Help/Info on custom keyboard case creation

[Me]

I want to create a custom keyboard case in either alu or plastic, and it def will have some sort of acrylic in it(for lighting purposes), but I have no experience with making cases. I am thinking about 3d printing it, but I am not sure. I am still in middle school so I only have very little experience with CAD, but I am a fast learner so with guidance I can learn pretty easily. I am mostly looking for info, but I will also be happy to collaborate on this with someone. Thanks, and anything will help, it doesn't have to be major.

[Leslieann]

There are some good 3d printed cases on Thinkgiverse.

You will want to used PLA, ABS shrinks too much and destroy your tolerances, PET doesn't stick to glue very well so it would really need to be done in one piece per layer. You might get a 60% to fit on a common printer but nothing larger. Also doing such large strong parts can take a loooong time and has a tendency to break print beds.

[nevin]

what size board were you planning to build?
are you planning to use a particular pcb?

...btw, love the user name. never thought that would still be available. nice!

[Me]

Thanks on all the info guys and thx about the username too. I am planning on building a 75% board and I have already ordered the pcb. It is the ymd75 pcb and I figured I would be able to get a case by the time it comes if I 3d print it, cuz its coming in about a month. I will check thingiverse, but I wanted to ask. How much would a 75% 3D printed case in PLA cost realistically?

[nevin]

do you have a printer or thinking of using a service?
- does your school have a maker space with a 3d printer?

some online 3d file sites have options to purchase prints through one of their preferred vendors
- there's also shapeways.com (nice but a bit on the pricey side)
- and many other places that will print your files

my suggestion would be to wait until you got the pcb so you could get measurements from the pcb and design a case around it. without having the pcb in hand there's no way to guarrantee what you design will fit. i don't know where the support posts are on this form factor. the support posts are pretty universal for 60% (most pcb makers follow the norm, but they can really put them wherever they like)

i did some quick searching and i wasn't able to find any usable resources for this particular pcb.

here's a KLE for one of the supported layouts
http://www.keyboard-layout-editor.com/#/gists/ef5601256afcf62a944640dcaa195faf

which you can copy/paste into either of the plate generators below to get you a rough size of what you'll be working with
http://builder.swillkb.com/
https://kbplate.ai03.com/

- you could also do a sandwitch style case like a lot of keeb.io's boards (very cost effective)
- ponoko.com has decent pricing for cutting all kinds of flat materials (like an acrylic middle layer for a sandwich style case...)

the ymd75 default case is a tray mount case (pcb screws to bottom of case via various standoffs)

did you get the solder version or the hotswap version? if you got the hotswap version you will want support under the pcb's hot swap sockets. bad things can happen when you push a switch into a socket on a pcb with very little support.

[Leslieann]

Thanks on all the info guys and thx about the username too. I am planning on building a 75% board and I have already ordered the pcb. It is the ymd75 pcb and I figured I would be able to get a case by the time it comes if I 3d print it, cuz its coming in about a month. I will check thingiverse, but I wanted to ask. How much would a 75% 3D printed case in PLA cost realistically?

The problem is size.
A TKL needs a printer with a bed of about 18inches while most printers are only 12inches, most printers of that size tends to be either custom built or commercial. Commercial 3d printing cost quite a lot more.

Question...
How much experience in CAD and 3d printing do you actually have? If you answer none to 3d printing expect 3-5 versions before you get it usable, if you have no CAD experience as well add another 3-5 versions.

As for cost, just a quick rough guess, about $150 each if you can find someone local with a custom built printer to do it for you. You have a basic setup fee and then plastic but on large prints like this, even the locals will want extra because the chance of failure on such a large print is very high. Plastic shrinks and when it does it causes lifting, a heated bed helps but heating an 18in bed is costly. A commercial shop is less likely to have issues getting it done but they will cost 2-3 times as much.

But you thought 3d printing was supposed to be cheap, that is cheap. To have this done in aluminum would cost 3-5 times as much and instead of waiting days would take weeks or months to have done at a shop.


One last note, there are online vendors that will give you an estimate, probably 2/3rds what I quoted. That will NOT be your actual price, I mean it can be however those online calculators are designed for small objects not really meant for heavy use only as working test models. They work on the premise of very thin walls and very low infill (my shop used to do work for one of them). When working at such larger scales or something meant to be used you need more than 10-15% infill and super thin walls. My own keyboard used 30% infill and 3 perimeters, even that was a compromise though it did work well. They will also tell you 3 day turnaround, when working at this scale that's almost impossible. Each segment will likely take 1-3 days in printing each and if you only have one printer and count on at least one failure (common at this scale) 3 days is just unreasonable. All of this was one reason we stopped doing work for these sorts of sites, it wasn't practical for larger prints which was what we specialized in. Personally, I probably wouldn't take the job for less than $200 because I know this will be a pain in the neck and would expect a 1-2 week turnaround.

[Me]

Multiple things. Things went wrong with the shipping and it turns out the pcb isn't coming after all. This wasn't necessarily a bad thing though, because while I was looking around for a different pcb I came across a technique called hand-wiring. I think that it will be tedious, and I will probably mess up a few times, but I think it will allow me a lot more customizability and affordability. Mainly the former. If everything goes according to plan(which it prob wont), then I will create a Cad file using builder.swillkb.com and KLE, edit it to make it be more 3d printer-friendly if I am going the way of the 3d printer, and then hopefully be able to either 3D print the case at my dad's work's 3D printer, or get a quote for a cutting service and get it made there. This will allow me to get dimensions right off the bat, and also have a custom layout which will be a little more suited to my needs. Is this a good plan?

Also, I am taking into account that it will be hard and will most likely involve some failures, but I like to challenge myself that way, so plz take that into account when replying.

[Leslieann]

Don't start with the KLE if you plan on 3d printing it, there's no sense in starting from scratch, check Thingiverse, there's plenty you can use to start from and eliminate 90% of the work involved.

[nevin]

thanks for dropping in with your experience Leslie.

i've helped a couple users here on the forum with with little or no experience with handwiring & firmware. i'd be willing to help you in those areas as well. for 3d files & printing advice Leslieann's your person. Leslie just finished this project (sorry, had to give you a plug, project was too good & lots of good info)

here's some resources & reading if you go the handwire route. yes, it's tedious, but there's a lot less that can go wrong with handwiring then trying to design a custom pcb.

hardwiring articles
https://geekhack.org/index.php?topic=87689.0
https://deskthority.net/viewtopic.php?t=1067

firmware
qmk on github  https://github.com/qmk/qmk_firmware
qmk docs  https://docs.qmk.fm/#/
kbfirmware  https://kbfirmware.com/ (some users i know have had some luck with this, good as a teaching tool to wrap your head around what's all in the firmware. based on an older version of qmk)

once you figure out what size/layout you want to do we'll have to figure out what parts you'll need to order other than the obvious switches & keycaps.
- micro controller - will depend on number of keys on board and how complicated you want to make the matrix
- diodes, trrs jacks if split board, usb breakout board
- wire, solder, tools (if needed)

if you're not sure on what you want your keyboard to be, KLE is a great place to quickly work out some ideas.
plus, i find the layouts helpful for wiring diagrams & pin notations for handwire projects
and, if you want to use VIA you'll need your layout in KLE anyway

[Me]

Thx for all the info guys. Ill def check out those hand wiring rescources you posted. Oh, I just looked at them and I have already seen the geekhack one, but thx anyway. Those handwiring jobs are clean, wow! For the firmware im def going qmk just cuz its more popular so there will be a lot more info about it(at least I think). Im still on the fence about the controller, because I'm not sure whether I should get the teensy, or another one. Also, before I found handwiring, I always thought that usb c was better, but many of the controllers don't have it. Maybe it just hasn't been standardized yet, but should my descision be based off of that? I'll leave a link to the layout that I made. It's basically 75% but without the nav cluster as I don't really use it and then it makes the board a little more compact. For the solder tools I already have that covered, and I think that 1N4148 diodes are best based off of what I've heard. I'm assuming that I could just use any 18 gauge or less wire, so that's not a problem. What are trrs jacks?

http://www.keyboard-layout-editor.com/#/gists/f770168380a3f7ed92078f26fd76880d

[Me]

Oh yeah, and the "M1" and "M2" in the layout are just 2 xtra keys that I had that I guess can be programmed to anything using qmk

[Me]

oh wait i just realized I have to find keycaps for that nooooooooooooooooooo

[Me]

nm I can edit the layout a tad and then a standard 104 keycap set(wish i went dsa) will have all the caps for it

[Me]

And one last thing. Another option would be to carve a housing out of wood. Does this sound/feel/look good?

[Leslieann]

It's basically 75% but without the nav cluster as I don't really use it and then it makes the board a little more compact. For the solder tools I already have that covered, and I think that 1N4148 diodes are best based off of what I've heard. I'm assuming that I could just use any 18 gauge or less wire, so that's not a problem.

It's a 60% with F-keys, using a modified layout.

oh wait i just realized I have to find keycaps for that nooooooooooooooooooo

WASD can make most of those, not sure about all of them but most of the rest, yes. So long as they make a key in that size for that particular row they can make it.
Just so you know I was really opposed to going without F-keys but so long as you have a function layer for them it's not as bad of a transition as I expected. Adjusting to the nav cluster being "built in" has been more difficult, not because it moved but because it's easy to grab the wrong key (left arrow instead of down seems to be my most common mistake).

Another option would be to carve a housing out of wood.

Wood is a great option, especially if you have no access to anything computer controlled and with this being so compact there's less worry about warpage over time.

Since you have access to a 3d printer, you actually could combine the two to make a very easy case, Use a thin plank for the bottom, glue on thin sides, making a nice little tub then print the standoffs that you need to lift the pcb off the wood (you need clearance for the pins). You could use flexible filament for the mounts for a bit of dampening or even engineer a perimeter gasket mount without having to do a ton of wood work. This also makes it easier to replace if you ever need to, after all, it's just a tub with a couple holes.

You could get by using all 1/4in or 3/8in wood and just enough for each section, meaning you could go more exotic, you could even shave costs further by using common woods for the bottom since it's not visible while sitting on the desk. I'd personally consider Bamboo for the sides, it's extremely durable, resistant to stains, nice looking and not super expensive. Unless you can find old growth Pine it will be a poor choice for the sides, it's too soft and it soaks up stains. If you want to make this really easy, go look in the trim section at your local hardware store, you can probably get Oak or Cedar in the thickness you need, meaning you're need to do very little cutting. I;d still prefer Bamboo or something even more exotic but these would be an easy way to go if your fabrication ability is limited by skill or equipment.

[Me]

Thx for all the info. I prob will do what you said, but I have a couple of Q's. Number one, I was thinking about getting a plate cut from a lasercutting service(plz recommend one if u can), and then create the other part of the housing using the technique you specified. Number two, I just don't fully understand it. Could you go over how to do what you said?(I like the idea of bamboo too). For that material of the plate I was thinking stainless steel since people seem to like it and it's less expensive than brass. Is that a good option or should I go with something different, like alu or polycarb?

[Me]

Also are u saying that I should go 60%?

[Me]

And one last thing. You talked about a pcb in your post, so I'm just clarifying. I am handwiring this kb.

[Leslieann]

Technically you crammed a 75% into a tall 60% case.
If you eliminate the F-keys you crammed a 65% into a 60%, this is why you have such odd key sizes.

Adding just one collum of keys means to the right you could use an off-the-shelf 65% or75% pcb and plate, and that will make your project sooooo much easier, as will finding keycaps. I believe laserboost does plate cuts but don't be afraid to call around town, a waterjet or laser cutting service can knock one out fast and surprisingly cheap. It's not like CNC where you have long waits for one-offs.

A plate means you can hand wire, but one way or another you need a plate or a pcb, this can be metal, a pcb or 3d printed plastic. People have hand cut plates but every one I've seen done the person always ends up saying it was not worth it. I considered using a die and press to do it, but by the time you get a press and die set done you could have just had one laser or water jetted.

On the case..
Most people start with a big chunk of wood then use a router to cut out the center, just like a solid aluminum case. Do not do this, it's wasteful and you end up with end grain on the sides and that rarely looks as good as the regular grain.  You want to use thinner strips and glue them, as shown below.

Zoom in, you can see this person glued thin strips to create walls, you never see any endgrain and you can use thin, cheap wood strips from your local home improvement store.


Image taken from here:
https://deskthority.net/viewtopic.php?t=5169

[Me]

Allright, thx, you've been a great help and now I know what I am going to  do with this kb! It will be my first custom so wish me luck!