Criticise my modular pack design idea. 28s * 1s20p

Looking for some constructive criticism on this modular pack design idea I would like to follow through with my build. If the idea is bad and I've missed something important then please state so.

Some background information. The attached design is a 2x3 format for 1s6p, this is just a quick design, the final intention is packs in 2x10 format for 1s20p. Then they would all be series'd together to form the overall 28s20p pack. Intention is to use VTC6 cells but got offered some VTC5As for a steal. Entire pack will peak at 700A for 10s boost which should be fine. I also intend to individually cell fuse using some fuse wire, size is yet to be decided.

Bike intention use is off-road enduro and so therefore the battery pack won't be pushed to the limits often.

I'll include the pictures now and then an explanation afterwards.



So on the left you have the "holder". On the right you have the copper busbar. The copper busbar will sit inside the holder with the 18650 cells on top.

The busbar is 1mm thick and is always 24mm wide. Reason for this busbar placement and design is to help with structural rigidity and heat dissipation. I've calculated that this busbar should be able to handle 150A continuously with a 30c rise. I may double the busbar thickness to 2mm to essentially double the continuous rating.



This is essentially how the busbar and cells will sit inside, you will have to imagine the same busbar and holder being on the other side of the cells.



Here is the underside. The reason for those empty circles it to leave the spot-welding of the fuse-wire till last as well as to leave it visible to regularly check them.

Things I'm already thinking of changing:
- Doubling busbar thickness to go from 150A to 300A continuous capability.
- Making the holder itself nearly twice as high as it currently is. Currently only holds about 5mm of the cells.

Edit:
Forgot to mention that the end of the copper busbar is essentially the positive/negative terminal in which some busbars would be used to connect all of the modules together in series.

You seem to think cells won't short through their wrapping?

john61ct said:

You seem to think cells won't short through their wrapping?

Click to expand...

Could you explain some more as I don't understand what you mean. If you mean the positive side of the 18650 cell isn't that why insulating washers/rings are used?

I could cover the busbar side facing the cells with some sort of insulator if it's that big of an issue/risk.

As long as the cells themselves have non-plastic insulator rings under the shrinkwrap, that completely cover the can's upper ring around the positive end, then you shouldn't need anything else.

If they don't, or you aren't sure, you can buy insulator rings made from Nomex, or "fish paper", something that is hard / thick enough and not compressible to allow pressure on them to push the conductors thru them, especially under heat. Plastic rings can deform under pressure especially when heated up, and allow a short to occur.


Regarding the design, I'd just recommend ensuring that your holders are sized to allow for the shrinkwrap on the cells, but still be a fairly tight fit; should make it easier to secure everything together prior to welding.


I forget who, but in one of the 18650-cell-fuse threads, someone is building an ultrasonic welding tip into a CNC setup, to automate the process in a way that should be superior to spotwelding the fuse, from what I understand out of the discussions.

So is it just compression which keeps the cells in contact with the busbar? Most large 18650 packs have the cells constrained by a couple of plastic capture plates and the bus plate goes to the outside. Holes in the busplate allow a wire bond to extend down through the busplate, through the capture plate and onto the end of the cell.

Yes, the plastic holder inside, conductive elements outside would be my druthers.

amberwolf said:

As long as the cells themselves have non-plastic insulator rings under the shrinkwrap, that completely cover the can's upper ring around the positive end, then you shouldn't need anything else.

If they don't, or you aren't sure, you can buy insulator rings made from Nomex, or "fish paper", something that is hard / thick enough and not compressible to allow pressure on them to push the conductors thru them, especially under heat. Plastic rings can deform under pressure especially when heated up, and allow a short to occur.


Regarding the design, I'd just recommend ensuring that your holders are sized to allow for the shrinkwrap on the cells, but still be a fairly tight fit; should make it easier to secure everything together prior to welding.


I forget who, but in one of the 18650-cell-fuse threads, someone is building an ultrasonic welding tip into a CNC setup, to automate the process in a way that should be superior to spotwelding the fuse, from what I understand out of the discussions.

Click to expand...

Yep, I guess I forgot to mention it but the intention is to have a an insulator ring, then the shrink wrap followed by another insulator ring just to be safe.

I'm going to 3D print this small one to check how tight the fit is and adjust accordingly till I find the perfect size. I'm currently using an 18.1mm diameter (what they're advertised as) but will check the cells with a vernier caliper once I receive them.

I've seen that thread and am following it closely but I think I will end up manually doing it using a spot welder.

jonescg said:

So is it just compression which keeps the cells in contact with the busbar? Most large 18650 packs have the cells constrained by a couple of plastic capture plates and the bus plate goes to the outside. Holes in the busplate allow a wire bond to extend down through the busplate, through the capture plate and onto the end of the cell.

Click to expand...

Theoretically the busbar itself should not touch the cells due to the use of insulator rings. The main idea is to use fusewire going from the "terminals" of the cells to the busbar itself. Pretty much like how you've described it with the exception of the busbar itself being on the inside rather than the outside.

I might be missing something important/obvious as to why the busbar is usually on the outside but I haven't realised it. Only thing I can think of is potential shorting but I would think that two layers of insulators would cover that.

Banging around on technical trails is a very different context from smooth roads.

Also, plan on the occasional accident.

john61ct said:

Banging around on technical trails is a very different context from smooth roads.

Also, plan on the occasional accident.

Click to expand...

I really appreciate your criticism and you're making good points but would appreciate the constructive part as well aha.

The modules will sit in a tight metal enclosure so even though the bike as a whole will be taking bumps and bangs not much should reach at a cell level. I understand the points you're making and I'm doing my best to resolve them, I would have thought the two layers of insulators would be enough but I haven't tested it and don't want to find out the hard way.

In that case, would you for example think the only solution is to redesign the modules to have the bus-bars be external. So for example the cell, then the holder, then the bus-bar itself, essentially having 1mm of the plastic holder between the cell and bus-bar, then use fuse-wire to connect the two?

Edit: Since making this post I thought about the external bus-bar and I just can't get it to work. So I guess I will stick with the bus-bar being internal. One obvious solution is to use some sort of non-conductive solution which is the same shape as the bus-bar therefore allowing any sort of contact to essentially become impossible.

Yes I'm just pointing out how critical the issue is. I don't have the experience to advise on the specific solution in your case.

In my design thinking, I've always considered the dielectric cell holders (spacers) innermost, the conductors very well isolated / protected from the cells' casing, and then a dielectric and physically strong outer shell.

Thermally conductive but dielectric insulating potting compound/sealant/adhesives are of great interest to me.

Sorry I can't be more specific.

Cylindrical cells arranged in straight rows and columns like that wastes a ton of space, and space is the biggest premium on our ebikes. Instead of those holders, I'd suggest gluing the blocks of parallel cells in a staggered alignment like Tesla does. Don't use hot glue like some do, as it simply doesn't adher well at all to the plastic slieeves on the cells.

Forget the pressure connection as well. You might be able to get away with it on a low powered ebike for the street, but at higher current and lots of vibration off road problems are guaranteed. There are very good reasons no manufacturer in the world builds battery packs that way.

Why 28s? It's a difficult voltage to match with most controllers.

John in CR said:

Cylindrical cells arranged in straight rows and columns like that wastes a ton of space, and space is the biggest premium on our ebikes. Instead of those holders, I'd suggest gluing the blocks of parallel cells in a staggered alignment like Tesla does. Don't use hot glue like some do, as it simply doesn't adher well at all to the plastic slieeves on the cells.

Forget the pressure connection as well. You might be able to get away with it on a low powered ebike for the street, but at higher current and lots of vibration off road problems are guaranteed. There are very good reasons no manufacturer in the world builds battery packs that way.

Why 28s? It's a difficult voltage to match with most controllers.

Click to expand...

I agree with that first part, have been redesigning it to be of the hexagonal variant leaving about 1mm between cells. I couldn't personally find much information regarding how to correctly glue the cells which is why I haven't personally pushed towards it. Would you be able to explain some more regarding the glue or link me something?

I don't know if I made it clear but there isn't meant to be a pressure connection. The cells will be connected to the busbar with fuse wire. The intention of the plastic holder is just to really hold everything in place.

Reason for 28s is because I'm using a Sevcon Size 6 and Zero 75-7R motor, best way to make the most of them both is utilise the Sevcons full voltage.

Appreciate the criticism and advice!

c70r said:

John in CR said:

Cylindrical cells arranged in straight rows and columns like that wastes a ton of space, and space is the biggest premium on our ebikes. Instead of those holders, I'd suggest gluing the blocks of parallel cells in a staggered alignment like Tesla does. Don't use hot glue like some do, as it simply doesn't adher well at all to the plastic slieeves on the cells.

Forget the pressure connection as well. You might be able to get away with it on a low powered ebike for the street, but at higher current and lots of vibration off road problems are guaranteed. There are very good reasons no manufacturer in the world builds battery packs that way.

Why 28s? It's a difficult voltage to match with most controllers.

Click to expand...

I agree with that first part, have been redesigning it to be of the hexagonal variant leaving about 1mm between cells. I couldn't personally find much information regarding how to correctly glue the cells which is why I haven't personally pushed towards it. Would you be able to explain some more regarding the glue or link me something?

I don't know if I made it clear but there isn't meant to be a pressure connection. The cells will be connected to the busbar with fuse wire. The intention of the plastic holder is just to really hold everything in place.

Reason for 28s is because I'm using a Sevcon Size 6 and Zero 75-7R motor, best way to make the most of them both is utilise the Sevcons full voltage.

Appreciate the criticism and advice!

Click to expand...

I built some packs with A123 26650 cells years ago and used a simple cheap and quite strong polyurethane glue available down here that expands some when it sets off...just 2 ribbons of glue down the sides of 1 layer of cells and stack the next layer with each cell touching 2 others. I let the parallel blocks touch each other since they are the same voltage, and I separated the series blocks from each other with a layer of 3M duct tape on each. After soldering all of my electrical connections I sprayed on 3 layers of conformal coating so I could flow air down each side of the stack over the ends of the cells for cooling, which proved unnecessary. I included passive top of charge balance boards with those packs, with that and the controller LVC as the only BMS functions. Those cells were harvested from warranty return 36V Dewalt packs back in 2008, all all are still in service, though capacity is down to 70-80% of nominal capacity.

I can appreciate wanting to get the most out of that motor/controller combo. I hope you have access to a dyno and the programming stuff to do so. There's no way I'd go with 18650's unless you plan to fit 100ah or so. Tesla gets away with it because they use thousands of cells. Instead I'd go with a 16s and 12s Chevy module from the same low mileage car for 50+ah, or a 2p28s pack of the bolt together 26ah Panasonic cells Ford uses in some cars. I leave the pack building to robots these days.