I’ve thought a bit more on how to store the batteries in the canister and I’ll briefly talk about the pros and cons that I see with the different solutions I’ve seen or come up with.
There are two issues. The lid which closes of the battery compartment should be removable, this means there needs to be an O-ring and a screw. The problem is that this means that there will need to be room for those things. With a diameter of 6.5 mm of the outer wall (measured from a borrowed LED-canister torch with 60 mm outer diameter) and the 18650-batteries this leaves little room for an O-ring or a place to place a screw. One solution I thought of was to have a wall of non-machined material that would become thicker in the middle to support screwing into it. After doing a drawing it became clear that the wall could be about 1.5 mm. Here I find a lack of real world knowledge is an inhibitor to making a clear decision at this point.
At this point I turned to a 2D Static Structural test case in ANSYS to see how the wall thickness compresses the body. I did this test with essentially taking a look at the cylinder on just a plane and then took a quarter of the body, as it is symmetrical.
I did the test with 6 and 12 bar and got the following results:
Wall thickness (mm)
Deformation at 6 bar (mm)
Deformation at 12 bar (mm)
I didn’t bother doing any convergence study, just a quick hack with a fine mesh. The point of this test is to see in what magnitude the deformation is in any case. The pressure would at some point transfer to the batteries and I don’t really know what kind of mechanical stress they can take.
After seeing the small size of compression I figure that if I should try to machine with an internal wall of 1.5 mm between the batteries and a bigger core in the middle to screw the lid into.
A M2 screw would fit and the screwing force is not very important for the sealing, more so the O-ring. How to fit the O-ring into this configuration is the next thing to figure out, as there is some lip around the perimeter, but not everywhere and it doesn’t really fit like it is designed as of now. Anyone how how a sealing reacts when the pressure change is more or less instantaneous?
This became longer than I expected, I’ll update with more about target metrics in the next post!