GH-122.2016 Case Options

[samwisekoi]

I made the cutouts exactly to the .75" edges, and there is plenty of room even for SA keycaps.  Remember, the keycaps have to be sized to provide space between them in the middle of the keyboard, so they will work without that extra spacing,

Regarding an all-metal case, oddly enough I have an 8x20" chunk of .250" aluminum for a base plate right here.  Not looking forward to cutting and sanding it.  But yeah, all-metal would be awesome!

 - Ron | samwisekoi

[samwisekoi]

OK, the GH-122 looks really good in the Unicomp PC-122 case ($20) with Nuclear Data Green SA keycaps.

 - Ron | samwisekoi

[Data]

Indeed it does.   


So are you thinking the extra 1/32 buffer is unnecessary?

[AKmalamute]

It occurs to me you could make the case / plate out of acrylic ... but it would have to be a little fancier than is normally done.

 This was talked about somewhere around here recently. Make a ~3mm "plate" with oversized holes. This will sit around the switches and flush against the PCB. Make another plate, 1.5mm, exactly like you would a plate. This will be reinforced by the lower piece of acrylic so the flexing of "super thin" plastic will be limited to the tiny section just around each switch. You'd probably want another ... what could you fit, say 2mm? On top of that again with the oversized holes, then bolt everything together, along with the lower piece and feet to make the whole thing act like one contiguous block of half-inch acrylic.

 Of course, I haven't tried any of this and to my knowledge no one else has either ... but it seems like it would work.

 Will Ponoko's P3 even cover it, at nearly 20 inches on a side?

[Data]

Ponoko P3 is 31.102 x 15.118.  So it will fit but I'm pretty sure you can only do one layer per sheet.

$$$

There must be cheaper alternatives to Ponoko, though.



Made some progress tonight, only the hard part remains.  I'll work on it some more tomorrow.  This is the "top" superimposed over the "plate".  I like to separate it by layers in the same file so I can see how they fit together and share common elements like edges and screw holes.  Lots of "guide" layers not shown here.   

[samwisekoi]

I do indeed think the 1/32" buffer is unneeded.  And in fact, the extra 2/32" would be helpful to the thin sections of material between the blocks.

I used Pololu for my laser-cut acrylic.  And yes, the price for plates this large is stupid expensive.  3D printing is worse.  I would like to see about vacuum-formed plastic top plates to provide more edge shape than a simple sandwich of acrylic slices.  What I would really like is a wedge-shaped base, but I can't find a way to do that in any material.

One thing I have learned is that for whatever reason this PCB doesn't like Swill "cases".  The foam completely deadens the switches so Browns feel like Blacks.  Greens are ok, and I have not tried Blues, but my normal Browns feel dead, dead, dead.  (To be fair, I am doing back-to-back testing against an F-122, but still.)  The board feels much better free-floating in the Unicomp case.  Therefore I'd suggest an air-gap under the board no matter what case design anyone tries.

Having said that, a big piece of Neoprene with a cut-out for the Teensy makes a good test and assembly setup.  This was what I used:
http://www.amazon.com/Polyurethane-Open-Cell-Firmness-Backing-Thickness/dp/B003VYAXXQ

 - Ron | samwisekoi
(Auto-typed by my GH36 Matrix Keypad.)

[0100010]

Appears Big Blue Saw is a lot cheaper for waterjet polycarbonate for a top cover compared to Pololu in laserjet acrylic for the same design.

[Data]

PC is probably a lot stronger than acrylic too.  I'll have to check that out.

[Data]

How many ISO users were there in this batch, Ron?

I'm gonna have to get creative with the plate to support that.

[Data]

ANSI 125 completed.  The red line is the PCB.



Anyone know if this object will work for 2u AND side-by-side 1u switch placement?  I realize you'd have to use PCB-mount stabilizers, but is that the only limitation?

[0100010]

Not sure if that would work.  Seems like the cutouts for the 1u side by sides would be too high for the stabs; but if they are PCB mounted maybe it is not an issue?

[Data]

Yeah, that's what I was saying.  Plate-mounted stabs would be useless, but for PCB-mounted it wouldn't matter -- and this PCB supports them.  Trying to accommodate both the double 1u switch and the single 2u switch with stabs.  I just laid one over the other and merged them.

[0100010]

Should be good to go.  Technically you don't even need the costar cutouts on the bottom either if you are going Cherry PCB mount.

[Data]

I think I'll have BigBlueSaw cut one in PETG for me so I can do a fit test.  It's only about $72.

Edit: Full plate is a whopping 20.250 x 7.250 inches.   

[Hak Foo]

"Only" about $72.  At this rate, a full GH-122 will end up being a $300 board end-to-end.  OTOH, on a price-per-square-meter metric, it still wallops Korean customs.

[Data]

We're only talking layered cases here.  If this was CNC machined it would be a small fortune.   

Just for ****s 'n gigs I ran a quote on a 3pc skeleton case (top/plate/bottom) in 304 stainless (waterjet) and it came to $197.70.  That's about what I was expecting.  I guess it's worth it to spend $72 to make sure I got the drawing right if it means avoiding a $200 mistake.

[bazh]

This is interesting if making into a CNC'd aluminum key, the thing is huge


Sent from my iPhone using Tapatalk

[Data]

This is interesting if making into a CNC'd aluminum key, the thing is huge


Sent from my iPhone using Tapatalk

That would be awesome (and heavy!) but way too expensive.

[bazh]

This is interesting if making into a CNC'd aluminum key, the thing is huge


Sent from my iPhone using Tapatalk

That would be awesome (and heavy!) but way too expensive.

Yeah that would be easily $700-1000 for the case only

[blighty]

OK, the GH-122 looks really good in the Unicomp PC-122 case ($20) with Nuclear Data Green SA keycaps.

 - Ron | samwisekoi

Was this PCB cut to fit the top of the case?

ANSI 125 completed.  The red line is the PCB.

Show Image

Show Image

This is looking good so far.  Is there a chance you could mock up a stepped caps lock and a 1.5 mod bottom row (with or without blockers on top plate)?

[Data]

OK, the GH-122 looks really good in the Unicomp PC-122 case ($20) with Nuclear Data Green SA keycaps.

 - Ron | samwisekoi

Was this PCB cut to fit the top of the case?

ANSI 125 completed.  The red line is the PCB.

Show Image

Show Image

This is looking good so far.  Is there a chance you could mock up a stepped caps lock and a 1.5 mod bottom row (with or without blockers on top plate)?

I was considering an "ANSI" and "ISO" plate with the most common options on each.  It's impossible to support everything though.

[njbair]

This is interesting if making into a CNC'd aluminum key, the thing is huge


Sent from my iPhone using Tapatalk

That would be awesome (and heavy!) but way too expensive.

Yeah that would be easily $700-1000 for the case only

Better way to spend $1K than a Korean custom IMO.

[samwisekoi]

I'm moving over here now for everything except shipping.

Looking at all of the plate options here and in the other thread, it seems like a PCB + .375" form factor is the way to go.  I've done some work on the switch/stab cutouts, and think that a single plate could be made to cover everything except ANSI vs ISO Enter clusters.

So if we can agree on a plate form factor (i.e. agree to use Data's form factor), we can cut the cost per plate tremendously, even if we have to have two versions.  This is especially true if we use cover plates that are specific to detailed layouts.

This weekend I'll work on two sections of the hard bits on the plate:

#1 Top rows.  The left and right blocks are easy; but the big 30-position top block has too many options unless the plate just has a couple of long 0.550" rectangular holes up there.  Rather than that, I think there are a few sensible layouts (F-122, 2xANSI, etc.) for the top, and I'll try to come up with something more helpful than two giant slots.

#2 ISO vs. ANSI Enter clusters.  This is the big one, and if there is some way to create a plate with snip-and-remove sections so we can consolidate the two major variants, it would be a huge win.  I'll try to accomplish that.

I'll feed the results of both efforts back here in the form of PNG/PDF/DXF/VSD files as before.  Hopefully Data or someone can use my work to make a more nearly universal design or designs.

 - Ron | samwisekoi

[Spaceman1200]

Wouldn't something like this work on enter? it would limit to pcb mount stabs, But i am fine with that as i have plenty of both

[Data]

Wouldn't something like this work on enter? it would limit to pcb mount stabs, But i am fine with that as i have plenty of both

Show Image

Potentially, yes. We've already compromised on the stabilizers in other areas so an Enter cutout like that might work. I'll have some time to mess with it tomorrow.

Edit: Ron, do you have a vector version of this sweet GH-122 logo you can share?

[Data]

I didn't end up working on the drawing at all this weekend.  Star Citizen alpha 2.4 proved too much of a distraction. 

I should have more time this week.  I'd like to order a sample plate before Friday so I can do a fit check.

[bazh]

Wouldn't something like this work on enter? it would limit to pcb mount stabs, But i am fine with that as i have plenty of both

Show Image

I've done something like this with my Weaven group buy round 1, the problem with the enter cutout is that it's left with the "|\" key not fully supported for ANSI user ( have to use a PCB mount switch to ensure the firmness), ISO user will have the same issue with their vertical enter key so we ended up making 2 separated plate for the round 2 :D

[Data]

Good to know.  Thanks bazh.

[Data]

OK, I was wanting to send this off to Big Blue Saw tonight for a test cut in PETG.  This is as close as I can get to a "universal" ANSI plate. 



But the space bar looks really funky for some reason.  Did I do this right?



Units are inches.  I only dimensioned the layout with the 7u spacebar because I'm already 99% confident that my 6.25 layout is correct and solid.  But it still looks really weird.

[Data]

Well if I screwed it up I'll find out when it gets here.  )

Test plate ordered.

[Hak Foo]

I am curious how it will hold up to sag.  That's the big problem with the GH-122 (or the Cherry 8200 for that matter) -- big boards without a super-stiff plaqte sag.

[Data]

Yeah, without any center-mounted supports I'm thinking steel is probably the logical choice for this beast.  I'm not expecting any load-bearing properties from PETG, although it is a fairly rigid plastic.  I only used it for this order because it's cheap.

[Data]

Order shipped, it's not going far so I'm hoping to have it by Friday. 

[Data]

Test plate received.  I'm really glad I did this because I found one pretty major error and a couple of other things that concern me.  Also this PETG material is awesome.

Here's an album:  http://imgur.com/a/RSeQ6





Issue #1: I cut the Right Shift hole completely wrong.  It'll have to be redrawn.  I'm pretty sure I just inverted the positions of the 1.75u "Short" Shift and its 1u companion key.  Everything is fine for the standard 2.75u Right Shift.

Issue #2: I can't say for certain but I'm concerned that the holes in the PCB for the space bar stabilizers might be misplaced for the 6.25u space bar.  They look OK for the 7u but they're way off for the ANSI 125 layout.  Either I did some Florida Math here or we have a problem.


- More photos in the album!

Issue #3: The circles marking the locations of the lock indicator LEDs are a little off center vertically.  I'll measure again and adjust them up a smidge to compensate.

Issue #4: The massive battle scars at the top of the plate are a serious weakness.  Can we talk about that area?  Two giant holes basically remove all structural integrity from that top cluster.  Did we ever figure out if there's a way to shore it up a little and still keep it accessible for the most common layouts?  I'm open to suggestions here.

Thanks to anyone still reading.  I feel like I'm mostly on my own here, so unless I get some feedback I'm just going to forge ahead and make something for me.

[Hak Foo]

So... what did it cost for a plate like that, and how is the sag factor of the material?

In terms of fixing the big top openings, I personally think the top two banks will break down into one of four options:

* Full 30 positions
* 26 positions (two normal 12-function rows plus escapes)
* 25 positions (only one escape)
* 24 positions (2x12 block in a conventional 122 style)

[Spaceman1200]

The holes for the 6.25 pcb mount are indeed way off but your plate mount should work fine. for the top clusters, i'd be game for 122 style

[Data]

So... what did it cost for a plate like that, and how is the sag factor of the material?

In terms of fixing the big top openings, I personally think the top two banks will break down into one of four options:

* Full 30 positions
* 26 positions (two normal 12-function rows plus escapes)
* 25 positions (only one escape)
* 24 positions (2x12 block in a conventional 122 style)

Plate was $68 +shipping.  It's not quite as rigid as acrylic, so it doesn't make a great plate at the thickness I purchased (0.063" / 1.6002mm) and given the size of this board -- there's noticeable sag around the top cluster and over the full length and width of the plate.  But it is fairly rigid so it would work for cases at greater material thickness or with smaller overall dimensions.  It's denser than acrylic and optically clear.  I like it mainly because the risk of stress fractures is essentially zero.  BigBlueSaw has it up to 1/4" thick.

I'm playing with the numbers you listed now to see if we can work something out.

[Data]

OK, I made these 4 samples.  Does this cover what most people are looking for?

FULL 30


TRADITIONAL F-ROW V1


TRADITIONAL F-ROW V2


MODEL F-122

(F-block centered over number row keys)

[Hak Foo]

Personally, I'd prefet Tradfitional F-row v1, but I'd probably take any of them to maximize economies of scale.

[Data]

I realized after posting it that the F-122 layout is essentially the same as the full 30.  The switches are in the exact same places, there are just fewer of them.  So the only locations where we need to accommodate the half-position switches are around the F5-F8 block and the two left-most rows where some might want to put an Escape key.

Provided there are no surprise additions to these 4 suggested layouts. 

[blighty]

I'd probably go for any of them as well, but I'd prefer ANSI 150 instead of 125.   

[Data]

I'd probably go for any of them as well, but I'd prefer ANSI 150 instead of 125.

Switch positions for both bottom row layouts are already in the drawing and the test plate.

[Data]

Looks like Imgur is down (again) for some reason.

Anyway, here is the adjusted plate.


Supports all the above top-row layouts, plus:

• ANSI 125 bottom row
• ANSI 150 bottom row
• Off-center stepped OR centered Caps Lock
• Split Backspace
• Short Right Shift
• ANSI Enter
• 2x 1u OR 1x 2u (horizontal) in the left column (using PCB stabilizers)

And yes, I'll do one for ISO if there's any interest  The only differences would be the Enter key layout and the split Left Shift.

Did I miss anything?  Does someone want to validate my spacebar placement so I can stop being paranoid about it? 

[AKmalamute]

I like it. And the PETG, is it strong enough to have a thinner plate, so the keyswitches will 'clip' into place?

I might request the final .dxf and order 1.5mm PETG and then get 5mm silicone with larger holes so it won't interfere with switch adherence but still support the whole of the plate. Not sure if that would be net less expensive than a SS-16ga plate but I don't expect to be running one of those GBs, it sounds like.

[Data]

I like it. And the PETG, is it strong enough to have a thinner plate, so the keyswitches will 'clip' into place?

I might request the final .dxf and order 1.5mm PETG and then get 5mm silicone with larger holes so it won't interfere with switch adherence but still support the whole of the plate. Not sure if that would be net less expensive than a SS-16ga plate but I don't expect to be running one of those GBs, it sounds like.

Thanks. I'll gladly make the DXF and DWG files available to whoever wants them.

You make an interesting point about the GBs. Engicoder and I were talking about that earlier. There were so few of these made/sold that it's very likely everyone will just do their own thing, and a few might never even get made. There are so many options with this board it's unlikely that any two people (among 20?) will want the same finished product.  So it kinda sucks for saving money but at least there will be some unique boards. 

Edit: And yeah, PETG would work fairly well for a plate, especially if it was a little thicker.  At 1.6mm there's just not enough material to prevent sag.  This board is huge!  It's especially noticeable around the two top rows -- a problem I hope to have eliminated in the latest design.  In the areas that are better supported, everything snaps in nicely like you'd expect.  I have little doubt that the material would work well for cases in other areas -- it's definitely superior to (clear) acrylic.  Metal is still king for plates.

[Data]

Looks like Imgur is down (again) for some reason.

Anyway, here is the adjusted plate.

Show Image

Supports all the above top-row layouts, plus:

• ANSI 125 bottom row
• ANSI 150 bottom row
• Off-center stepped OR centered Caps Lock
• Split Backspace
• Short Right Shift
• ANSI Enter
• 2x 1u OR 1x 2u (horizontal) in the left column (using PCB stabilizers)

And yes, I'll do one for ISO if there's any interest  The only differences would be the Enter key layout and the split Left Shift.

Did I miss anything?  Does someone want to validate my spacebar placement so I can stop being paranoid about it?  :rolleyes:

Going to confirm the two main bottom row layouts on my test plate when I get home, then I'll probably order this as part of a 3pc sandwich case tonight.

[Data]

Bottom row is good.

ANSI 125 fits but a plate-mounted stabilizer is required.


ANSI 150 fits but a PCB-mounted stabilizer is required.


I also made sure the stepped, off-center Caps Lock was correct.

[Data]

OK, here are some files for whoever wants them.

AutoCAD 2013 DWG "master" file
 GH-122 Case - Universal.dwg
AutoCAD R12-format DXF - universal plate only
 Plate ANSI Universal Prime.dxf
AutoCAD R12-format DXF - Data's Cuts (this is what I'm submitting to BBS)
 GH-122 Case Prime - Cut Layout.dxf

A note on the DWG file --
I put everything on layers so you can select different layouts by just turning the appropriate layers on and off.  They're labeled so it should be self-explanatory once you start looking at it.  I also left my guide layers and some dimensions in there but you shouldn't need them.  If you use this to create your own plate/case make sure you move everything to Layer 0 and delete the other (empty) layers in your DXF before you save & submit for cutting.  No ISO support yet, but it shouldn't be too difficult to add.  For those without access to CAD, PM me if you need something special and I'll try to help.

[Hak Foo]

What did you use to draw these files?  They open in LibreOffice but seem to be scaled wildly off, and FreeCAD seems to only find about half the shapes (the screwholes and the big cuts in the top plate, but not the individual switch cuts or the plate borders)

The LED cutouts on the plate seem to render as a single line, not an actual box.

[Data]

AutoCAD 2013

Which files are you opening?  The DXFs are AutoCAD R12 format because that's what BBS likes. The site has never balked at any of my files, but I can export them in a newer DXF format if you think that would help. The shapes are all polylines and circles, nothing fancy.  Does FreeCAD handle polylines?