Slim solderless 216 Wh 13S1P pack knocked me off my horse (21700 Moli P45B)

[nefles]

Hi,
I've learned a lot here, so want to give back with this 13S1P DIY build - in case someone else might be wondering if this is feasible
I have a 750 Wh commercial pack, great for long hauls, but a pain in the back for those short 20 odd km rides I end up doing a lot all the year round (also commutes). Pain in the back, coz my bike is a really funny shape, leaving no easy space to fit a pack to the frame (it is a so called "Uber V", ie a 25 yo Cannondale "Super V" modified with a lefty fork, and a Scalpel rear triangle so disk brakes can be fitted all round, and finally a rear hub brushless motor + 25A controller + cycle analyst):

So the idea was to see if I might get enough juice from a single string of cells, such as 13S1P to reach the max 48V my controller can deal with. And 1P only because the only place that's not in the way, is in the gap between the front wheel and under the main diagonal tube, which is so oversized (70mm diam) that cells could fit orthogonally and not stick out. This 500x70x40mm space could fit 13 cells side by side followed by the BMS.

Problem is getting enough amps from 1P only, 25A is hard on a cell, so I turned to the slightly larger 21700 Molicel P45B that seemed to pack a punch, with up to 30A continuous discharge, so should handle my controller's max 25A, yielding in theory 1200W (more than my hefty 750 Wh that's just 36V). With a capacity of 4.5Ah x 48V = 216 Wh, which should push for the entire length of my routine 20-30 km trips in hilly Provence (usually 5-8 Wh/km). I opted for neodyme magnets to press 0.1mm copper bus bars onto the cells, because I did'nt have a spot welder, and didn't want to loose any of each cell's capacity during welding, plus I might want to experiment several shapes for the pack, plus it is easier to replace a bad cell. Here's a few pics of the making :


The BMS is a JK Smart Active Balance BMS BD4A17S4P - (8S - 17S) - 40A:


The case looks fine, and could be run over by a bus, but I'm not satisfied as it is heavy, and also I made it a little too wide and high. It is made of aluminium/plastic sandwich sheets (al-plastic-al, 3mm thickness) with aluminium angle rails with 3mm grooves, everything glued apart from one end cap, and top panel which slides out. The pink shrink wrap was a test.


I am happy to report that the pack behaves exactly as planned: 24.25 A according to Cycle Analyst (over 1100W), and over 180 Wh with a 4.17-3.15V discharge (I don't need any more, and don't know how bad it is for the cells to go further down in voltage). On a normal ride the temp probe in between the cells doesn't go much above 30°C. When I emptied half the pack in a straight continuous 400m climb, temp peaked over 40°C, so no issues there.
In fact the pack has so much juice, it nearly knocked me off the bike on the first trial. I don't need that much, so I'm gonna lower the throttle max amps
Also the bike is far better balanced now, it used to be back heavy with that hefty hub motor, much better grip at the front, much less tendency to do wheelies in steep climbes. And oh boy is it nice to forget that back pack, and have free shoulders whilst riding
The pack's nickname will be "Into the wild" :

 

[gromike]

Interesting use of magnets, something I would've never thought of.

 

[nefles]

Interesting use of magnets, something I would've never thought of.

I wouldn't have thought of this either, it has been reported here a few times, such as here 40T, neodymium magnets, plasti dip cycle test
What I was worried about, is wether the magnet pressure would let 25 A through. I considered using all kinds of conductive pastes, whic are rare, cost a bomb, makes the contact more slippery, and must be dosed just right ar risk shorts if it gets squeezed out from the positive end and reaches the sleeve... Currently it works just fine in "dry" mode BMS reported "cell wire resistance" as 0.3 Ohms, I don't know haw good/bad that is...

 

[raedy07]

I wouldn't have thought of this either, it has been reported here a few times, such as here 40T, neodymium magnets, plasti dip cycle test
What I was worried about, is wether the magnet pressure would let 25 A through. . Currently it works just fine in "dry" mode BMS reported "cell wire resistance" as 0.3 Ohms, I don't know haw good/bad that is...

i also experimented with mags, but only on stationary batteries. wasn't sure how they would do with vibration. but being additionally taped into position seems to work well for you - do you know d holding-power of your mags?
another idea was to use springs, i guess even >0.2mm nickel or copper strips could be used. just never found d right springs.

 

[BlueSwordM]

I would not recommend using magnets at all on a portable application...

You never know when the magnets will slide and cause a short somewhere.

@raedy07 Springs would be a good option, but you'd be force to use phosphor bronze springs at the minimum, or beryllium copper at the slightly higher end.

 

[Diggs]

Interesting battery concept. I like it. Some people don't understand the power of neodymium magnets. Small ones can pinch to the point of drawing blood. Large ones are dangerous.

 

[Zambam]

I would not recommend using magnets at all on a portable application...

You never know when the magnets will slide and cause a short somewhere.

@raedy07 Springs would be a good option, but you'd be force to use phosphor bronze springs at the minimum, or beryllium copper at the slightly higher end.

Great idea if you do not have a spot welder or do not want to solder (provided the cell terminals are magnetic. Too bad the cells I will be building packs with are not magnetic, otherwise I'd be trying the magnet trick. These cells also require expensive spot welders or laser welders that are outside of my budget so I'll be soldering )! Did you see that he also taped the magnets/ copper strips so they cannot shift? Is that acceptable then?

I'd like to see someone do a max current test of this setup.

 

[BlueSwordM]

Great idea if you do not have a spot welder or do not want to solder (provided the cell terminals are magnetic)! Did you see that he also taped the magnets/ copper strips so they cannot shift? Is that acceptable then?

I'd like to see someone do a max current test of this setup.

Yeah, I did see that he taped the magnets with good tape.

I'm just worried about anything that isn't a soldered spring connection at the bare minimum.

 

[Zambam]

Yeah, I did see that he taped the magnets with good tape.

I'm just worried about anything that isn't a soldered spring connection at the bare minimum.

Your concern being current carrying capacity?

 

[docw009]

Good idea! Magnets have also been used and discussed by others here.

yet another solderless DIY battery pack via NIB's

Subsribed

endless-sphere.com

Magnets holding copper bus onto cell, no-weld no-solder

This subject came up in another thread, and I have moved it to its own thread, due to the renewed interest.... Original thread found here: http://endless-sphere.com/forums/viewtopic.php?f=3&t=104395 Although metal magnets conduct electricity, they have high resistance, so I would experiment...

endless-sphere.com

I think where magnets might have issues is on layouts which use multiple strips of conductor. Keep it all one single layer. And the cells do have to be locked in place. No movement.

I believe you used the insulating washers on the positive ends? Those are considered essential for welded nickel, which have no pressure on the negative edge of the can where it overlaps the cell end. Probably even more important for magnets, which exhibit heavy pressure, If somehow they moved over the edge, now you're beating on the vulnerable part of the cell,

 

[offGridDownUnder]

A thought about the magnets shifting...

I can imagine making an in-line holder that all the cells slide into to maintain their relative positions. How about a second (and third) in-line holder on either side to also hold the magnets in position and alignment, rather than just relying on the magnetic attraction?

The cell holder would be almost as wide as the cells are long, and the magnet holders would be much thinner - a bit less than the thickness of the magnets used and then wrap the entire assembly as the maker has done. Then the cells and magnets could not move from their positions.

A variant - the two outer holders for the magnets are closed on one side and hold the magnets (in addition to the magnetic attraction) so that wrapping (or making a compression case for) the magnet holders compresses the entire assembly.

If the magnet attachment approach provides acceptable electric characteristics this would be an approach that permitted easy disassembly to check or replace cells.

 

[Zambam]

Another idea: Make donut insulator which the centers the magnet in the plastic battery holders so it cannot shift. A dab of hot glue in the middle of the copper strip (between 2 cells) keeps the strip from shifting.

 

[amberwolf]

BMS reported "cell wire resistance" as 0.3 Ohms, I don't know haw good/bad that is...

That's 300milliohm, which is probably a few dozen times higher than the actual cell resistance...so pretty bad.

I've only used magnets to secure wires to cells under test for pack repairs--I wouldn't use them in any pack build; there's too many possible ways for the connection to fail, including shorting multiple cells together. There are potential ways to mitigate these issues,some of which have been tested in various parts of the linked pages from other posters; I'd rather just avoid having to worry I missed some potential failure mode that isn't handled by those.

 

[Hillhater]

Wire / pack resistance of 0.3 ohm would suggest a voltage sag of approx 8.0v at 25 amps ??

 

[leffex]

How has the pack behaved. How is the performance of the pack now and has it worked flawless. What is your verdict, nefles
?