CNC Aluminum Phantom Build Log

[lxrogers]

Hi guys,
I just started a keyboard project and I often come to this forum for inspiration and resources, so I thought I would post my progress here.

I am making a keyboard using the Phantom PCB and Teensy controller. The case is going to be the fun part, as I'm CNC'ing it out of a two solid chunks of aluminum.

Render:


Components

More

• Cherry MX Red Switches
• Teensy USB Controller
• Phantom PCB
• CNC Aluminum Case
• Keycaps TBD

Starting Pictures

More

Keyboard Components

Stock

I hope you guys enjoy the build log! I have a few questions, too.

1. What color should I anodize the case? I'm currently leaning toward a more subtle grey-blue, as I'm not going for a crazy vibrant green vibe.
2. Based on that color, what color keycaps should I get?
3. Does anyone have a Phantom or generic Tenkeyless cad of the cover? I don't know the exact dimensions around the areas of the island function keys.

For me, I'm thinking of going with a grey-blue and white engraved PBT caps.

--LXR

[JackMills]

If by cover you mean the mounting plate, I think you can find the CAD files here at the Phantom wiki

And the idea of going with grey-blue anodizing combined with white keycaps sounds good to me. I personally like an orange case with black keycaps.

[lxrogers]

I mean the part that usually goes on top of the mounting plate (although on mine they will be one piece).
Like this one from my Ducky:

[JackMills]

Ah my bad, I think it should be the same as a Filco tenkeyless, but I cannot find any usable CAD files for it. Only screenshots of CAD designs

Apparently photekq did a design for an aluminum filco/phantom case with a cover, maybe you can ask him for some info

[exitfire401]

Looking forward to how this progrsses! Good luck 

[Findecanor]

You could deduce that from the plate files. Each 1×1 key is square with the switch mounting hole in the dead centre.
However, you might want to shave off half a mm or so from the edges to give each key near an edge as much space to the edge as to its neighbour.
Also, some keycaps have rounded corners at the bottom, while others have sharp corners and if you are really anal about design you might want to make the case match.

[lxrogers]

Machining Update

Machining the pocket for the PCB:



This part was fairly straightforward. I had to do some measuring to find out where I could put the support extrusions. I'm also planning on using these to screw the mounting plate onto the case.

Machining the fixture plate for the bottom part:

This part is necessary to hold the bottom part in while I contour the bottom of it. I added extra holes to a fixture plate I already had. You can see it uses two dowel pins to constrain in the XY plane, and two screw holes for constraining in the Z direction.

Rough pass for bottom contours:

This part was really fun. The rough pass made a cool topographic style. After this, I will go in with the curved end-mill to smooth it out.

I will have better pictures next week as the battery for my digital camera died and I had to use my phone camera.

[lxrogers]

You could deduce that from the plate files. Each 1×1 key is square with the switch mounting hole in the dead centre.
However, you might want to shave off half a mm or so from the edges to give each key near an edge as much space to the edge as to its neighbour.
Also, some keycaps have rounded corners at the bottom, while others have sharp corners and if you are really anal about design you might want to make the case match.

ahh yes but the spacing is what I want to get right I think I will laser cut some mounting plates and covers and see if I can prototype the spacing to get it just right.

[Zeal]

Paging CPTBA... Beep...Beep! 

[CPTBadAss]

Hi yes, I was alerted to machining porn and I'm now here. I just went six to Midnight. All that lube and shiny metal and chips? Awww yeee. Really excited to see the rest of your progress!


I'm really digging that shape in your model.

[Joey Quinn]

Can't wait to see the finished case, that render looks awesome. Also thanks for posting the machining process it may be my favorite part.

[JackMills]

That contour will be gorgeous, even the topographic contour is nice. I can't wait to see this case when it's finished

[lxrogers]

I got the new battery for my camera so here is a better picture of the rough pass.



Unfortunately I won't be able to work on the machining this week because the shop is closed for the whole week for Thanksgiving Holiday. But maybe I will be able to make some progress on the CAD for the mounting plate or the cover part.

[HPE1000]

Awesome progress, this should look great 

[lxrogers]

After a long delay from vacation I have many updates! I'll spread them out.

First I worked more on the contouring. I realized that because of a boundary issue in HSMWorks I didn't actually finish the rough pass. Here it after the full rough pass:





Now it was time to use the 1" ball-end mill with carbide inserts!! This should smooth it out. Calculating the feeds and speeds I got a whopping 160 feet/min, which kind of scared me so I turned it down a bit. Even still, the thing vibrating like crazy making a HUGE NOISE. The cutout on the bottom basically made the thing a speaker. It was nuts.

 Here is the result:



As you can see, some of the rough pass is still there. This threw me for a loop. I'm pretty sure that the contour with the ball-end mill was the correct tool path, so perhaps the rough pass was deeper than it was supposed to be? After much research I still couldn't figure out what was wrong. However, everyone in the shop who saw it thought I did it on purpose saying it looked really cool. So I'm going to go with doing it on purpose! Haha!

Next I started working on the mounting plate. It involved testing and researching a lot to line of the plate holes and the raised parts of the frame/case.

I shall return soon.

--LXR

[Evo_Spec]

Very cool!

Watched to see updates!

[lxrogers]

Next I started working on the mounting plate. I was super worried that the plate dimensions wouldn't work out. I started off with this CAD which was imported from the Phantom .dxf files:



However, after checking out the Cherry MX CNC thread: (https://geekhack.org/index.php?topic=65747.0;topicseen ) I realized that the design might not machine correctly depending on the end mill size.

I wanted to avoid having to draw my own sketch as much as possible, because then I wouldn't be sure if the holes would line up with the phantom, but I had to go for it and trust the math. This is the design I came up with:



You can see the half circles on either side, which allow for the tool to go all the way to the edges and not leave a rounded corner. The diameter of the extra semicircle is 1/16th, which is the size end mill I would be using. I needed to test this (and the layout), so I got some 1/16th acrylic (about this thickness Cherry prescribes for its mounting plates) and laser cut them:





As you can see the switches seemed to fit just fine. As a side note, the reason I used a 1/16th instead of something like 1/8th, which would have fit, was because I wanted to make the material in between the switches as thick as possible to make the plate stiffer. This means I'm in for a long ride because a 1/16th end mill will have to go pretty slow because the CNC I'm using has a maximum spindle speed of 8100. (The smaller end mills want to spin as fast as possible to reach higher feed rates). With that spindle I'm looking at 8 ft/min, which means cutting this thing out will take around 5 hours.

More on that soon!

[scubaste]

This is awesome. I could look at pictures of this process all day. Not many people have access/ability to make something like this happen. Good work, keep us updated!     

[CPTBadAss]

Lxrodgers, do you think that a faster or slower process time influences how the bit wears when cutting these materials? Are you worried about breaking any bits?

[lxrogers]

Lxrodgers, do you think that a faster or slower process time influences how the bit wears when cutting these materials? Are you worried about breaking any bits?

It definitely does. The 1/16th drill bit, according to FSWizard (an online feeds & speeds tool) should feed at 8 ft/min using a spindle speed of 8100 which is the max the machine I'm using can spin. I had a deadline for this project since I'm doing it for class, and ended up in a super stressful situation!

I had about 4 hours left on the machine to finish making the mounting plate, for the entire school term, however HSMWorks said the operation would take 4.5 hours using 8ft/min. I knew I had to go all or nothing because a keyboard with missing keys is only marginally better than a keyboard with no keys. So I ran the tool path at 120% feed knowing there was a good chance it could break my tiny little 1/16th guy. It was double ended, so if it broke I could try only one more time.

In the end it didn't break. I think High Speed Steel (material of tool I was using) on aluminum is a pretty good situation, and you will only really break the tool if you are way off with your feed and spindle speed. But you're right, running the feed faster than it should increases the chip load on the tool (basically how much it slices off per rotation) which will cause it to wear down a lot more.

When machining a steel project with 3/16 bits, I broke about 4 of them haha.

Needless to say I almost died during that shop session from willing the machine to go faster and holding my breath while checking to see if the tool broke yet. I'll post pictures of this saga soon!

[CPTBadAss]

LOL. That sounds hilarious hahaha. Hopefully everything worked out and nothing broke too bad. The shops I worked in at school were very chill if you broke bits though. So I think you should be good?

And that's cool that this project taught you about chip load and feed rates .