipe's ID is a bit too small to accept the 7 cells so first I turned the OD smaller in a lathe to make it more flexible, then I slit it lengthwise so it can be spread open enough to insert the cells. Here's how it looks with a group of cells inserted in one end, you can see how the split pipe tightly holds the cells.


It's held together with 2 end caps and 3 tie rods made of 10-24 threaded rod partly covered with shrink tube. Seven springs tension the rows of cells. There are 6 contact plates, the positive plates have raised bumps to assure they do not make contact with the cell case. Each contact plate has a 4mm female bullet connector for attachment to the wire harness. Here's a photo of the parts.

Here's a photo of the assembly and the assembly after taping.


These parts could be rearranged to make a 5s7p pack, and two packs put together for 10s7p.


I'm working on getting it mounted in my bikes trunk and will also be discharge testing it.
 

Attachments

  • 001 5-25-15 Tube.jpg
    001 5-25-15 Tube.jpg
    39.2 KB · Views: 3,628
  • 002 2-26-15 parts.jpg
    002 2-26-15 parts.jpg
    43.3 KB · Views: 3,628
  • 003 2-26-15 assembly.jpg
    003 2-26-15 assembly.jpg
    38 KB · Views: 3,628
  • 004 2-26-15 taped assem.jpg
    004 2-26-15 taped assem.jpg
    36.9 KB · Views: 3,627
looks far interesting, if I get it right you will add a 1s35p lithium paralleled cells to a 8s LiPo pack.

Which is the capacity of each pack?

How will you manage voltage outputs, HVC / LVC?
 
I am doing a discharge test today and expect the data will help me decide a LVC. Here's a photo of the test I just started.

The hard case bricks in the main pack are stacked 2 high. Each pair has 4s balance leads joined to an 8s balance connector, then all paralleled to one 8s balance lead for the pack. I have a celllog plugged into that connector to read total voltage for the pack. There is a DVM across the tubular booster pack. The two packs are connected in series with an 8.0 ohm load.

About every 15 minutes I read the voltages and enter them into a spread sheet along with the current time. The spread sheet calculates amps and amp-hours.

I expect to get about 28 amp-hours. So at about 36 volts it will draw about 4.5 amps and take over 6 hours. I will post results later.
 

Attachments

  • 005 5-27-15 Discharge Test.jpg
    005 5-27-15 Discharge Test.jpg
    49.5 KB · Views: 3,625
The following chart shows results from battery test. I divided the main pack voltage by 8 so it could be compared to the booster pack. Based on this test if the packs start at 4.1v per cell and I consider LVC to be 3.7 v per cell of the main pack then total LVC will be about 33.4 (3.7 x 8 + 3.77 = 33.37). So that gives 27.9 amp-hours which is more capacity than I ever use so it should not be an issue.
 

Attachments

  • 006 5-27-15 Discharge Chart.jpg
    006 5-27-15 Discharge Chart.jpg
    26 KB · Views: 3,625
These are the Makita cells from Doctorbass. In my test I did not fully charge or fully discharge them, only used .84 amp hours per cell. So you are right, I could get by with 28 cells.

The 4s hardcase batteries have lots of capacity for a bargain price, but I like the versitility of building my own 18650 packs. My bike has a Q100 motor with a small s06s controller, so 12s is more volts than I want and 8s is just a bit too little.

Next I am going to build a 9s7p version to test out some variations on this concept.

My current lipo pack has a lot of life left, so I have lots of time to work out the build technique, then I will think about better cells or just use more of these makita cells. I have a recumbent bike with plenty of room to mount batteries.

But, today I am going for a long ride.
 
In the last three weeks besides riding, I've also been working on improvements to my battery packs. The 1s35p booster pack works great and I like the extra performance from going from 8s to 9s. Next I built two 5s7p packs and connected them for 10s7p. Here's a photo of the assembled pack and a photo with one pack disassembled.

I changed from using little piles of solder to form the positive contacts on the plates to soldering small brass nuts to the plates. I can locate them more accurately and they have equal heights.

I replaced the drilled wooden spring plates with plastic bushings, the bushings have the same OD as the cells and can be nested in line with the cells. They are easier to make and line up the springs more accurately.

I changed from 3 to 2 tie rods and cover the tie rods with a tough plastic tube so the copper plates can't cut through to the tie rods.

I am using 4mm bullet connectors soldered to the plates at each end to provide the main connections.

Yesterday I was able to go 30 miles with this pack. I'm sure I ride slower than most of you and I also pedal constantly. But, I am very happy with these cells and how the pack has done it's job.
 

Attachments

  • 007 6-15-15 10S7P Version.jpg
    007 6-15-15 10S7P Version.jpg
    39.2 KB · Views: 3,625
  • 008 6-17-15 10s disassembled.jpg
    008 6-17-15 10s disassembled.jpg
    46.2 KB · Views: 3,625
Looks well engineered and kool, but a lot of work. Very clever to insulate the rods with plastic tubing. I built some tubing packs with Nicads many years ago and soldered the cells end to end. Can't do that with LiPo though. Any plans to commercialize this system?
otherDoc
 
You're right it's a lot of work. That's one reason I will not be commercializing it, this is a hobby to me. After building a few versions, I have gotten faster at making the parts, and once they're made I can easily replace cells. Other than just doing this for the heck of it, someday I will transition from hard case 4s bricks to 18650 cells. With this method I can build 9s or 10s, with the bricks I am limited to 8s. Before retiring, I worked 43 years as a mechanical engineer, developing ideas never stops being fun.
 
Since I last posted, I have been using the 2s35p version of this pack in series with with my 8s HK pack. I have come to like the extra speed of 10s. I mount the the two tubes of that pack in my trunk. I don't like the way it fits, so here's another more compact pack that I like a lot. I found a 4 by 4 pvc fence post in a home store. A short piece makes a nice housing for a 2s27p pack.

Here are the parts before assembly along with the tool I use for marking end caps and copper plates for drilling.

This is how I solder the brass nuts to the copper plates. They are held in place with stainless screws which are removed after soldering.

These are how 27 cells fit in the housing.

Then the spring plate add springs are added

Here's a side view of the assembly

Here's an end view of the assembly.

And here's the pack in my trunk, it fits much better and 27p is still more capacity than I need.
 

Attachments

  • 009 7-28-15 parts for 2s27p.jpg
    009 7-28-15 parts for 2s27p.jpg
    43.1 KB · Views: 3,627
  • 010 7-28-15 copper plate.jpg
    010 7-28-15 copper plate.jpg
    37.5 KB · Views: 3,627
  • 011 7-28-15 cells in housing.jpg
    011 7-28-15 cells in housing.jpg
    37 KB · Views: 3,627
  • 012 7-28-15 springs in housing.jpg
    012 7-28-15 springs in housing.jpg
    36.9 KB · Views: 3,626
  • 013 7-28-15 side view of pack.jpg
    013 7-28-15 side view of pack.jpg
    25.8 KB · Views: 3,626
  • 014 7-28-15 end view of pack.jpg
    014 7-28-15 end view of pack.jpg
    25.2 KB · Views: 3,626
  • 015 7-28-15 Pack in trunk.jpg
    015 7-28-15 Pack in trunk.jpg
    50.8 KB · Views: 3,626
Nice :)

Perhaps you might like to consider foam instead of springs to reduce complexity? Origin of the idea here: http://endless-sphere.com/forums/viewtopic.php?f=14&t=57810
 
Punx0r said:
Nice :)

Perhaps you might like to consider foam instead of springs to reduce complexity? Origin of the idea here: http://endless-sphere.com/forums/viewtopic.php?f=14&t=57810

I read that entire thread and reread most of it again because I wanted to fully understand it. It's an excellent idea and workmanship. That method should work on the 2s pack i just posted. On the earlier packs that I built I have 5 layers of cells, I think it takes more than foam to squeese 5 cells and 6 copper plates together. My intension is to eventually replace my HK pack with a 23 inch long 8s pack of 18650 cells. Anyway, that's why I used springs instead of foam or rubber.
 
Dude, commercialize it. I double dog dare ya.

The price of a tab welder, plus the labor is the #1 thing keeping people from adopting these cells en-masse in the ebike world.
If it works well, you can sell a ton of these over a long period f time, as the 18650 cell form factor isn't going anywhere soon, and the cells keep improving.. so a quick swap of cells from time to time would be awesome, rather than going and welding a new pack together.
 
This no solder idea seems to be more logical. I just cant picture it. Its something I believe I have more confidence in doing then magnet no solder.

Whats is the layout of each piece? How is each part hooked up?
-----------------------------------------
I've been trying to figure out the layout, just cant......hopefully I am not too off....
The white thing with +'s first (Is that for placement of battery or is that a battery terminal contact? OR is that to mark where the nuts go on metal sheet?)
then nuts on sheet (Does the white + posts go through the nuts?),
batteries contact + or nuts?

Other end is, sheet (no holes in it) with spike, springs with spring holder, case cover?
Whats the sheet with holes in it, with spike? "the tool I use for marking end caps and copper plates for drilling." So the white thing with +'s is the tool, sheet with holes is not used because its the step prior to sheet with nuts?
Then its compressed for firm electrical contact with 4 corner bolts.
-----------------------------------------
Yeah I've thought about this now, I'm totally giving this a go. I like how you used the bolt and nut to solder nut onto metal, then take out bolt. Any specific kind of solder?
Whats the thickness and type of the metal?
What kind of nuts are you using, brass or something else?
Is the spring, suppose to be a certain diameter for contact onto battery terminal, and you use it for the gnd side because the can itself is grounded right, or would the spring move around too much, wouldnt the nut contact system work best on both ends, or is there fear of damaging the battery terminal.


IdleMind said:
Here are the parts before assembly along with the tool I use for marking end caps and copper plates for drilling.
009%207-28-15%20parts%20for%202s27p_zpsac060vn8.jpg


This is how I solder the brass nuts to the copper plates. They are held in place with stainless screws which are removed after soldering.
010%207-28-15%20copper%20plate_zpsqposhwcn.jpg


These are how 27 cells fit in the housing.
011%207-28-15%20cells%20in%20housing_zps9hkxjgl9.jpg


Then the spring plate add springs are added
012%207-28-15%20springs%20in%20housing_zps7zz22flv.jpg


Here's a side view of the assembly
013%207-28-15%20side%20view%20of%20pack_zpsjaxtl50a.jpg


Here's an end view of the assembly.
014%207-28-15%20end%20view%20of%20pack_zpsyf5ngufb.jpg


And here's the pack in my trunk, it fits much better and 27p is still more capacity than I need.
015%207-28-15%20Pack%20in%20trunk_zpskaxedapl.jpg
 
That is a nice looking bike. Did you build it?
otherDoc
 
markz said:
This no solder idea seems to be more logical. I just cant picture it. Its something I believe I have more confidence in doing then magnet no solder.

I think what you are calling spikes are small screws that are used to hold the brass nuts for soldering and then removed. Here's another photo of the marking tool along some brass nuts and 10-32 stainless screws.

The mrking tool is shown with the pointy ends of the marking screws up and they are extended just beyond the sufrace of the plywood. First cut a copper sheet to loosely fit in the frame in the photo. Then place the marking tool inside the frame on top of the copper sheet. Now tap the marking screws with a hammer to make small impressions in the copper plate. Next drill .140 holes at all these markes. Then assemble the nuts and screws in the holes. You will have what is shown in the photo in my last post.

Now use a propane tourch to heat the plate and nut enough to melt solder. Then touch your solder at the nut to plate interface and the solder will flow and attach the nuts to the plate. The solder won't stick to the screws so now remove them.

Only the plates that contact the plus end of the cell need to have nuts, so in this case two plates have nuts and one is nutless.

The nuts are are just there to form a bump on the copper plate, they could have been formed into the copper plate as another builder demonstated.

This is a 2s pack so there are two layers of batteries with 3 copper plates.
 

Attachments

  • 016 8-3-15 Marking tool.jpg
    016 8-3-15 Marking tool.jpg
    32.6 KB · Views: 3,625
docnjoj said:
That is a nice looking bike. Did you build it?
otherDoc

It is a home built bike. It started life as a gas assisted bike, there's a build thread here.
http://www.motoredbikes.com/showthread.php?37613-Build-of-Recumbent-with-RS-35cc

There is also a thread showing its conversion to e-power here.
http://endless-sphere.com/forums/viewtopic.php?f=2&t=61982

Electric assist on a recumbent changes everything for an older guy. Use the motor to get up to 10mph and you're not wobbling around trying to balance while getting up to speed.
 
I am really liking this idea. Personally I don't need my connections to be spring loaded, but the PVC fence post and pipe arrangements are something I wouldn't mind giving a try. I'm still working out how to configure my 18650 builds, but I'll be considering this a great encasing method!
 
dequinox said:
I am really liking this idea. Personally I don't need my connections to be spring loaded, but the PVC fence post and pipe arrangements are something I wouldn't mind giving a try. I'm still working out how to configure my 18650 builds, but I'll be considering this a great encasing method!


You might be interested some thoughts on these springs.

First, the spring OD is a bit larger than ideal which leaves too little material between the springs when making a plywood spring plate. In the 7p round tube version the cells are packed tight together and I had too much break-out between the holes. So I used thin wall plastic bushings which I made with a lathe. In the square tube version the cells have a tiny bit of space between them and I was able to make a wooden spring plate. I drilled the holes with a 3/4 spade drill that that was ground down to just larger than the spring OD. So if you do decide to use springs, I think you can find some with smaller OD.

Second, the springs are compressed to give about 9 lbs of force. In the round tube design the tube is slit to allow it to be spread open to fit the cells, this means the cells are squeezed which is nice to hold them tight together and prevent cell movement. The springs are strong enough to overcome the tubes radial squeeze and still provide contact force. The OD of the tube was turned smaller before slitting to avoid too much radial squeeze.

Third, notice in the square tube version there are 4 small wooden spacers under the first row of cells to evenlly space them to fill the width of the tube. In this version the cells fit the tube closely, but not tight. The spring force is enough to hold them in position and prevent fretting between the cells.
 
I've been using the 2s booster pack thru mid november and no problems. It's getting too cold for me to take my usual 60 mile ride and I don't like riding my recumbent in the winter. I have a nice wllmart bike that I bought 10 years ago that never gets riden. So i am adding a small front hub motor and making a winter battery pack for 10 mile rides. It is not a tubular pack, but based on the same concepts as I used for my other packs. Here's how far I've gotten.

This will be 8s7p and will use springs to compress each row of cells. I am working on the harness to connect the balance leads to the parallel plates. This will make more sense when I get more photos in a few days.
 

Attachments

  • 027 12-08-15 Winter Pack1.jpg
    027 12-08-15 Winter Pack1.jpg
    47.4 KB · Views: 3,612
The wire harness is done and I have successfully test fitted the parts together. I need to sort out a few bad cells and make a cover befor test driving.
 

Attachments

  • 028 12-09-15 Wire Harness.jpg
    028 12-09-15 Wire Harness.jpg
    47.9 KB · Views: 3,612
  • 029 12-09-15 Assem with Harness.jpg
    029 12-09-15 Assem with Harness.jpg
    50.5 KB · Views: 3,612
Thats a similar idea I had too, using copper nuts and bolts. Adding it to the copper sheet is smart.
I was going to use Acrylic or ABS and make a channel for each row.
 
Great idea, I love it.. A couple of modifications come to mind due to what I already have on hand. Pure copper foil tape with conductive adhesive, two layers put sticky side toward each other with brass or copper washers in between the layers to make the bumps.

Neo disc magnets for the spring compression on one end of your cell stack, copper foil tape again with magnets sandwiched between the adhesive layers and other magnets arranged in the end cover to compress the stack. I'm only thinking this because I have a couple of hundred of them right now. :)

I had a different kind of build planned but I like your idea enough to give it a try I think.
 
Back
Top