Hey, using the STEP files linked in this thread I went to order a metal replacement back for my librem 5 from some fabricator online, and they didn’t want to do it and said that the tolerances required to create the STEP file as given were most likely too fine for the end result to be useful based on what I was asking them to do.
I went and looked at the part they were concerned about and it does not physically match the backs that my Librem 5’s actually came with. There is a finely shaped hole in the part in the STEP files whereas the real backs from my real Librem 5’s are simpler and more solid.
Is there a STEP file that more closely matches the physical backs that ship with the Librem 5’s? I’ll attach a picture.
I was extremely intellectually lazy and asked chat gpt what services were good for solving the problem. This is not an endorsement of ChatGPT; I know there are multiple articles online that describe how it was created by paying Kenyans between $1.32 and $2.0 to read sexually abusive material and click “Yes” or “No” based on how abuse it was, essentially suggesting that ChatGPT was literally created by abusing people in third world countries.
But somehow that’s all too easy for me to forget sometimes, and so I happened to use ChatGPT. However unfortunately, that is what I did.
It suggested something called Xometry which looked like a fairly legit looking site, but I guess I’ll find out. Their human who contacted me that making this thing with metal was not so good of an idea suggested either SLS or HP MJF, but after some back and forth I told them to try HP MJF. I made it clear that it’ll be on me if it doesn’t have correct dimensions and doesn’t work. But either way, the guy said that SLS would have the same or similar issues, where the 3D printing tolerances are not as fine as what is in this STEP file from purism, because the file was probably created for the purposes of a mold.
[Edit: The interaction left me wanting to download some CAD software and make my own much simpler and less elegant STEP file of the back plate that could possibly be made of metal because I still think a metal-back phone would be cool. But when I looked at freecad on my Librem 14 briefly thus far I didn’t immediately figure out the interface, so I didn’t do that yet.]
Pricing was very similar for Xometry, but the main challenge for me when ignoring the money side was their human thinking that the parts would most likely not resolve well in the 3D printing, and that the resultant part would be even more warped if 3D printed with metal because the metal 3D prints go in an oven after being created, to finalize them, and some stuff like that.
Did you end up having any success in getting a metal backplate for your L5?
I did some “modifications” to my back plate long back, and just this morning found a chip in the plastic along the outer edge of the backplate.
They said that they could 3D print it with nylon but almost rejected the idea, based on the premise that required tolerances in the file are too small. But I told them I would just eat it and wanted to try it. So they added some note to the order that best effort was OK.
I’ll have more info when it arrives.
Edit:
Exact price on Xometry for Nylon 11 job created with HP MJF (and the “best effort” note) is costing me $59.24 supposedly.
My Nylon print from MakeXYZ was not great, it looks good as an SLS print, it just has some tolerance issues so it doesn’t seat to the back of the phone.
Hey! I’m writing this from L5 with Nylon 3D back; just arrived.
It fits well enough to click into place along most edges, but when I try to lift it off of the device it feels like it is falling off without effort as if it weren’t fully attached to begin with. But it’s really close – much better than I expected when I told them I wanted to just try it even if tolerances were bad.
But what I’m realizing now is that I didn’t read their materials data sheet extremely carefully – so, what I don’t know, is:
If I overheat this phone, can it melt the back now?
There is a particulate dust that came off of the back onto the device. When i remove the back, there is a “dusty” look left on the L5 from the 3D print part’s material surface losing some microscopic amount of material. Should I be worried this particle dust may be toxic, assuming I continue to carry this thing with me everywhere in my daily life?
As someone mentioned in another thread, I dont know if this back is “less RF transparent” than the original - so I do not know if it causes worse signal. But I was able to post this message with it on
I tried to take some pictures while holding by the original device chasis to highlight/test that the 3d printed portion wasnt falling off, because I think its friction fit seems weaker than on the original back.
Looks cool. How does the texture feel? Is the back thicker than the original?
That’s unlikely. Nylon is usually printed around 250’C [edit: this seems to have powder melting point of 202 °C/396 °F according to that linked sheet, still well above what would be normal for backplate]. Granted, if the high heat is long term and there is a thin section of material, there could be some effect, minor warping. But for melting, I’m pretty sure something else is really wrong at that point - like “dude, your phone’s on fire”
I think it feels fine. Almost sort of “fuzzy” but I’m less worried about that than about if it’s somehow toxic or if it rips when using it many times.
I tried attaching and detaching it from the back a whole bunch of times and I have a sense that the tightness of cling to the device is already reduced after just playing with it for 30m. So that might be the next biggest issue. It was already quite loose. The probability of it falling off on its own when I didn’t intend to take it off seems higher than ideal.
If it’s otherwise fine, and it’s only about the small tolerance of the clip, you might be able to fix it. First thing I’d try was to gently heat the edge to bend them inwards a fraction. And I do mean a fraction (both side 0.2-0.5mm should be on target). Maybe some hot flat surface (180-200’C, no open flame) that you can push against.
Ah, I was thinking more like something that spans the whole length of the side (to control that all clips bend the same - not sure if that’s really relevant but may prevent non-symmetry warping). Normal soldering iron tips are usually short (and round). Maybe, if you have a flat tip, you could do it just at the clips.
I had another Q, do you think that the mid-frame CAD files will be available at some point? In addition, the aluminum frame would be nice to see as a CAD file too. Hopefully the L5 gets to be fully open at some point soon.
Yes, I expect that it will become available, date TBD. This part is actually broken on my personal L5 (my device has survived numerous dramatic drops, but the internal frame has grown tired), and Purism strives for user-serviceability and longevity through easily-replaceable parts (replacement battery shipments being a little more nuanced).
So, I did not do this. My intuition based on the shape was that the very small inward facing almost invisible tabs were more important and were losing material from friction. So I put some tape over them in the size of where they fit into the main L5, and now the backplate seems to attach reasonably tightly like how I want.
As a result it seems pretty perfect now. Haven’t encountered much in the way of signal issue (RF transparency), it’s easier to remove than the standard plastic backplate because it is less tightly connected, and hopefully it’s stronger.
Holding nylon with the slight feel of softness instead of hard plastic is maybe elegant feeling in a certain way. Seems like maybe a fun style. Definitely makes it unique. If it starts to tear apart like the standard plastic backplates, I’ll be sure to post here.
Edit: Sent from mobile while using the 3D printed backplate.
Update: We received the supplier’s response for the original plastic L5 rear cover today and it sounds like they are charging a shockingly high price per unit. Unless there is significant interest for a group buy (200+) and folks are willing to pay $15/ea + shipping, we will likely reject the quote and search for an alternative supplier offering relatively low MOQs.
