@Vanarian is your man if you ask about magnets and if they reliable for your application.
1 Like
Where the magnets would be? If the magnets work in shearing (you slide them when pulling out) I would say that 6kg is not enough, if you pull them it may be
I would do what I did on my ebike battery, leave places to put a lot of magnets, and fill in as necessary
@Andy87 @Zach Thanks haha Iām honored!
AFAIK magnets donāt wear out due to vibrations themselves! Itās all about temp management (keep them under the Curie point).
Giving you a proper reply when Iām home!
2 Likes
I would change it so the touch each other, make a cavity and glue them in with epoxy, 1mm the force drops a lot
1 Like
This (and Iām home!)
Letās say you have 1x 6kg magnet, at 1mm distance you may drop around 3.5kg, 2mm you drop at 1kgā¦
But strictly speaking thatās still a lot of force. Hereās the catch : you can roughly multiply your magnetic force per your number of magnets. So even with losses due to clearance, you can still achieve strong link. Add as much magnets as you need, you can order from Supermagnete to try it out.
Also even with a lateral slide mechanism instead of vertical one (you slide the magnet out of axis instead of pulling away straight in its axis) it will hold everything in place easily.
You should use the strongest possible magnets so you have a sturdy assembly with somewhat easy lateral slide in/out.
Thatās exactly what I do with my battery packs. Iām more worried about the connectors wearing out however. Are you using xt60/90 or custom variant?
2 Likes
Use this to calculate the forces, really handy tool
2 Likes
My only concern with gluing is the adhesion. If the CA fails to bond properly to the TPU or the magnet, weāre looking at a failure mechanism. Might be worth researching nonetheless, especially considering how strong modern day CAās are
1 Like
Firstly, thank you for your input! So after doing a bit more research, if i use the above magnets, in a linear slide mechanism, with a 0mm gap between magnet pairs, iād only be seeing 3kg of force (1.5 x 2 sets of magnets). I think this should still be sufficient, but im not 100% certain. I think my best bet is to test really.
Fun fact of the day; A magnets force will not change depending if it is attracted to a magnetic surface (eg iron) or if it is attracted to another magnet or opposing polarity.
Dual XT60ās. Hopefully should be more than sufficient for the time being. The problem with this was the amperage requirements for each pack, which was explained in some earlier posts. You thinking they could wear out and become problematic?
2 Likes
That is one fear I have with my design, while the magnets are pretty well held in place, the steel plate that they attract is glued to a 3D printed piece
I would design it as if the connectors had no holding force, while the friction on XT90 is really strong when new, after a bunch of insertion they loose most of the force
1 Like
Man, this looks like a sweet idea! I feel like this concept could lend itself well to flexible battery packs. If you built the āspineā in a hinged way then the individual packs could move a bit individually.
1 Like
Update! Weāre Prototyping
So, this is printed in PETG, and I think i might keep the magazines in PETG, since even if printed in TPU, the cells are rigid and would not allow much flex of the āmagazineā anyway. Printing the mags in PETG is a) easier and b) provides more impact resistance to the cells.
I need to order the Magnets, in order to test their adhesion to TPU and PETG, to ensure that wont be a mechanism of failure.
Furthermore, Have to ensure that there is enough structural rigidity with those tabs to ensure that they will be adequate to support the mag without snapping.
So far, really pleased. Really excited to finally see it come to life
15 Likes
Good to see your progress on this.
Have you considered attaching a 1S BMS to each of the packs? Perhaps something like this?
Each P-group would course you would need pos and neg charge leads, maybe 20AWG, I would expect they would attach in a similar fashion to the P-group case as the discharge connectors, with the discharge bypassing the BMS.
This would mean that no central BMS is necessary, as each pack can charge while plugged in on the board, or on a separately dedicated charger depending on the userās preference.
Charging might be an issue though. Havenāt thought it that far through. Plus you have 11 more points of failure with the addition BMSes on the P-groups.
Thoughts?
1 Like
I did! Was looking at integrating a TP4056 Module, which would also mean the packs could individually be charged on the go with a run of the mill usb-c cable, however, those boards are only capable of 1A charging, which for 1.6 Ah packs means a really long charge time.
This in tandem to the fact that there is no easy way to monitor State Of Charge per pack, meant that if one were to try to charge a pack over usb, they would have to charge each one to full in order to ensure that when all the packs were reconnected, that they would be the same voltage.
Additionally, due the potting, i began to realise that if i integrated a charge module in each pack, that there was a good chance that heat could become an issue.
On that note, i wish there was an easy implementation for thermistors in regards to off the shelf BMSās that would allow for 12 seperate inputs, and could take action based on temp.
Thinking that this may be doable via raspberry Pi or something, which then communicates to BMS via CANbus, but would be awesome to just have one integrated package
3 Likes
This project is awesome! What are you planning on using for the bus connection in the 4P groups?
I just made a 12s4p out of p42aās, and I opted for NESE because (amongst other reasons) it can handle the massive current that these cells can pump out. You talked about nickel earlier in the thread, but finding a nickel solution that can handle the current without getting hot (especially important if you are potting the p-packs, i would thing) just seems like a big headache.
Could this system work with tinned copper braid like the @winfly packs? Or are you decided on spot welding nickel for space constraints?
2 Likes
Copper braid is what im looking at right now. Gonna have to do some testing to see about how im going to connect it to the cells, As I know it can be particularly tricky to solder with and given my space constraints, I need it to be low profile.
Weāll see 
2 Likes
Would using winfly style compression be an issue?
Not really, itās just that they havenāt been designed that way. Worth a thought though 
1 Like
https://uk.rs-online.com/web/p/potting-compounds/9184992/
In other news, prototyping is underway. Had the bright idea to possibly integrate temperature sensors in each pack, and then attempt to develop a monitoring system that can communicate to the VESC over can to trigger a limo mode if cells start getting hot, or if any of them reach cutoff.
Who knows if thatāll happen for v1
2 Likes