[opensource] Toad | 70% keyboard

[TalkingTree]

Update: A new version, made from the ground up, has been completed. The thread about it is here.
Original post:


60% is my favourite form factor but sometimes a function row is handy and I often do key combinations that include F keys.
I don't like 75% keyboards because they lack the gap between function and numbers row and because all the keys are too close to each other.
Since 70% keyboard are so rare, let alone with a gap, I made one from scratch. This includes PCB, plate and case.
I'm very proud to show the results of this project because I used a lot of community resources in the process that it really feels like it's something made together.


Part 1: PCB
The PCB was the most challenging part because everything I came across with at university had nothing to do with electric current, capacitors and resistors, therefore my knowledge of electronics was very limited and I truly needed some help there.
That help came from the gentlemen of The Living PCB Design Thread bpiphany, pomk and vvp whose inexhaustible patience and accurate explanations actually made this possible. Should you guys ever read this, thank you from the bottom of my heart.



It's basically a 60% board with Poker mounting holes plus the function row along with two extra mounting holes.
For the sake of it, it features an universal layout, two LED indicators in a dedicated spot which are also traced to the Esc and Caps Lock switches for windowed caps.
The three large holes in the function-numbers gap were originally placed there to mount the PCB in a modified Filco TKL case but, since I had some measurements off, I decided not to mangle the case and to design one from scratch.

I know this PCB has a lot of design flaws, such as the USB connector on the top side which hits the plate, but I guess you live and learn.
Months ago I wouldn't ever pictured myself designing a PCB, so this project actually made me a novice from the wannabe I was. At the end of the day this PCB works and I'm happy with that.



Stabilizers, along with more expensive parts not used in this project, were provided for free by my buddy LeandreN.

I made a few revisions of this PCB, this is the very first one and it runs the QMK firmware. They were all produced by EasyEDA.


Part 2: Plate
I designed the plate with swill's superb tool which spared me the sorry task of measuring everything, making unavoidable errors and wasting a lot of money. You have my infinite gratitude.

The plate has 3 mm rounded edges and features MX+ALPS (_t:2) and Cherry stabilizers cutouts. I then added mouting holes with the custom polygons feature. The larger ones on the top right are for LEDs.

I had a batch of five plates cut at HEK's out of 1.5mm thick brushed 304 stainless steel and then I painted them matte black with a spray can.




Part 3: Case
Designing a case was a tough task because I didn't have my mind clear about a model since the beginning. I knew I wanted something simple, along the lines of a Filco Majestouch case but all the different options I went through didn't seem to satisfy me.

The initial idea was to cut and glue a Filco TKL case but eventually I decided against mutilating it.
Then I tried to make a case out of wood, but everything I came up with was so ugly that made me wanna baptize my project toad and no, I'm not showing.

I kept the name and finally I decided to have it printed.
I designed it with Tinkercad and, again, swill's builder helped me a lot in this process. I converted the case type to sandwich and imported the SVGs, so I only had to design some minor things like the case feet, PCB standoffs and USB cutout.
A flat case was the cheapest to print but I nonetheless wanted some angle, so I also printed some cone shaped feet which I screwed at the bottom half with salvaged screws from an old PC case, I think. I also added some rubber strips to prevent (too much) slippery.
Here's a comparison between printed feet and the aluminum one that I tried to copy.




I don't own a 3D printer so I submitted my order through 3D Hubs which was then made by 3DM Print with a Gimax3D printer in PLA at the resolution of 200 µm.


Part 4: Layout, switches and Keycaps
For the layout, I use a ANSI enter with split Left Shift and Backspace and a 7u bottom row. The characters encoding is the standard italian.
This KLE will better explain how my layout is organized.




Switches are Cherry MX Silent Blacks bought at sennin32's 'group buy

Most keycaps are OG Cherry italian PBT dyesub from a G81-3000/SAI.
Backspace is from wodan's GMK HADapter & HADditionals kit.
ANSI Enter is a PBT dyesub and it's from Shadovved's EnjoyPBT Dyesub/Blank Keysets.
R1 Delete key is from the 1800 kit of xiaodian317's GMK Classic Retro & GMK GEEKHACK Pack.
R1 §ů was dyesubbed by imsto. It's slightly lighter in colour than the other keys but I can live with that.


All the pictures shown in this post and a few others are available at this imgur link.

Thanks for reading.

Update 1: Source
I decided to make the project opensource. You can access the repository on EasyEDA here.
For the firmware please check the XMMX project as they share the same code.

Plate files are here.  toad_universal.zip (28.16 kB - downloaded 572 times.)

[Tally810]

Very cool project dude.  The case came out nice imo.

Sent from my SM-G928V using Tapatalk

[xondat]

Lovely. Needs an aluminum case though )

[_haru]

This is amazing! Layout is pretty cool too

[Leslieann]

Awesome work!

A lower layer height, 0.1mm as opposed to the 0.2mm you used on the feet would have come out far far better, though, they would have cost a bit more as it would have doubled the print time. Lower layers are more important on angled and curved surfaces as it hides the stair stepping effect a printer leaves when doing them. If you want to fix them (or clean up the case), you can lightly sand the plastic, then spray with sandable primer, sand, prime, repeat. After just a couple coats it should be a nice smooth case that looks like something you would buy off a store shelf. Thicker scratches you can use Bondo scratch filler.

0.2mm, while many consider it okay, or even decent, quite a few companies, including mine, consider this to be an extremely rough print. For something we intend to show off we never go thicker than 0.125 and if we have the time, we will go down to 0.1, 0.08, 0.05 or even less. While we have done 0.03 and 0.01 and it looked fantastic, the time required was simply unrealistic. For example a print that took 1 hour at 0.2 will take 200 hours at 0.01, when you consider we often do 36 hour prints at 0.125, a that same print would require 4500 hours, or about 6 months of non stop printing. Most home 3d printers will wear out before it reaches that point.

[Leslieann]

Lovely. Needs an aluminum case though )

This is what you do BEFORE sanding it out for milling, if he ultimately even wants that. Plastic does have it's benefits (like not shocking the hell out of you with static buildup!).

This case probably cost around $100 (a very rough estimate just for showing why), and had it not worked, he would have been out another $100 every time he made a change. Whereas, with aluminum, he would very likely still be waiting for the first one and would have cost him $500-$1000 per try. 3d printing isn't great at all things, but cheap, fast prototypes is where they shine.


Keep in mind, fast and cheap are relative, faster and cheaper isn't necessarily fast or cheap. While a cnc milling machine can make this faster than a 3d printer, you can wait weeks or months before a shop will tackle your small one off job. Same with price, paying hundreds for a few dollars worth of plastic seems like a lot, but it's a faction compared to what a machine shop will charge.

[Sifo]

this is the layout i've wanted for 6 years but was too lazy to make. good job

[ygor]

[TalkingTree]

Thanks everyone for their kind words, I do appreciate that.

Lovely. Needs an aluminum case though )

Aluminum case is surely something I'm considering but plastics is great for prototyping. I'll probably do some more PCB revisions and case prints before thinking to move to aluminum. The true benefit of plastics is that you can file or sand excesses wheter it's needed. Surely it can be done with aluminum as well but I question the results.

If you want to fix them (or clean up the case), you can lightly sand the plastic, then spray with sandable primer, sand, prime, repeat. After just a couple coats it should be a nice smooth case that looks like something you would buy off a store shelf. Thicker scratches you can use Bondo scratch filler.

That's precious, thanks, I'll try that.

0.2mm, while many consider it okay, or even decent, quite a few companies, including mine, consider this to be an extremely rough print. For something we intend to show off we never go thicker than 0.125 and if we have the time, we will go down to 0.1, 0.08, 0.05 or even less. While we have done 0.03 and 0.01 and it looked fantastic, the time required was simply unrealistic. For example a print that took 1 hour at 0.2 will take 200 hours at 0.01, when you consider we often do 36 hour prints at 0.125, a that same print would require 4500 hours, or about 6 months of non stop printing. Most home 3d printers will wear out before it reaches that point.

I've read a bit about 3D printing and I was aware that 200 microns would have been rough, especially with PLA.
Does printing ABS gives any better look in your opinion?

This case probably cost around $100

€ 63.61 shipped, roughly $73.

[Leslieann]

I've read a bit about 3D printing and I was aware that 200 microns would have been rough, especially with PLA.
Does printing ABS gives any better look in your opinion?

Abs gives no better finish, however it's easier to smooth.
You can use what's called an acetone dip to smooth it chemically and it it doesn't soften as easily as PLA, which if you even sand it hard will worm up enough to soften and become gooey. Keep this in mind if you sand your case, start on the bottom, it's kind of tricky, but not difficult.

You can use a lighter to flame polish ABS and PLA (and others), we do this a lot on rough areas or to gloss an area after sanding, however, you need to use a lighter, which means it doesn't work well on large surfaces. Using a torch or something will simply melt the whole thing.

There is also a new material similar to PLA which can be polished chemically. Chemically polished prints look better than injection molded as there is no seam.

So why not just use ABS?
Because ABS is a F***KING terrible plastic to 3d print. Yes, technically it's stronger, which is the big claim people using it use as a reason, the problem is, that is based on molded strength, not 3d printed strength. ABS doesn't adhere to the previous layer well, so it tends to delaminate, so while yes, it;s technically stronger, if done properly, when done wrong, you can literally crush it in your hands. And it's VERY easy to get wrong, opening a door in your house can cause enough of a draft to ruin the print. So you end up with people building cardboard shields, plexiglass cases, etc.. trying to keep air from blowing across it. You also need a heated bed because otherwise ABS has a tendency to warp because it shrinks, which also needs to be accounted for in the design (up to 8%) and the shrinkage can cause all sorts of odd print problems such as an hourglass shape. ABS is actually falling out of favor for various high temp PLAs, nylons and such at this point for these reasons. Not to mention the toxic smell it produces when printing.

I learned on ABS, which made my learning curve WAY steeper, but when I did finally switch to PLA, it was child's play. There are some machines that do ABS well, but it's rare to really see one that not just looks good, but actually has proper strength and no warping or distortion, something is almost always bad. and those that do, tend to not do PLA well. Everything in 3d printing has tradeoffs and pretty much no machine does it all, which is precisely why there are so many styles and types of printers.

€ 63.61 shipped, roughly $73.

Pretty much what I expected (I figured $60-70 plus shipping), but I rounded up just to make things easier and not give people wrong expectations.

Price is heavily dependent on who does the work, the amount of infill, type of plastic, size of printer required... While a stretch, $100 isn't really out of bounds.

[TalkingTree]

So why not just use ABS?
Because ABS is a F***KING terrible plastic to 3d print. Yes, technically it's stronger, which is the big claim people using it use as a reason, the problem is, that is based on molded strength, not 3d printed strength. ABS doesn't adhere to the previous layer well, so it tends to delaminate, so while yes, it;s technically stronger, if done properly, when done wrong, you can literally crush it in your hands. And it's VERY easy to get wrong, opening a door in your house can cause enough of a draft to ruin the print. So you end up with people building cardboard shields, plexiglass cases, etc.. trying to keep air from blowing across it. You also need a heated bed because otherwise ABS has a tendency to warp because it shrinks, which also needs to be accounted for in the design (up to 8%) and the shrinkage can cause all sorts of odd print problems such as an hourglass shape. ABS is actually falling out of favor for various high temp PLAs, nylons and such at this point for these reasons. Not to mention the toxic smell it produces when printing.

Your knowledge of 3D printing is impressive and I thank you for sharing it. Since I'm not printing myself I guess I'll keep have my prototypes printed in PLA.
Again, thank you.

[Leslieann]

Your knowledge of 3D printing is impressive and I thank you for sharing it. Since I'm not printing myself I guess I'll keep have my prototypes printed in PLA.
Again, thank you.

Thanks and you're welcome, I've been doing it for a few years now.
PLA is fine (and better for the environment), don't let people tell you it's bad. So long as you don't leave it in your car on a 100F degree day or out in the weather it will last a long time. Most people trashing it are just quoting the same things they have heard parroted by other people.

Be careful chasing perfection, there's almost always a way to make something better, get it to where you are happy with it and enjoy it.

[InvidiousIgnoramus]

Now this is a form factor I could get behind.

Endgame is a myth

[TalkingTree]

Because of all the recent and unexpected interest for my Toad, I've decided to give a preview of what's going on.

I'm slowly progressing towards a second version of my Toad and I'll be showing the PCBs that I have just received.
Most notable differences from the first revisions are ALPS support and split spacebar. Tech-wise, I moved all the SMD components to the bottom side so to have access to them even with a plated in place.

[ygor]

That looks really good, man.

[InvidiousIgnoramus]

That is looking really slick now, haha

[wolf]

Excellent work and a great take on the 70% layout.

Congrats on being able to get the case printed.  I don't have that option here - the only place I found that does 3D printing can only make small items and I don't own, or have access to, a 3D printer.  I ended up getting my case cut out of aluminium - top panel, five "side" layers and the base - surprisingly inexpensively: NZ$148.52 or approx US$107.50

[TalkingTree]

the only place I found that does 3D printing can only make small items and I don't own, or have access to, a 3D printer.

Have you tried 3DHubs? It was resourceful to me. You can try and split your model with dovetails and glue it together.

[SamirD]

I just saw this and very, very clean board.  I'm not even a fan of compact designs but yours makes me want to try yours out.  And such outstanding results from a novice at electrical stuff makes me think I may have a chance to design my own board someday.

[Sifo]

such outstanding results from a novice at electrical stuff makes me think I may have a chance to design my own board someday.

that is also what i took away from this

[TalkingTree]

such outstanding results from a novice at electrical stuff makes me think I may have a chance to design my own board someday.

I would encourage everyone to give it a try. At the beginning I made errors that made the board unusable, it costed me money, but in the end I made it. One just has to try and not to give up.

[Marvellion]

This layout is amazing, really good thinking.

[SamirD]

I don't even know where I would begin as I'm so comfortable with an M that there currently is nothing better for me.  I keep searching though...

[Gumball]

That looks awesome! I want one 👀👀

[luismonterieart]

Wow really impressed with your work so far! 

[need]

Love how deep the keycaps sink in, and the simple design. It's definitely better than 90% of CNC cases out there that have way off proportions and meaningless shapes.

[TalkingTree]

That looks awesome! I want one 👀👀

Thanks. I'll be releasing the source as soon as I'm done with v2, so chances you can have one are good.

Wow really impressed with your work so far! 

Thank you mate.

Love how deep the keycaps sink in, and the simple design. It's definitely better than 90% of CNC cases out there that have way off proportions and meaningless shapes.

Thanks. The top half of the case is 10mm high so the keycaps sink slightly more than some other keyboards which, I reckon, are 7mm high.
I'm reworking the whole project and I'm trying to design an aluminum case whose smooth surface would hopefully make the entire board look much prettier.

[oli]

That looks awesome! I want one 👀👀

Thanks. I'll be releasing the source as soon as I'm done with v2, so chances you can have one are good.

Wow really impressed with your work so far! 

Thank you mate.

Love how deep the keycaps sink in, and the simple design. It's definitely better than 90% of CNC cases out there that have way off proportions and meaningless shapes.

Thanks. The top half of the case is 10mm high so the keycaps sink slightly more than some other keyboards which, I reckon, are 7mm high.
I'm reworking the whole project and I'm trying to design an aluminum case whose smooth surface would hopefully make the entire board look much prettier.

What are the chances you'll add ISO compatibility in plate and PCB in future revisions?

[TalkingTree]

What are the chances you'll add ISO compatibility in plate and PCB in future revisions?

PCB already supports ISO, even in rev. 1. As for plates, I'll be offering various templates but I won't be selling any myself.

[oli]

What are the chances you'll add ISO compatibility in plate and PCB in future revisions?

PCB already supports ISO, even in rev. 1. As for plates, I'll be offering various templates but I won't be selling any myself.

Can't wait! Not only do I agree with you in terms of the nice functionality a dedicated F-row adds, but it's also so damn aesthetically pleasing to look at.

[Unforgivable]

That looks awesome! I want one 👀👀

Thanks. I'll be releasing the source as soon as I'm done with v2, so chances you can have one are good.

Wow really impressed with your work so far! 

Thank you mate.

Love how deep the keycaps sink in, and the simple design. It's definitely better than 90% of CNC cases out there that have way off proportions and meaningless shapes.

Thanks. The top half of the case is 10mm high so the keycaps sink slightly more than some other keyboards which, I reckon, are 7mm high.
I'm reworking the whole project and I'm trying to design an aluminum case whose smooth surface would hopefully make the entire board look much prettier.

Open source?

You’re a saint, sir. We need more people like yourself in this world!

Thank you!

[Sifo]

That looks awesome! I want one 👀👀

Thanks. I'll be releasing the source as soon as I'm done with v2, so chances you can have one are good.

HYPE

[TalkingTree]

Open source?

Yes under a Creative Commons BY-SA 3.0 license.

[vtachkov]

I was planning on diving into custom keyboard design myself and what you've done here is really inspiring. I too know very little about digital circuits but I've already started reading through the living PCB design thread. From what I've read I wouldn't have guessed that you were a novice, considering how often I saw you responding to and helping others in that thread. Then again, I'll paraphrase a quote from one of my old college professors: The best way to learn something is to teach it to someone else.

Would you mind if I PM you down the road as I get ready to start working on my own projects?

[TalkingTree]

I wouldn't have guessed that you were a novice, considering how often I saw you responding to and helping others in that thread.

I've been googling a lot by myself and found many answers to my questions around the internet, yet the most invaluable help I've found was in this very community and mostly from the gentlemen cited in the very first post.

Would you mind if I PM you down the road as I get ready to start working on my own projects?

Be my guest.

[KetchyKech]

this form factor is amazing -- spaces between f keys are huge for me too -- hopefully we see this form factor pop up in future GBs 

[TalkingTree]

I truly wanted to release this project as opensource including the case as of now but I'm new to parametric modeling so it will realistically take me several more weeks to get something done.
For the time being, I made the PCB source available. Please have a look at the end of the original post.

[mike-y]

Thanks for keeping up with this and for the new links.  I've only barely gotten into KiCAD, and looking at your work on the pcb helps a lot with the understanding of it.

[TalkingTree]

Thanks for keeping up with this and for the new links.  I've only barely gotten into KiCAD, and looking at your work on the pcb helps a lot with the understanding of it.

I'm glad it's useful to someone. Good luck with your project.

[TalkingTree]

Speaking of the case, I'm aiming to simple shapes once again but this time in aluminum.
So far I can say I'm done with the plate and 90% done with the top (it's missing the threads on the bottom side).

The render is missing the bottom half so it might look funny.


Any input is greatly appreciated.

[xondat]

Personally would get rid of the horizontal spacebar hole as it adds flex to the whole bottom row etc. Would also round the corners/increase chamfer size but that's personal preference. LED holes look a little small, but I'm not educated on that topic so won't go further.

[emenelopee]

I dig it. Nice to see a 4:3 aspect among all the widescreen-shaped keyboards out there.

[_haru]

I dig it. Nice to see a 4:3 aspect among all the widescreen-shaped keyboards out there.

I need a 16:10 keyboard to match my 1600p monitor.

[TalkingTree]

Personally would get rid of the horizontal spacebar hole as it adds flex to the whole bottom row etc.

I have three mounting points there, do you think the bottom row would still flex?

Would also round the corners/increase chamfer size but that's personal preference.

The switch opening is on a 19*19 mm base and the fillets of the inside corners are 0.5 mm. I'm afraid a bigger fillet wouldn't leave room for the cap to move.

LED holes look a little small, but I'm not educated on that topic so won't go further.

The opening you see is to allow the light through, the LEDs themselves are hidden in a 7mm cylinder beneath that. That's a feature I've seen on many recent designs and I find it quite nice.


Thank a lot you for your input.

[Pwner]

Personally would get rid of the horizontal spacebar hole as it adds flex to the whole bottom row etc.

I have three mounting points there, do you think the bottom row would still flex?

Show Image

Would also round the corners/increase chamfer size but that's personal preference.

The switch opening is on a 19*19 mm base and the fillets of the inside corners are 0.5 mm. I'm afraid a bigger fillet wouldn't leave room for the cap to move.

Show Image

LED holes look a little small, but I'm not educated on that topic so won't go further.

The opening you see is to allow the light through, the LEDs themselves are hidden in a 7mm cylinder beneath that. That's a feature I've seen on many recent designs and I find it quite nice.

Show Image

Thank a lot you for your input.

You could do something like this for the spacebar:


Cleans it up and is a bit more sturdy.

.5mm radius is fine for the corners and the leds are also fine.  You could probably even get away with something like 5mm for the cylinders unless you are using larger leds of course.

[TalkingTree]

You could do something like this for the spacebar:

Show Image

Cleans it up and is a bit more sturdy.

That's been duly noted. Thank you.

.5mm radius is fine for the corners and the leds are also fine.  You could probably even get away with something like 5mm for the cylinders unless you are using larger leds of course.

Would a shorter radius change the light effect in your opinion? For instance I'm using either 3mm or 2x3x4mm LEDs.

[Pwner]

You could do something like this for the spacebar:

Show Image

Cleans it up and is a bit more sturdy.

That's been duly noted. Thank you.

.5mm radius is fine for the corners and the leds are also fine.  You could probably even get away with something like 5mm for the cylinders unless you are using larger leds of course.

Would a shorter radius change the light effect in your opinion? For instance I'm using either 3mm or 2x3x4mm LEDs.

Probably not by a big margin, but in my mind it should focus a bit more light out the led hole.  What size is the hole that the light will actually shine through?

[TalkingTree]

What size is the hole that the light will actually shine through?

1mm.

[TalkingTree]

Following xondat and Pwner's suggestions, I removed the horizontal spacebar opening in the plate for myself; it now looks like this.



The universal plate still features it as it's needed for some plate mounted stabilizers.

Files for the universal plate are in the OP.

[TalkingTree]