NOW OPEN SOURCE!!! swill's plate building tool [builder.swillkb.com]

[harlw]

Wow - this is great. Thanks OP! Bookmarked.

I just had a plate lasercut in wood and the holes turned out slightly too large so I'm excited to have the detailed info and tools in this thread.

slowclap.gif

[swill]

Wow - this is great. Thanks OP! Bookmarked.

I just had a plate lasercut in wood and the holes turned out slightly too large so I'm excited to have the detailed info and tools in this thread.

slowclap.gif

Be sure to account for the kerf of the tool you are using to get the correct sized cutouts. Check the advanced settings for more details. If you have questions, just ask and I will do my best to support your needs.

[phishy]

I'm so glad to still see this thread going strong.  I still havent programmed my teensy but my keyboard is completely built for the most part.  I need a few more spacers for the edges.  If anyone can help me out navigating through this teensy stuff I'd appreciate it.

[swill]

I'm so glad to still see this thread going strong.  I still havent programmed my teensy but my keyboard is completely built for the most part.  I need a few more spacers for the edges.  If anyone can help me out navigating through this teensy stuff I'd appreciate it.

Show Image

Looking good sir.  I am sure you will love that layout.  I love my FC660C...

[BigBlueSaw]

I updated the material list based on builder suggestions and added keyboards to the main example page:
http://www.bigbluesaw.com/big-blue-saw/big-blue-saw-special-info/waterjet-and-laser-cut-keyboard-plates.html

http://www.bigbluesaw.com/examples

[Zustiur]

If anyone can help me out navigating through this teensy stuff I'd appreciate it.

I can't help much from my phone but if you tell me your matrix I can give you pointers on using the TMK firmware when I get home.

[edit]
I got most of my firmware knowledge from http://deskthority.net/workshop-f7/how-to-build-your-very-own-keyboard-firmware-t7177.html
If you follow that guide, note the following:
The guide forgets to tell you to download 'programmer's notepad' which you'll use to edit the code.
I never got winAVR to work properly. You may be better off downloading cygwin
The LED code in the guide may be wrong, or it may just be that I was using the LED on the teensy rather than a separate LED.

[nephiel]

The LED code in the guide may be wrong, or it may just be that I was using the LED on the teensy rather than a separate LED.

For the built-in LED on the Teensy I use something like this:

Code: [Select]

if (usb_led & (1<<USB_LED_CAPS_LOCK)) {
        // set pin as output
        DDRD |= (1<<6);
        // Teensy 2.0 built-in LED is active high
        PORTD |= (1<<6);
} else {
        // Hi-Z
        DDRD &= ~(1<<6);
        PORTD &= ~(1<<6);
}

[skullydazed]

I finally got around to setting up virtualbox and an ubuntu VM so I could run the python version. I spent way too much time trying to figure out how to run python code that imports freecad without running it in the freecad python window, and I just don't think it's possible on OS X.

The first modification I made was to add ALPS-only cutouts:

Code: [Select]

+        elif t == 4: # alps compatible switch, not MX compatible
+            points = [
+                (7.75-k,-6.4+k), (7.75-k,6.4-k),
+                (-7.75+k,6.4-k), (-7.75+k,-6.4+k),
+                (7.75-k,-6.4+k),
+            ]

This lets me create an alps+costar plate that supports multiple layouts, like this:



I'm not 100% certain that will work, but I'm certain enough that I ordered a plate from BBS to validate my change.

[swill]

I finally got around to setting up virtualbox and an ubuntu VM so I could run the python version. I spent way too much time trying to figure out how to run python code that imports freecad without running it in the freecad python window, and I just don't think it's possible on OS X.

The first modification I made was to add ALPS-only cutouts:

Code: [Select]

+        elif t == 4: # alps compatible switch, not MX compatible
+            points = [
+                (7.75-k,-6.4+k), (7.75-k,6.4-k),
+                (-7.75+k,6.4-k), (-7.75+k,-6.4+k),
+                (7.75-k,-6.4+k),
+            ]

This lets me create an alps+costar plate that supports multiple layouts, like this:

Show Image

I'm not 100% certain that will work, but I'm certain enough that I ordered a plate from BBS to validate my change.

I am glad that it was not just me that could not get that to work on OSX.  I spent at least a solid week trying before I gave up and said '**** it' and installed ubuntu and ran it there.  I was trying to make that code run on a server anyway, so it was not the end of the world for me to use Ubuntu.

Awesome.  If you want to do a pull request I can merge your changes into my repo.

Just in case others want to do alps only cutouts, it can be done with the current tool, it is just not super obvious.  Here is some documentation of the feature that I use to make alps only cutouts.

https://geekhack.org/index.php?topic=65189.msg1685418#msg1685418

You can use that to validate your cutout calculations if you want.  I can do the math later to verify the values will work.

Love to see that you started playing around with the python version.  Thanks for posting back.

I have not had a chance to work on this recently because I have been swamped with real life, kiteaid.com and summer...  :)

[neverused]

Swill, did I miss your brass build log? I remember you getting a plate but can't remember seeing it completed...

[swill]

Swill, did I miss your brass build log? I remember you getting a plate but can't remember seeing it completed...

I have not finished it because I have been too busy. Need to work on it.

[neverused]

Yeah parenthood can do that, good luck finding time! I've got a second one due anytime now and am trying rush building another board before he shows up.

[swill]

Yeah parenthood can do that, good luck finding time! I've got a second one due anytime now and am trying rush building another board before he shows up.

Haha, ya. A baby, new house, renos, it all takes time from the important stuff. 

[neverused]

They grow up faster than you realize though, soon you'll sirens your time chasing kids. Great work with the tool, I got to see your build someday soon!

[skullydazed]

I am glad that it was not just me that could not get that to work on OSX.  I spent at least a solid week trying before I gave up and said '**** it' and installed ubuntu and ran it there.  I was trying to make that code run on a server anyway, so it was not the end of the world for me to use Ubuntu.

Personally, I always like my stuff to be platform agnostic, but yeah it's not the end of the world. Speaking of server stuff I haven't actually dug in, but is there any reason it runs as root besides binding to port 80? I know it doesn't matter too much for something on a VM but if I can strip it down to a regular user we can do fun things like spinning up multiple copies of freecad to make multi-user possible on a single system.

Awesome.  If you want to do a pull request I can merge your changes into my repo.

Just gotta polish up a few things and I will do that.

Just in case others want to do alps only cutouts, it can be done with the current tool, it is just not super obvious.  Here is some documentation of the feature that I use to make alps only cutouts.

https://geekhack.org/index.php?topic=65189.msg1685418#msg1685418

You can use that to validate your cutout calculations if you want.  I can do the math later to verify the values will work.

Nice, looks like mine are exactly .1mm smaller than yours. I based mine on the matias datasheet so I assume you've padded it just a little to allow for non-perfect cuts?

I have not had a chance to work on this recently because I have been swamped with real life, kiteaid.com and summer... 

Haha, ya. A baby, new house, renos, it all takes time from the important stuff. 

Priorities, man!

They grow up faster than you realize though, soon you'll sirens your time chasing kids. Great work with the tool, I got to see your build someday soon!

My own little one is 10 months old now and can crawl across the room in the time it takes me to glance at the dog and back again. As much as it's already become the new normal and I have a hard time remembering what life was like before her, I already miss the days where I could just sit with her on my lap and have her actually stay there.

[swill]

I am glad that it was not just me that could not get that to work on OSX.  I spent at least a solid week trying before I gave up and said '**** it' and installed ubuntu and ran it there.  I was trying to make that code run on a server anyway, so it was not the end of the world for me to use Ubuntu.

Personally, I always like my stuff to be platform agnostic, but yeah it's not the end of the world. Speaking of server stuff I haven't actually dug in, but is there any reason it runs as root besides binding to port 80? I know it doesn't matter too much for something on a VM but if I can strip it down to a regular user we can do fun things like spinning up multiple copies of freecad to make multi-user possible on a single system.

Ya me too.  I have to admit, that was really the only reason I didn't launch that code earlier.  I spent a crap ton of time trying to make it work natively on OSX.  Supper annoying...  The reason it has to be run as root is only because it is running on port 80.  You can run it on port 8080 and a regular user should have the permissions to create a webserver on that port.  Port 80 is a bit special and requires root (or sudo) to listen on.  I was launching the application in my /etc/rc.local file which automatically runs as root, so that didn't matter for me.

You won't be able to make it run multiple copies of FreeCAD on the same box.  FreeCAD is pretty annoying in that it uses global variables which will cross between instances of FreeCAD.  So you will never be confident that what you are drawing is not being corrupted by another user's instance.  I solved this by running multiple instances of the server in LXC containers (you could use Docker if you wanted) and served the multiple users requests from different containers.  I built my cluster of containers to automatically scale based on the amount of traffic.  In the end Python was just a little too slow for my liking doing the actual draws, so I went looking for another language, which is why I rewrote it in Golang. 

Awesome.  If you want to do a pull request I can merge your changes into my repo.

Just gotta polish up a few things and I will do that.

No rush.  I have been super busy recently, so I am probably going to be a bit slow merging it. 

Just in case others want to do alps only cutouts, it can be done with the current tool, it is just not super obvious.  Here is some documentation of the feature that I use to make alps only cutouts.

https://geekhack.org/index.php?topic=65189.msg1685418#msg1685418

You can use that to validate your cutout calculations if you want.  I can do the math later to verify the values will work.

Nice, looks like mine are exactly .1mm smaller than yours. I based mine on the matias datasheet so I assume you've padded it just a little to allow for non-perfect cuts?

My measurements came from jdcarpe.  He has experience with such things, so I trust him.    Also, I have found that the tooling usually ends up cutting a little smaller than what you specify if they adjust for kerf (i think it is due to the tool speed).  I think my measurements are a little tight from the single test I have done, but it should be functional.  We will have to test to be sure...

I will add Alps only as a switch type to the current tool as well.  Also, I am planning to add Alps only stabs, but I have to confirm the correct spec.

I have not had a chance to work on this recently because I have been swamped with real life, kiteaid.com and summer... 

Haha, ya. A baby, new house, renos, it all takes time from the important stuff. 

Priorities, man!

The baby is sleeping right now and I have a bit of time and I was considering pulling out my soldering iron and getting a session in, but I am knackered.  I went to a wedding last night and I am still a bit hungover.

They grow up faster than you realize though, soon you'll sirens your time chasing kids. Great work with the tool, I got to see your build someday soon!

My own little one is 10 months old now and can crawl across the room in the time it takes me to glance at the dog and back again. As much as it's already become the new normal and I have a hard time remembering what life was like before her, I already miss the days where I could just sit with her on my lap and have her actually stay there.

Ya, it definitely changes things.  Mine is now old enough that we can go into the yard and kick a ball around, which is a lot of fun.  He only just turned 2, but he is pretty coordinated so we can go have fun getting energy out outside now. 

[tinimi]

Hi all.

I'm currently building my first custom keyboard. It's plate mounted (hand wiring) TKL board. I've generated template from http://builder.swillkb.com/
Today I've got plate from laser cut. And found problems with stabilizers

I've bought stabilizers from WASD keyboards. First of all, looks like my stab holes are too long - stabs fixed not very well. I would need glue to fix them properly.
But another problem is space-bar. On backspace everything is fine. But on space-bar curved stablizer wire part hits middle of keycup! I don't understand why. Cut-outs for backspace and space are the same. They must be different? Looks like they must be a bit above current position.

Could you help me, how to set up them correctly?

My drawing (dxf) is in attachment, and images with dimensions (i've remove another keys - they are ok).

Looks like stab cutout must be 14.0(or less!) in height (instead of 14.2).
And aligned to the switch cutout. Without 0.5499 gap at top and 0.751 at bottom.

[swill]

Hi all.

I'm currently building my first custom keyboard. It's plate mounted (hand wiring) TKL board. I've generated template from http://builder.swillkb.com/
Today I've got plate from laser cut. And found problems with stabilizers

I've bought stabilizers from WASD keyboards. First of all, looks like my stab holes are too long - stabs fixed not very well. I would need glue to fix them properly.
But another problem is space-bar. On backspace everything is fine. But on space-bar curved stablizer wire part hits middle of keycup! I don't understand why. Cut-outs for backspace and space are the same. They must be different? Looks like they must be a bit above current position.

Could you help me, how to set up them correctly?

My drawing (dxf) is in attachment, and images with dimensions (i've remove another keys - they are ok).

Looks like stab cutout must be 14.0(or less!) in height (instead of 14.2).
And aligned to the switch cutout. Without 0.5499 gap at top and 0.751 at bottom.

Where did you get the plate cut?  Did they adjust for kerf or did you?

What thickness plate did you get cut, 1.5mm or 1.6mm?  The costar stabilizer holes are 14.2 because with 1.6mm plates they don't clip in perfectly and it causes the uprights to be pushed in a bit which causes rubbing.  All the stabilizer cutouts are identical (as you can see from your measurements), so the spacebar should not be any different than any of the 2u stabs.  What hits?  The wire on the switch?  You said the wire hits the keycap?  Im not sure what is hitting here.

With a little more info I will try to help...

[tinimi]

Plate is 1.5mm
I did no kerf adjustments. And I'm not sure, if cutting service do it. Laser cut has very low kerf.
Middle of the wire (it is curved, from wasd) hits key cup.
http://www.wasdkeyboards.com/index.php/products/keyboard-parts/cherry-mx-stabilizer-spring-for-spacebar.html

[KHAANNN]

What do you mean by keycup?

The wire hits some keycaps tho that's expected: https://geekhack.org/index.php?topic=73871

You can pretty much remove whatever the wire hits, even re-bend it slightly to change it's movement behaviour (I've never tried sanding/drilling the switches themselves, but I'm going to try it too on my next build)

[swill]

Plate is 1.5mm
I did no kerf adjustments. And I'm not sure, if cutting service do it. Laser cut has very low kerf.
Middle of the wire (it is curved, from wasd) hits key cup.
http://www.wasdkeyboards.com/index.php/products/keyboard-parts/cherry-mx-stabilizer-spring-for-spacebar.html

If you did not account for kerf and the cutter didn't either, then that is likely why the stabilizers are not fitting nicely.  Do you have a caliper so you can measure what the actual opening dimensions are?

As KHAANNN mentioned, some keycaps do have issues hitting the wire.  What keycaps are you using?

[tinimi]

As KHAANNN mentioned, some keycaps do have issues hitting the wire.  What keycaps are you using?

Tai-Hao PBT Double Shoots: https://www.keychatter.com/wp-content/uploads/2015/02/keychatter_2015-02-12_03-28-02.jpg

[swill]

As KHAANNN mentioned, some keycaps do have issues hitting the wire.  What keycaps are you using?

Tai-Hao PBT Double Shoots: https://www.keychatter.com/wp-content/uploads/2015/02/keychatter_2015-02-12_03-28-02.jpg

I have never tried those, so I am not sure. Are they cherry profile and reasonably thick caps?

[tinimi]

As KHAANNN mentioned, some keycaps do have issues hitting the wire.  What keycaps are you using?

Tai-Hao PBT Double Shoots: https://www.keychatter.com/wp-content/uploads/2015/02/keychatter_2015-02-12_03-28-02.jpg

I have never tried those, so I am not sure. Are they cherry profile and reasonably thick caps?

As reported - Cherry. But I'm not very professional to confirm.
They are thin. With thick keycaps that would be more problem - as on 2u keycaps there are very thin gap between keycap and wire.

[azhdar]

As KHAANNN mentioned, some keycaps do have issues hitting the wire.  What keycaps are you using?

Tai-Hao PBT Double Shoots: https://www.keychatter.com/wp-content/uploads/2015/02/keychatter_2015-02-12_03-28-02.jpg

I have never tried those, so I am not sure. Are they cherry profile and reasonably thick caps?

As reported - Cherry. But I'm not very professional to confirm.
They are thin. With thick keycaps that would be more problem - as on 2u keycaps there are very thin gap between keycap and wire.

it is OEM profile.

[tinimi]

14.2mm x 3.3mm is a cut-out size for 1.6mm plate. What size must be for 1.5, or 1mm plate? Without gaps.
And gap from top of the switch is 0.55mm, bottom 0.75mm. Are this correct? May be set to 0 ? and 14mm heights, like a switch hole.

[swill]

14.2mm x 3.3mm is a cut-out size for 1.6mm plate. What size must be for 1.5, or 1mm plate? Without gaps.
And gap from top of the switch is 0.55mm, bottom 0.75mm. Are this correct? May be set to 0 ? and 14mm heights, like a switch hole.

I am asking what the measured size of the opening is (by using a caliper).  If you did not specify a kerf value then your openings are probably 0.2mm (or more depending on the laser) bigger than what you see in the kad file, so it would be good to get an actual measurement of the opening.

Can you ask the laser cutter if they accounted for kerf and what the kerf size is for their laser?

[tinimi]

14.2mm x 3.3mm is a cut-out size for 1.6mm plate. What size must be for 1.5, or 1mm plate? Without gaps.
And gap from top of the switch is 0.55mm, bottom 0.75mm. Are this correct? May be set to 0 ? and 14mm heights, like a switch hole.

I am asking what the measured size of the opening is (by using a caliper).  If you did not specify a kerf value then your openings are probably 0.2mm (or more depending on the laser) bigger than what you see in the kad file, so it would be good to get an actual measurement of the opening.

Can you ask the laser cutter if they accounted for kerf and what the kerf size is for their laser?

I've just measured with calliper - exactly 14x14 for switch and 14.2x3.3 for stabilizer. May be 14.25 (my calliper can measure only by 0.1, but it looks like "a bit more").
But - stabilizer plate insert is just 13.7 long!

And - photo of wire attached. I'm not sure, that moving stab by 0.5 cat save this situation.

[tinimi]

I've found, that on both space bar, and 2u keys stabilizer insert (in keycap) rubs stabilizer insert in plate. In top part of it. So plate insert must be move upper. Looks like without 0.55mm gap at the top at all.

[KHAANNN]

I've found, that on both space bar, and 2u keys stabilizer insert (in keycap) rubs stabilizer insert in plate. In top part of it. So plate insert must be move upper. Looks like without 0.55mm gap at the top at all.

my initial findings were similar, however on all occasions I managed to make the current cutout work

If you move the cutouts upper, the wire might start touching the switch instead of the insert rubbing the insert-slider

[tinimi]

Just to clarify. Keycap insert is not symmetric. One side is short, anoter is long. Longer part upper? Short part to the wire?
(Hope, you understand me without paintings)

[KHAANNN]

Just to clarify. Keycap insert is not symmetric. One side is short, anoter is long. Longer part upper? Short part to the wire?
(Hope, you understand me without paintings)

True

[tinimi]

BTW, swill, you are doing great job by placing this site online. It's awesome tool. Easy, simple and powerful.

[tinimi]

I've done hand made straight wire (without curve in the middle) - and now it don't hit spacebar. But I need to move stabilizer mounts a bit upper - probably by 0.2mm.
Thanks for your help! Next time I will choose cherry plate stabs 

[KHAANNN]

I've done hand made straight wire (without curve in the middle) - and now it don't hit spacebar. But I need to move stabilizer mounts a bit upper - probably by 0.2mm.
Thanks for your help! Next time I will choose cherry plate stabs 

I also hand bend the wire, but I bend the curved one to make it clear better and have a better keypress experience, without the clumsiness

You've experienced and solved Costar issues now, wait till you experience Cherry ones, they might not be as easily solve-able as the Costar's, Costar stabs are also easy to replace, easy to re-work, that's now the case with the Cherry

The only advantage of Cherry seems to be the keycap clearance, which is a significant advantage, I will also retry Cherry stabs for my next keyboard, but 90% I will stick to Costar's

[StormyMonday]

Hi swill, I chose the "medium" cutout type with 12 edges. The cutouts were precisely  lasercutted into 1,5mm steel. My original MX switches can move freely from left to right for about ... ~0.6mm, I would estimate. They sit tight enough that this is not a real issue. I just ask, if I shall change the cutout design to 15.0mm*14.0mm for the next batch, ie. shrinking the horizontal cutout width by 0.6mm, what do you think?

Do you really need the side cutouts for switch opening? I VERY rarely use them, as I normally just desolder the switches if I want to replace them with another type.

I HIGHLY recommend using the simple square cutouts if you can go without the cutouts for switch opening. And I also recommend picking one type of stabilizer hole and sticking with it, whether it be Cherry or Costar.

The 15.6 x 12.8 Alps rectangle merged with the 14 x 14 MX square is a compromise, designed to allow for use of Alps switches or MX switches on the same plate. It has the added benefit of allowing for easy switch opening, by having the sides cut out for the tabs on the MX switch to open. But it has the disadvantage of introducing instability in the switch hole, if the cuts are not made to EXTREMELY tight tolerances. Like 0.03mm tolerance.

@swill, maybe make the square holes to be the default, rather than the "rectangle on square design." Really, even the "H" cutout would cause less problems, while being more expensive. This tool is designed for ease-of-use, not necessarily advanced options by default, in my mind.

You guys both make great points. I personally chose the square holes for my build for the reasons JD pointed out. I will make the H the default, that's a good call.

Thanks for the feedback Eszett, I hope you are happy with your build.

If I may be so bold, one of these days, when you have some spare time    please consider giving us some dedicated Alps mounting holes. That would make the drawings less complicated, the fabrication less costly and the need for precision less exacting.

[StormyMonday]

I have no clue to be honest (I asked for 1.5 mm steel), it was cheap as hell (~25$ for what you see) and it came greasy as **** so I wanted to clean it up a bit so I can try some switches. I already found someone who will powder coat it for another ~20$ but I've already tweaked my layout and I think I'm gonna discard this one.

This is revision  2 :

Show Image

The thickness of powdercoat can vary dramatically and it can also be quite thick. Usually, it is preferred to mask off any critical areas (screw threads and the like) prior to its application. I don't want to scare you away, just make you aware that if you start off with a 1.5mm plate, it won't take much film thickness for your components to no longer fit. I would suggest that you share these concerns with the person who will perform the application.

You might also visit your friendly neighborhood gunsmith. There are many speciality coatings that have been developed for firearms that are not only protective, but are also attractive. They also, necessarily, have a very thin film thickness, otherwise the mechanism of the firearm would bind.

[StormyMonday]

I have no clue to be honest (I asked for 1.5 mm steel), it was cheap as hell (~25$ for what you see) and it came greasy as **** so I wanted to clean it up a bit so I can try some switches. I already found someone who will powder coat it for another ~20$ but I've already tweaked my layout and I think I'm gonna discard this one.

This is revision  2 :

Show Image

By the way, I really think that you are onto something here. I've always admired the FC-660C/M from afar (I even have a basic design in the keyboard editor), but I was never sure that I could live with it. Your design may push me over the edge.

Thanks for sharing.  :thumb:

[StormyMonday]

I figure I might as well document this feature now that it may become relevant for people's prototyping.  I actually added this feature a month or so ago at Melvang's request.  The feature is basically the ability to change the size (and shape) of the square switch cutout.  NOTE: This functionality ONLY works on the square cutout shape, the settings will be ignored for all other shape cutouts.

I alluded to this functionality in this post when talking about building a simple and inexpensive lube station out of plastic.

I think this is relevant to the current conversation because it may work out to be the cheapest way to prototype hand wired cases.

So here is the idea.  You have a 1.5mm switch plate laser cut out of acrylic.  We all know that 1.5mm acrylic will never hold as the only support for a keyboard plate.  So we would have to add a supporting layer underneath the switch layer.  I would suggest using something like 3mm acrylic or maybe even something like wood.  Of course, in order to add this layer, the openings have to be bigger than the switch plate so the switches will clip into the 1.5mm switch plate.

To illustrate, the switch plate would be 1.5mm and would look like this:
(Attachment Link)

You would then have a support layer with 16mm square holes in, say, 3mm acrylic or wood which would go under the switch layer and would look like:
(Attachment Link)

When you put the switch layer on top of the support layer, it would look like this:
(Attachment Link)

I suspect we can get some relatively inexpensive prototype cases built this way.  I would be interested in doing it in 100% acrylic and putting my LEDs on the bottom of the switch in a hand wired setup.

Ok, so how does this feature work already?!?  Get to the interesting part...

So if you add a background color to your layout and then look at the raw data, you will see this new object at the top of the raw data:
(Attachment Link)

We are going to piggyback off that object and add our config into that object to change the shape of the square.  In this case we are going to grow the square by 2mm in both width and height resulting in a 16mm square hole, so we would put the following:

Code: [Select]

{backcolor:"#505050",grow_x:2,grow_y:2},
Now the interesting thing here is that you can grow in both a positive or negative direction.  So if you wanted to cut an Alps Only plate and keep it on the cheap, you could do the following:

Code: [Select]

{backcolor:"#505050",grow_x:1.6,grow_y:-1.1},
This would result in the following cutout:
(Attachment Link)

I think that is enough for this post.  If you have more questions or would like to discuss the feature more, fire away...

Sorry, it looks like it is possible to do an Alps only switch plate, after all.    I don't know how I missed this, but I can't wait to tinker with it.

EDIT:
My objective today was to generate a drawing for a plate, which is suitable for mounting Alps / Matias switches.

As instructed, I added a background color to a design in the keyboard-layout-editor. This did not, however, add a {backcolor: "#xxxxxx"}, line to the top of my raw data file. I went in circles for a few minutes trying to figure out what I was doing wrong. Finally I just copied and pasted {backcolor:"#505050",grow_x:1.6,grow_y:-1.1}, as a new first line for my raw data file. I imported this into the "Plate & Case Builder" tool and voilà, it spat out mounting holes for Alps switches.

Hopefully, this will save someone else a couple of minutes of head scratching.

I imported the resulting *.DXF file into SAGCAD, which is supposed to be a friendly CAD program, but since all of the documentation is in Japanese, I haven't found it to be all that friendly. I have managed to verify that the generated mounting holes are 15.6mm x 12.9mm, but I think that I'll give LibreCAD a go, next.

NOTE:
If anyone else decides to trod this path, be aware that the thickness for an Alps-compatible plate is only 1.0-1.2mm.
Source: http://deskthority.net/wiki/File:Forward_Electronics_--_spec_--_SKBMFA.pdf

Since Swill's tool generates a 2D drawing, it does not contain a thickness specification. Make sure that your fabricator does not assume that a 1.5mm plate is a one size fits all solution.

[Melvang]

I have no clue to be honest (I asked for 1.5 mm steel), it was cheap as hell (~25$ for what you see) and it came greasy as **** so I wanted to clean it up a bit so I can try some switches. I already found someone who will powder coat it for another ~20$ but I've already tweaked my layout and I think I'm gonna discard this one.

This is revision  2 :

Show Image

The thickness of powdercoat can vary dramatically and it can also be quite thick. Usually, it is preferred to mask off any critical areas (screw threads and the like) prior to its application. I don't want to scare you away, just make you aware that if you start off with a 1.5mm plate, it won't take much film thickness for your components to no longer fit. I would suggest that you share these concerns with the person who will perform the application.

You might also visit your friendly neighborhood gunsmith. There are many speciality coatings that have been developed for firearms that are not only protective, but are also attractive. They also, necessarily, have a very thin film thickness, otherwise the mechanism of the firearm would bind.

Cerokote is what you are looking for here.  Last time.e I checked, certain colors can be applied as thin as .006" thick.

[swill]

MAJOR UPDATE!!!

This update has been a LONG time in coming.  I had to do a major rewrite for how I handle the dimensions of the plates and how absolute positioning changes those dimensions.  Many pieces of the drawing code have been completely rewritten with this update.  Hopefully it all is worth it, because the builder now supports rotated clusters and absolute positioning of clusters.

So for example, here is a draw of the Ergodox layout:



Code:

More

[{x:3.5},"#\n3",{x:10.5},"*\n8"],
[{y:-0.875,x:2.5},"@\n2",{x:1},"$\n4",{x:8.5},"&\n7",{x:1},"(\n9"],
[{y:-0.875,x:5.5},"%\n5",{a:7},"",{x:4.5},"",{a:4},"^\n6"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"!\n1",{x:14.5},")\n0",{a:7,w:1.5},""],
[{y:-0.375,x:3.5,a:4},"E",{x:10.5},"I"],
[{y:-0.875,x:2.5},"W",{x:1},"R",{x:8.5},"U",{x:1},"O"],
[{y:-0.875,x:5.5},"T",{a:7,h:1.5},"",{x:4.5,h:1.5},"",{a:4},"Y"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"Q",{x:14.5},"P",{a:7,w:1.5},""],
[{y:-0.375,x:3.5,a:4},"D",{x:10.5},"K"],
[{y:-0.875,x:2.5},"S",{x:1},"F",{x:8.5},"J",{x:1},"L"],
[{y:-0.875,x:5.5},"G",{x:6.5},"H"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"A",{x:14.5},":\n;",{a:7,w:1.5},""],
[{y:-0.625,x:6.5,h:1.5},"",{x:4.5,h:1.5},""],
[{y:-0.75,x:3.5,a:4},"C",{x:10.5},"<\n,"],
[{y:-0.875,x:2.5},"X",{x:1},"V",{x:8.5},"M",{x:1},">\n."],
[{y:-0.875,x:5.5},"B",{x:6.5},"N"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"Z",{x:14.5},"?\n/",{a:7,w:1.5},""],
[{y:-0.375,x:3.5},"",{x:10.5},""],
[{y:-0.875,x:2.5},"",{x:1},"",{x:8.5},"",{x:1},""],
[{y:-0.75,x:0.5},"","",{x:14.5},"",""],
[{r:30,rx:6.5,ry:4.25,y:-1,x:1},"",""],
[{h:2,_rs:180},"",{h:2,_rs:180},"",""],
[{x:2},""],
[{r:-30,rx:13,y:-1,x:-3},"",""],
[{x:-3},"",{h:2},"",{h:2},""],
[{x:-3},""]

Or you can create each of the hands individually (code included below):

Left Hand:



Code:

More

[{x:3.5},"#\n3"],
[{y:-0.875,x:2.5},"@\n2",{x:1},"$\n4"],
[{y:-0.875,x:5.5},"%\n5",{a:7},""],
[{y:-0.875,a:7,w:1.5},"",{a:4},"!\n1"],
[{y:-0.375,x:3.5,a:4},"E"],
[{y:-0.875,x:2.5},"W",{x:1},"R"],
[{y:-0.875,x:5.5},"T",{a:7,h:1.5},""],
[{y:-0.875,a:7,w:1.5},"",{a:4},"Q"],
[{y:-0.375,x:3.5,a:4},"D"],
[{y:-0.875,x:2.5},"S",{x:1},"F"],
[{y:-0.875,x:5.5},"G"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"A"],
[{y:-0.625,x:6.5,h:1.5},""],
[{y:-0.75,x:3.5,a:4},"C"],
[{y:-0.875,x:2.5},"X",{x:1},"V"],
[{y:-0.875,x:5.5},"B"],
[{y:-0.875,a:7,w:1.5},"",{a:4},"Z"],
[{y:-0.375,x:3.5},""],
[{y:-0.875,x:2.5},"",{x:1},""],
[{y:-0.75,x:0.5},"",""],
[{r:30,rx:6.5,ry:4.25,y:-1,x:1},"",""],
[{h:2},"",{h:2},"",""],
[{x:2},""]

Right Hand:



Code:

More

[{x:5.1},"*\n8"],
[{y:-0.875,x:4.1},"&\n7",{x:1},"(\n9"],
[{y:-0.875,x:2.1},"",{a:4},"^\n6"],
[{y:-0.875,x:7.1},")\n0",{a:7,w:1.5},""],
[{y:-0.375,x:5.1},"I"],
[{y:-0.875,x:4.1},"U",{x:1},"O"],
[{y:-0.875,x:2.1,h:1.5},"",{a:4},"Y"],
[{y:-0.875,x:7.1},"P",{a:7,w:1.5},""],
[{y:-0.375,x:5.1},"K"],
[{y:-0.875,x:4.1},"J",{x:1},"L"],
[{y:-0.875,x:3.1},"H"],
[{y:-0.875,x:7.1},":\n;",{a:7,w:1.5},""],
[{y:-0.625,x:2.1,h:1.5},""],
[{y:-0.75,x:5.1},"<\n,"],
[{y:-0.875,x:4.1},"M",{x:1},">\n."],
[{y:-0.875,x:3.1},"N"],
[{y:-0.875,x:7.1},"?\n/",{a:7,w:1.5},""],
[{y:-0.375,x:5.1},""],
[{y:-0.875,x:4.1},"",{x:1},""],
[{y:-0.75,x:7.1},"",""],
[{r:-30,rx:3.1,ry:4.25,y:-1,x:-3},"",""],
[{x:-3},"",{h:2},"",{h:2},""],
[{x:-3},""]

With this change there is one sporadic issue that I still have to resolve, which I will work on over the next week or so.

Sometimes the stabilizers are not drawing correctly.  Basically, I currently have the switch cutouts and the stabilizer cutouts as different shapes and I union the shapes together based on the selected configuration.  Unfortunately the algorithm for the union has some issues and the union does not always work.  The algorithm is very complex already, so I have not tried to create a pull request for that module yet.  This problem happens mainly with the Ergodox layout because the angles seem to cause a higher rate of failure.  Sometimes changing the amount of padding for the plate changes the numbers enough that the union works, or flipping the stabilizer (eg: {_rs:180}) will make it draw correctly.  In short, double check your plates to make sure the stabilizers are drawn correctly for now.

Instead of doing the union, I will probably just manually draw out all permutations of possible switch + stabilizer cutouts as individual drawings and then just reference them based on the scenario.  This should resolve the problem and make the draw function less error prone.

I fixed a number of minor issues with this build, but this post is getting long, so I won't cover them in detail...

Since this is a major change to the code, please let me know if you find anything that you feel may be an issue.

[Zustiur]

Woohoo. Now I just need the pcb equivalent!

[swill]

Woohoo. Now I just need the pcb equivalent!

There are a few projects starting to work on that problem. I have been thinking about the problem and trying to see how I can add that to this tool, but it is not trivial. I will probably give it a shot in the near future, but don't expect anything for a while.

[Zustiur]

No pressure.
I did see that others had started doing that but when I actually wanted them I couldn't find any of them!

[swill]

No pressure.
I did see that others had started doing that but when I actually wanted them I couldn't find any of them!

This is probably the closest thing to a working solution so far.

Public site: http://kalerator.clueboard.co/
Github repo: https://github.com/skullydazed/kalerator

The main issue that you will probably have with this solution is that you need a licensed version of Eagle cad to be able to actually view a keyboard sized PCB.  It is definitely a good start though...

[skullydazed]

That's a pretty nice update! Great work!

I've been working on my own ventures, but I still need to send across my changes for alps cutouts. I got a sample back from big blue saw and once they adjusted the kerf for the 5000 series aluminum I specified it worked great. I also just got certified to use the laser cutter at my makerspace, so I'll be cutting out my own acrylic cases soon enough. 

I also have a little something else I'll be sending over with my pull request:

Code: [Select]

skully@builder:~/kb_builder$ ./kb_cli -f foo
FreeCAD 0.16, Libs: 0.16R5358 (Git)
Processing: foo
...
Writing STEP file......
Face...

*******************************************************************
******        Statistics on Transfer (Write)                 ******

*******************************************************************
******        Transfer Mode = 0  I.E.  As Is       ******
******        Transferring Shape, ShapeType = 0                      ******
** WorkSession : Sending all data
 Step File Name : /home/skully/kb_builder/static/exports/switch_foo.stp(71734 ents)  Write  Done
saving......
processing Shape (99.0 %)
successfully exported /home/skully/kb_builder/static/exports/switch_foo.dxf
Finished: foo
Processing took: 44.63 seconds
**** Overall plate size: 322.325x103.25 mm
*** Files exported for plate switch
static/exports/switch_foo.js
static/exports/switch_foo.brp
static/exports/switch_foo.stp
static/exports/switch_foo.stl
static/exports/switch_foo.dxf


[swill]

That's a pretty nice update! Great work!

I've been working on my own ventures, but I still need to send across my changes for alps cutouts. I got a sample back from big blue saw and once they adjusted the kerf for the 5000 series aluminum I specified it worked great. I also just got certified to use the laser cutter at my makerspace, so I'll be cutting out my own acrylic cases soon enough. 

I also have a little something else I'll be sending over with my pull request:

Code: [Select]

skully@builder:~/kb_builder$ ./kb_cli -f foo
FreeCAD 0.16, Libs: 0.16R5358 (Git)
Processing: foo
...
Writing STEP file......
Face...

*******************************************************************
******        Statistics on Transfer (Write)                 ******

*******************************************************************
******        Transfer Mode = 0  I.E.  As Is       ******
******        Transferring Shape, ShapeType = 0                      ******
** WorkSession : Sending all data
 Step File Name : /home/skully/kb_builder/static/exports/switch_foo.stp(71734 ents)  Write  Done
saving......
processing Shape (99.0 %)
successfully exported /home/skully/kb_builder/static/exports/switch_foo.dxf
Finished: foo
Processing took: 44.63 seconds
**** Overall plate size: 322.325x103.25 mm
*** Files exported for plate switch
static/exports/switch_foo.js
static/exports/switch_foo.brp
static/exports/switch_foo.stp
static/exports/switch_foo.stl
static/exports/switch_foo.dxf


Nice.  Glad BBS worked out for you.  I have not heard of any issues with their stuff so far.

Looking like you are adding a CLI to the builder.  I actually started that way and in the process of trying to make it a 'web service', I ended up removing it.  Now that the open source python version is really targeting personal installs for makers, it really makes sense to have a CLI again.  Maybe we can even support the new __repr__() as an input file type to the CLI.  Regardless, love seeing the activity... 

[skullydazed]

Nice.  Glad BBS worked out for you.  I have not heard of any issues with their stuff so far.

Looking like you are adding a CLI to the builder.  I actually started that way and in the process of trying to make it a 'web service', I ended up removing it.  Now that the open source python version is really targeting personal installs for makers, it really makes sense to have a CLI again.  Maybe we can even support the new __repr__() as an input file type to the CLI.  Regardless, love seeing the activity... 

I think it may have been an inexperienced operator or something. The holes were .2mm wider than I specified, and I wasn't sure the surface of the cut was right, but BBS hopped right on it and made me a new plate. The new one measures perfectly.

I am adding a CLI. One of the problems I have right now is coming back to adjust my previously generated plates. I'm never quite sure if I got the corner radius right, and now that I have to worry about kerf I'll want to make sure I save those settings. The CLI is step 1 of doing that.

[potatowire]

Nice.  Glad BBS worked out for you.  I have not heard of any issues with their stuff so far.

Looking like you are adding a CLI to the builder.  I actually started that way and in the process of trying to make it a 'web service', I ended up removing it.  Now that the open source python version is really targeting personal installs for makers, it really makes sense to have a CLI again.  Maybe we can even support the new __repr__() as an input file type to the CLI.  Regardless, love seeing the activity... 

I think it may have been an inexperienced operator or something. The holes were .2mm wider than I specified, and I wasn't sure the surface of the cut was right, but BBS hopped right on it and made me a new plate. The new one measures perfectly.

I am adding a CLI. One of the problems I have right now is coming back to adjust my previously generated plates. I'm never quite sure if I got the corner radius right, and now that I have to worry about kerf I'll want to make sure I save those settings. The CLI is step 1 of doing that.

This is awesome, and I am particularly happy since I am going to try to modify your Kalerator code to be KiCAD compatible...

[skullydazed]

This is awesome, and I am particularly happy since I am going to try to modify your Kalerator code to be KiCAD compatible...

That would be awesome! If you do this I'd love to work with you so that kalerator could output either EAGLE or KiCAD when generating your PCB.