Best filament for a battery enclosure?

Yeah it’s hard to draw the line really, for example let’s say both of the redundant BMS temperature sensors fail and it doesn’t stop the battery temperature at 60C. Well then in order to continuously drain a high current from the cells in order to get them to heat you would need to:

  • Have the throttle/pedal assist system (PAS) stuck on full or a high level
  • Be going up a hill or somewhere that puts a high load on the motor continuously
  • Have both the bike computer and motor controller temperature sensors also fail to detect the resulting high temperatures in the motor and motor wiring respectively and not cut the power.

And this is on top of the other safety features built into the battery such as individually fused cells to prevent a dead cell from shorting the others in its parallel group, or equally a dead short on the output would blow all the fuses on the cells if for some reason the BMS doesn’t cut the power first as it should, as well as enough cell-to-cell spacing to prevent a thermal runaway in one cell from propagating to its neighbors and then across the rest of the whole pack, etc.

So yes, I am relying on the BMS to do most of the heavy lifting with regards to safety (over/under current, voltage and temperature, as well as cell balancing), but there are other systems in the chain that have their own safety features, as well as built in hardware safety features, and so I think you could demonstrate that all of those combined means that you’ve done everything reasonable to ensure that this remote scenario of the cells getting dangerously hot does not happen.

There are of course a few things that would make the design even safer such as:

However, these would significantly increase the cost and complexity of the design and would be overkill considering that no ebike batteries have those features, and a lot of them don’t even have some of the basics such as individually fused cells or watertightness. So for now I’m going to stick with plastic internals and externals since that is the norm for ebike batteries anyway.

Also, I agree it is a bit frustrating that you can’t take items with batteries over say 100Wh on a lot of transportation, so for example taking a regular bicycle on public transport such as busses, trains or planes, or having it shipped somewhere is fine, but ebikes are either banned on public transport or incur extra costs if shipped.

It’s understandable on the one hand because the transport/shipping company have no idea what the quality of the pack is and whether it was put together with some low grade nickel plated steel strips with sharp edges in China with the cells just glued together, or whether it was built with attention to detail and the intent to be safe and repairable.

It’s ironic, because having just talked about the features I’ve implemented in the design, the final product in this case would probably be put in the same category as a dodgy chinese battery for general purposes simply because it’s a one-off DIY pack that I probably can’t afford to have UL certified or similar.

1 Like

@Snelso91, you will definitely want to use glue. ASA is prone to warping. You will also want to heat the bed and chamber at least 15-20 minutes prior to printing. Just use the glue and don’t look back, trust me. Not worth the headache trying to print with ASA without it.

I want to give you some information because I have already gained experience with this. Not only with the construction, but also in practice I have experience with the fact that the housing should also protect the batteries inside in case of falls or blows.

  1. my first favourite is PLA / PLA+. there are PLA types that are really good, that nothing can actually beat them. The only thing is the moulding temperature, which doesn’t play as big a role in the application as you might initially think (unless you want to use the housing next to a blast furnace). Otherwise, my first housing would already be heavily deformed, but it actually shows no deformation at all, despite having been used for a year, even in summer at temperatures of around 30°C (or a little more) in the sun, on a bicycle. There is PLA that you can neither break nor significantly damage after printing. But basically (almost) any PLA will do.

  2. my second housing is printed from PETG because PETG is more stable with thinner walls (for example in the housing). Another advantage of PETG is a slightly higher softening temperature compared to PLA and that it does not shrink as much. If a better internal fit is required for components, PETG is the better choice. With PLA, shrinkage over time would theoretically have to be taken into account to compensate for this disadvantage. But then the parts would not fit immediately after printing, but only after a few days or weeks. PETG is more difficult to process than PLA. Maximum stability is achieved at slow printing speeds (around 35mm/s to 60mm/s) and high nozzle temperatures (250°C to 260°C). *I do not yet have any experience with PETG housings in the event of impacts or falls, but I assume that it behaves similarly well to PLA.

To build the housing

You must always pay attention to the stability of the design: not too few walls (perimeter). Filling ensures that the housing retains its shape in any case. When designing in the CAD software, also consider how what you are designing could or will behave under load scenarios. The stronger/thicker a wall of the enclosure is, the more difficult it will be to break or damage. This is especially true for the outer walls of the enclosure. If you use too few bottom and top layers, outer and inner walls (perimeter) when printing, then it can happen that these are dented by some circumstance and get holes.

Good luck!

If you can tolerate a higher material cost, then PET-CF seems like it may be an all around “best” filament, provided you anneal it to get its full benefit. It only comes in black, but I’m assuming it can be painted if you need it to be some other color.

If it were me, I’d do any iterative prototyping in PETG to save money, but do the final print in PET-CF. Strong, and very high temperature resistance, beating out nylon and allegedly having little if any of the creep under load that I’ve read nylon has. Also, according to the youtube videos, it prints with a beautiful surface finish.

Carbon fiber does conduct electricity. Carbon fiber filled filament typically doesn’t, but it can depending on density of the fibers. I’d probably insulate electronic parts from the enclosure if using CF just to be on the safe side.

1 Like

Take a look at this filament SMARTFIL ABS FP Fireproof. Not expensive at all and gives a extra layer of security.

I would also print some inner layer in tpu for shock absorption.
I used to race drones and all our batteries and cameras were protected by some tpu enclosure.

About ABS vs ASA UV protection a good ABS works as good as a ASA filament.
I can recommend the ABS+ from Esun, i’m using it for almost 2years in a hidroponic tower that get’s exposed to 40+ C degrees in the summer and it’s like new.
PLA in my climate can’t even survive a single day in my summer conditions without getting deformed.

4 Likes

I like in Singapore and I know for sure PLA won’t last more tha few years exposed to our weather. I have seen it fail many times.

The UV and humidity will destroy it

Also if you SHIP your goods in a container and it gets stuck in a metal shipping container in the sun, or in the back of a truck … LOL yeah. That’s like “hot car” temperature and it’ll soften it.

Do not use PLA for outdoor goods that might be sold or used internationally. Your climate may take it, but others won’t.

edit Also OP is making a battery enclosure, which will be subjected to additional heating from the batts.

1 Like

Thanks for the suggestions.
Regarding PLA, I agree with the other commenters in that it seems too risky in terms of the low heat resistance temperature, especially in the long term. And for PLA+, it is still PLA just with additives, so correct me if I’m wrong but I wouldn’t expect it to have a significantly higher heat deflection temperature (HDT) than normal PLA. Similarly, I would expect both PLA and PLA+ to be too brittle for this application if they were to get into a crash?

As for PETG, I agree it is definitely a good filament, and it would have been my filament of choice if I was limited to an unenclosed printer such as the Bambu A1, my old bed slinger printer, etc.
But given the enclosure of the P1S, it seems like there would be better suited filaments such as ABS/ASA, PC, PET, etc that other commenters have mentioned.

Lastly, thanks for the design tips.
For this particular project, I have designed the outer housing to have 5mm thick walls in all directions to make the box as sturdy as possible. I will also be using a high strength printing profile such as using 4 perimeters with 140% line width for inner perims, sparse infill and solid infill for greater strength, as well as 0.12mm layers for better layer bonding.
I based those decisions mainly off a few useful videos from CNC Kitchen which discuss how line width, layer height, number of perimeters, etc. effect different forms of strength (tensile, bending, etc.)

I also chose 60% gyroid infill which is about the max percentage that most people suggest before diminishing returns afaik.
It also makes sense that not having 100% infill would improve impact strength as the material would compress more instead of shattering due the voids in the infill, as per the paper “3D printing of gyroid structures for superior structural behaviour” by Catarina Maia Moreira da Silva, where they found that 60-80% was optimal for maximum impact strength with gyroid infill.
I chose gyroid infill because it’s equally strong in all directions and in this application one of the biggest dangers is an impact, which could come from any direction.

you’ve got to be careful when setting your product requirements and especially when communicating them to users of the product afterward. several of the requirements you described are not achievable in the manner you described them:

  1. Must be impact resistant. If the bike gets into a crash I need to be 100% sure that the enclosure will not be punctured and crush the battery cells inside.

you can’t be 100% certain the enclosure won’t be punctured or that the cells won’t be crushed. you can take a functionally safe approach to this and work to make them fail in a safe manner if/when they are crushed. for example, design it intentionally so if enough force is exerted to crush the cells then the mechanical linkage to the ebike would also fail, causing the cells to fall away from the rider? you’d want long wiring too so the wires don’t keep the cells nearby where fire hazard to the rider is greatest.

  1. Must be durable. If it gets scuffed up over the years just from rubbing against things occasionally this mustn’t meaningfully compromise the strength of the part.

a sacrificial layer around the critical structure should help with this, but you’d want to use different colors perhaps and a disclosure that “if the outer shell degrades to the point that you can see the brightly colored inner shell, it is time to replace the pack as it is no longer protected against minor scuffs and scrapes”.

  1. Must be able to survive outdoors for an infinite period of time. So this means a decently high deformation temperature to resist sitting in direct sunlight in the peak of summer for hours, as well as things like enough UV resistance to survive outside.

seriously, infinite? plastics degrade over time, there is absolutely no way to defensibly use this language in a product description. as for deformation, i hope that the cells inside can appropriately be used to assist with mechanical rigidity. if so, then even PLA might survive well enough. otherwise you’ll need to constrain the environmental variables and pick a material based on those. e.g. you could set your max temp to the highest allowed temp of your electronics and battery pack (likely around 125C), and perhaps argue that leaving it in direct sunlight in the middle of Arizona (or similar hot place) would be unsafe for parts that have nothing to do with the plastics used… therefore are a violation of the usage guidelines. if that’s ok as a starting point, you’ll need to look for plastics that won’t deform until well above 125C. you might also want to find a passive temp logger (something like a sticker that permanently changes color when it gets hotter than 125C).

  1. Must not be electrically conductive. I’m not sure if carbon fibre filled filaments have this issue, but the filament cannot be electrically conductive as I will probably be using it for the internals as well which will be in direct contact with the battery cells including their terminals.

you’re likely fine here with pretty much any filament, however moisture could be an issue. moisture buildup inside the enclosure can lead to partial shorts; moisture can wick along the lamination lines (the horizontal ridges, one per layer, that form when you print something). you’ll need to think carefully about this and design something to prevent it from being an issue.

  1. Must be somewhat stiff, but I don’t think it has to be super stiff because it’s only holding the weight of the batteries and itself in a normal scenario (about 4kg). Mainly I’m saying this to rule out things like TPU which have incredible impact resistance/toughness, but which would deform too much under an impact or the static weight of the battery cells.

you’re likely fine with pretty much anything on this one too. I print with TPEE sometimes and despite being a flexible filament it’s easy to make it relatively rigid - i don’t suggest it here but really only because it has a lower glass temp than other plastics, but if heat weren’t an issue i think you’d be ok with a TPU or TPEE on the harder side of the scale. you might still want to use some flexible filament for the contact interface between the bike and the plastic, else vibration and shock will likely damage the enclosure more quickly.

Thanks for pointing that out. If true it could maybe be worked around, but it might make more sense to avoid the problem altogether. Maybe a filament made with glass fiber instead of carbon fiber would work almost as well? The SMARTFIL ABS FP that marcio.coragem mentioned above sounds very interesting also.

Thanks for the constructive criticism. I admit I did write the requirements here in a slightly over the top manner. Although I should clarify that none of the wording here is going into any kind of product description, as currently this is a one-off project for myself.

Yes, obviously it is impossible to 100% guarantee that something can survive any impact, especially one where the forces are not precisely defined. For example specifying something to survive the impact of building being dropped on it vs the impact of hitting something solid on the bicycle at 30 mph (such as a car or lamp post) are 2 very different scales. The latter is roughly the worst case scenario that would be reasonable in this context.

To clarify, what I meant to say is that it should be as impact resistant as is reasonable in this context (a 3D printed plastic enclosure). If I were to make the box out of thick steel instead, then it would obviously be able to survive much greater impacts, but the requirement is really about trying to mitigate this risk as much as possible by choosing the best filament possible that is printable on a P1S, as well as designing appropriately thick walls and using high strength slicer settings.

This wasn’t something I had considered yet as I expect it would take a lot of testing or simulation to get a correct sacrificial structure that holds the battery securely to the frame in normal use, but fails when a big enough impact is applied?

Currently the mount on the frame is designed to have something similar to female dovetail joints, with the male versions on the battery which would simply slide in from the side and be locked in place with a conventional key lock via a pin. Now if I make the dovetail joints thin enough, I could see them shearing at the corners which is where I expect the highest stress would be, but again this would probably take several simulations to get right.

Also, at the moment the main battery connector is an AS120, which is non-locking (essentially friction fit), so I would expect if there was a violent enough pull on the cable it would separate naturally and let the battery be ejected from the frame.

Yes, I already saw that @RMB suggested using a TPU sleeve as outer layer which would take the majority of the impacts as well as most of the scuffs too since it is ultra durable and impact resistant. The only thing I can think this might be harmful for is heat dissipation during the summer due to the extra insulation.

Yeah, again poor wording on my part sorry. My main focus was really that it has to have a high enough HDT to survive the worst case which would be outside in direct sunlight for the whole day during the height of summer, and must not become brittle from UV exposure like ABS does.

Currently the cells do act a bit to improve the mechanical stability of the pack, but that is mostly due to the plastic structure they are sitting in as well as the internal screws in the inner structure which compresses the end capture plates onto the cells to hold them in this structure. Therefore, if this inner structure were to deform significantly, then the cells could move a bit. However, this is somewhat by design, because one of the key aims of this pack was to make it highly repairable, which means that when the end capture plates are removed, any individual cell can be replaced after their fuse wires have been cut.

In the UK where I’m based, summer temperatures are only about 35C at most, but that’s not accounting for the effect of direct sunlight, which is why I suggested a rough maximum temperature of around 60C based on some googling as I mentioned in a previous comment. But even if this is incorrect, and the inside of the enclosure gets a bit hotter than that, I will likely have bigger problems from the cells degrading (but not catching fire), than the material deflecting.

Regardless, ASA and the other materials mentioned here (other than PLA and PETG) all have HDTs of 100C or higher, which definitely seems safe to use in summer, especially since people already make car parts out of them such as bodywork, which can be in the exact worst case conditions I’ve been describing.

Yes, watertightness was a big concern for me initially due the fact that I need the outer enclosure in particular to be watertight to prevent water ingress in the worst case of say the battery being dropped into shallow water in a crash or otherwise, and I know that it is incredibly difficult to make a 3D printed object truly watertight due to the layer lines and seams.

Therefore, my main method to mitigate this is probably going to be to apply an epoxy coating to either the interior or exterior faces of the outer housing, as well as using gaskets and screws to secure the lid, and sealing the cable hole with silicone sealant.

Also, regarding carbon fibre filament conductivity, I agree with you and @RandomKhaos that for the most part it is not conductive, but it is important to be cautious because there are some that are, at least partially, such as shown in this test, where Bambu Lab’s own PAHT-CF is shown to be conductive when using threaded inserts, which I do plan to utilise heavily in my design along with appropriate screws.

1 Like

That sure sounds like a lot of extra work. Maybe it would be easier to 3D print an ABS shell to act as a water barrier would be easier? If it were leaky then perhaps steeping it in an acetone fog would smooth the surface enough that it would become waterproof.

Yes, this was something I considered recently as being a possible alternative, but just from a cursory search now, it seems others such as prusa have tested it and it didn’t make the part watertight enough to be reliable, so it’s likely that I would still need to apply an epoxy coat even after acetone smoothing.

They even say in this article that one of the best combinations they had was using PETG for its mechanical properties, and then simply applying an epoxy coating afterwards made it completely watertight, and the combination of PETG + epoxy coat was going to be my original plan for a strong watertight 3D printed box, before I considered switching to a higher grade filament by starting this thread.

Also I agree it is an annoying post processing step, but unless you have any other suggestions, unfortunately I can’t think of an easier way to make a 3D print truly watertight whilst maintaining the same dimensions.

For example, it’s possible that annealing could improve watertightness, but it would likely introduct warping and shrinkage which might be difficult to compensate for when designing the part.

Similarly, another method I’ve seen which is useful for making transparent prints is to overextrude by a few % on purpose in order to remove any air bubbles from inside the print, but I don’t know if this can be relied upon to fully seal up the layer lines, even if it does work to improve optical clarity of transparent prints.

Also, overextruding implies that it will need some post processing such as sanding or drilling to bring it back to the right dimensions anyway, as holes would probably be shrunk and outer dimensions would be expanded.

Interesting suggestion, based on bambu lab’s filament guide and your comments, I can see that it has great bending strength, stiffness and high HDT, but it looks like the impact resistance is lower than most of the other materials including, ABS, ASA, PETG, and I’m not sure if the high stiffness and >100C HDT is necessarily needed in this case?

Also for annealing, I’m a little concerned like I said in my previous comment that producing dimensionally accurate parts that are unwarped will be difficult after annealing, but perhaps I’m misinformed?

Yes, glass filled would be another composite alternative, though it’s mechanical properties would be different. Example, increased weight and less rigid.

I’m rather interested in the specialized filaments like fire retardant or some of the ESD safe stuff. Could see use cases where you can mix these filaments to handle certain characteristics of the build.

Unfortunately, this is one of those important details that the reviewers seem to gloss over. I’ll be trying it myself sometime in the next 30 days, but until then I just don’t know for sure. It would be nice if I could just anneal it in situ while it is still stuck down on the build plate immediately after printing it, but so far I haven’t read of anyone even attempting that, so maybe I’m being overly hopeful. I would much prefer not to have to pack it in salt or plaster of Paris like CNC Kitchen did in his annealing experiments. His results looked rather crude, and i f it drifts in that direction, then I’m doubtful it will stay accurate. On the other hand, he wasn’t working with PET-CF, so I’m not sure how much of what he did would even translate into annealing PET-CF.

I haven’t used any of Bambus PC. Because all there fillament is low quality compared to what I use.

Now the Bambu lab PC. 99% PC 1% addative. So it’s a blend. They don’t say it’s a blend but if your not a hobbyst it’s clear that it’s a blend based off the temps. Hobbyst just see PC and hit order. But they do publish what’s it’s make up is. More then likely the addative is as usual petg to help fight against warping. 1% does help with that but there is a noticable impact in performance if you are looking for the PC propertys.

Other companies don’t even put out that info. They say it’s PC but the temps are a clear give away that it’s a blend. Like prlin. Nice PC blend that I don’t see how it could ever warp with the amount of PETG they put in and how much it sticks to pei. But they market it as PC and not a blend. The super hard PC is just more PC added to the blend.

They do this so hobbyist can say they use these engineering materials and they print great. Well it’s not really what you think it is. So you gotta go look for pure stuff. PC, PA etc some manufacturers are making it a blend and not saying anything so the guy with his ender 3 can use it and say he prints with it.

As much as I hate Bambus low quality fillaments they are honest…as much as Bambu lab is going to get. They put the info out there that it’s a blend but you gotta find it your self. So pretty honest for Bambu lab.

Now don’t get me wrong. Blends are fine as long as people know they are getting a blend. But most of the time they don’t. They want PC and wounder why it’s not acting like PC. But some times you want something with some resistance to heat, great first layer adhesion and a bit less brittle and blends are great for that. We were need to start calling out manufacturers for this.

Now a blend will work fine for what you want. But pure pc would be better. But pure PC can be a pain to print with if you never ran it. But it’s not that hard. Honestly no filament other than Delrin is hard to print. And even the. Delrin is not horrible it just requires a process and if you go too high on the temp it releases fromaldehyde gas. When using engineering grade stuff just ask an engineer how they run it and any tips they would give for your setup. Use that as a starting point. The whole people are having issues with this type of fillament so let’s make it a blend and not say anything is having a negative impact. Wrong materials for the job some times and the worst it’s a bandaid fix to get pass a print issue when the issue it’s self should be fixed. It’s holding the skill development of people up.

1 Like

I’ve heard a lot about the toxicty of PET v. PETG. I would be very cautious with PET if I were you. Very harmful vapors. The G is your friend. Still, ASA is your winner.

I am sure there is a national or international standard for lithum battery safety. You should probably find it. This must be especially true for e-bikes that are allowed to be taken on trains, bus etc.

I recently picked up a toaster oven from a garage sale that I plan on installing a PID controller in to attempt to start testing out annealing parts. CNC kitchens annealed PLA really opened my eyes to what might be possible. The best process I’ve seen is using extremely fine ground salt as the medium. It reduces the rough texture and makes the part look more like a matte finish.

Most instructions are to just heat it in an oven or furnace with supports attached to help reduce drooping. Typically start at a low temp, bring up to a medium temp, and then up to final annealing temperature. Bring temperature down in a similar approach. No matter how you anneal a part, you’d need to figure out how much shrinkage to expect so you can start at a larger size to end up at the right dimensions.