I can’t say that any method is the safest and best method as they all have good and bad things with them and so much depends on how well/carefully these different methods are used. Basically, any one installation could be great or lousy with any of the methods.
Rigidity doesn’t matter with NESE though since the cells are not permanently connected to each other. As long as they don’t lose contact at either end you’re okay.
I haven’t used NESE in a board so I don’t know how good a choice it is for that. Posting a separate thread asking about NESE (or searching here, there might be some feedback already) could get you some good info.
I haven’t personally used anything based on NESE or NESE-adjacent techniques (loose cells in holders), but my general feeling from all the posts and discussion I’ve seen on the forum here over the years, is that they all have a tendency to have micro-arcing or microabrasions from the cells shifting in their holders due to flexing and vibrations, which is concerning to me, enough that I don’t think I will ever consider using them for any kind of vehicle.
They are also generally bulkier, heavier, and more expensive due to the required holders.
They do have the notable advantage of being able to easily replace a defective cell if the pack is found to have one, but for me at least, the drawbacks outweigh the benefits.
@Battery_Mooch Hello! I didn’t forget about you, I just got a little lazy
The good news is just swapping out that single cell and re-building the P group fixed the issue. I’ve run the pack down twice and everything is in perfect harmony.
I have 2 of these packs, I’m giving consideration to combining them on a budget trike build, they are off a 4WD so have xt60’s both ends, is it as easy as hooking these up that way?
Nope, not yet.
The devil is in the details. The concept is fine but how it’s actually implemented (tolerances, heat handling, vibration/shock handling, etc.) is what IMO will determine if it’s a reliable product. That will come out over time.
The sort of cross braces on top are for rigidity so without the cells it’s very still along the length of the board perpendicular to the cells. The plan is to epoxy them in place, I’ll need to make sure I’m masking the terminals or something so I don’t clog those up. The top side is exposed so they’re not fully insulated, I was looking into thermal interface materials for lithium packs so if it turns out I need to cool them I can flip it over and mount to a heatsink on the enclosure.
Any obvious pitfalls here? It’s a 1P pack so I’ve spaced the cells out with walls between instead of sitting into and rubbing as they might in one of those jigs people use for glueing P groups. Does that sound reasonable or should I be aiming to widen the bores and fishpaper the cells? The part is designed to be printed standing on edge so the perimeters are along the critical axis and won’t be split apart. Fishpaper rings will be used this is just a test fit
I think if you manage to wear through 3mm of plastic with a battery cell, your board will have encountered a few bigger problems first. Like the bearings rattling out or the deck being shredded into its constituent splinters. Just route the balance wires away from the cans, I think. You can glue the cells into the holder to arrest relative motion, too.
Yeah the current plan is to epoxy the cells in, not sure if I phrased that clearly. The holder is printed in one piece per row of 8 cells, and the cells have just enough play in there ATM that I can add adhesive
If force follows stiffness, is a plastic frame going to be stiffer than sheet metal and spot welds?
I like @Flyboy’s theory about taking stress off the spot welds, but i still dont see how that is practically achievable. Directly welding a solid sheet of metal to a bunch of metal cylinders is a pretty fuckin’ rigid connection. I wouldn’t be surprised if even the JBweld directly to the cell cans fails to alleviate 100% of the force on the welds.
Assuming both are rectangular beams running the full depth of the group, JB Weld infill will be (4.1GPa/207GPa)/(70mm / (2*0.2)mm) = 3.5 times more rigid than the nickel, relieving about 3/4 of the strain. This will be altered a bit by nickel tab and infill geometry, and neglects the influence of the cells themselves in lateral bending.
Also relevant is the elongation at break, which is typically quite large for polymers and very small for tiny shear joints in metal. If a bad weld pops, the epoxy will still be there to prevent the failure from spreading.
I think the concept is sound, but maybe I could use chopped fiberglass in the future.
yeah I think there might be a bit of miscommunication
The issue with p groups not being held together well, or being held by flexy stuff like a compliant silicone, is that the adhesive takes up very little of the load and the cells are able to move relative to each other. So the load is transferred through the cells to the welds, where you don’t want it. A rigid connection is not necessarily a strong connection (as fly said, the total elongation before it snaps shows a low max load), and a little spot weld is ideally not supposed to be a structural join but just electrical.
Red is the movement of one cell relative to the other, tiny blue is the shear strain on the weld. This actually happens because the nickel is relatively rigid. If the nickel tab could flex easily it wouldn’t stress the joint. The aim here is to remove the red strain by constraining the cells
I think it actually will because of the dimensions (tiny tiny contact area on a spot weld vs the whole body of the plastic), but more importantly it’s not just plastic vs welds, it’s glued cells vs glued and reinforced cells. Imagine there’s appropriate glue in there as well, I couldn’t illustrate that cleanly but clarifying I’m not trying to make the plastic-less cells look weak I just didn’t have time to work out how to show that.
Yeah I haven’t messed with fiber filaments yet, seems like a fun idea. Unless you meant fully making the part in a FG mold as an epoxy composite? I’m already having a bit of difficulty with dimensional accuracy of prints (forgot that internal diameters like to shrink) so that sounds like a lot of pain
I feel like such a twat because I did ask for suggestions and then just said nah, but it would be beautiful if I was just completely wrong after a post like that
I meant mixing something like 3mm chopped fiberglass strands into the epoxy untill it becomes thick like paste. Could conceivably increase the elastic modulus of the glue by an order of magnitude, so the residual strain in the welds would go down from about 25% to about 2.5%. The only way to get to 0% would be inventing a new material with infinite stiffness, unfortunately.
