Introducing the GH36 Programmable Matrix Keypad

[samwisekoi]

PROTO_1 is done and ready to go out for fabrication.

Current list of prototype testers is:

PROTO PCB RECIPIENTS
 - samwisekoi
 - regack
 - metalliqaz
 - jdcarpe (2x)
 - domoaligato (2x)
 - SpAmRaY (2x)
 - HoffmanMyster
 - dorkvader

This list will be updated in the OP after this post.

Here is v140808b8 aka PROTO_1:

More

GH36 Matrix Gamepad Prototype PCB #1 by samwisekoi 2014

I have added additional mounting holes (total of nine now), as well as additional cut lines to provide 2x4, 2x6, 3x4, 3x6, 4x4, 4x6, 5x6, and 6x6 configurations.

All switch pads are twinned for reversible use, and all switches have LED resistors connected to a much larger ground plane.  However, only WASD/ESDF and NUMLOCK LED circuits are powered.  Most switches do not have LED positive traces, and the LED1 and LED2 pins are not connected, although they are labeled and available for hardwiring.

Future versions will have full LED support, and I would like to add the capability to create a 2x12 configuration to enable an F-122 layout using a GH36 plus a traditional PCB.  (We'll have to get PF1-24 keycaps made to match Nuclear Data Green for that config!)  And I will try to better match form factors of existing 60% and 75% keyboards.

Anyhow, this one is going out for fabbing today.  If you are on the list for a PROTO_1 PCB or two, please PM me your shipping address.

Thanks to one and all!

 - Ron | samwisekoi

[samwisekoi]

Order placed for 20 PCBs.  Had to rev it up to v140808b9 after going through the error check.  11 day quoted leadtime, although my last order got here much faster than that.

I'll get the design file up on Github when I can.

 - Ron | samwisekoi

[Grendel]

I also would be interested to build one up (as a left-hand controller.) Just finished a LH "ErgoPad" and while working very nicely there's room for improvement -- this could be it.

[regack]

Well, I hope it all lines up... I guess I can test whatever configuration or something...

[domoaligato]

I also would be interested to build one up (as a left-hand controller.) Just finished a LH "ErgoPad" and while working very nicely there's room for improvement -- this could be it.

that is sweet!

Would anyone be able to create a acrylic plate case design for these?
I would like to contact a local company to have a case made but I suck with cad

[jdcarpe]

I also would be interested to build one up (as a left-hand controller.) Just finished a LH "ErgoPad" and while working very nicely there's room for improvement -- this could be it.

that is sweet!

Would anyone be able to create a acrylic plate case design for these?
I would like to contact a local company to have a case made but I suck with cad

I'm on the case! )

Would you be using 5mm and 3mm acrylic for the layers, like the ErgoDox cases?

[domoaligato]

whatever you come up with bro. your the master cad guy and you have done this before. I was thinking like a split 60% litster case

btw I followed your twitter....

[samwisekoi]

Well, I hope it all lines up... I guess I can test whatever configuration or something...

Your plate should be fine.  This is a 5" square PCB with 36 0.75" switch footprints centered on the board.  Mounting holes are 0.125" diameter, 0.125" in from the edges and corners (on-center) plus dead center at 2.500".  (I'll make a mechanical drawing.)

I'm on the case! )

Would you be using 5mm and 3mm acrylic for the layers, like the ErgoDox cases?

However, let's not all rush out and make cases and plates just yet.  This first set is an electrical prototype only.  The next rev will have different outside dimensions and mounting holes to better match existing keyboard PCBs.

I have a sample PCB for just about everything except a JD40, and I'll grab one of those from Ming.

I was just going to mount the prototypes on stand-offs above a 5" square piece of acrylic and put sticky rubber feet on the bottom.

Still, this is a big-time DIY thing, so make your one-offs as naughty or as nice as you want.

 - Ron | samwisekoi

[domoaligato]

I like naughty   )

[regack]

Well, I hope it all lines up... I guess I can test whatever configuration or something...

Your plate should be fine.  This is a 5" square PCB with 36 0.75" switch footprints centered on the board.  Mounting holes are 0.125" diameter, 0.125" in from the edges and corners (on-center) plus dead center at 2.500".  (I'll make a mechanical drawing.)

ok, great, my standoffs are actually outside the PCB, I actually just threw this together, since the one-off will fit INSIDE an area of acrylic that I was going to have cut out, so I'm not wasting any material cost, just more cutting time.

[jdcarpe]

I have a sample PCB for just about everything except a JD40, and I'll grab one of those from Ming.

Actually, I'll have to rectify that situation...

[Hak Foo]

I'm interested in the beta/production round because it looks like it will be feasible to use it with ALPS switches, thanks to the mirrored holes.

[MOZ]

Interesting project, had I not been so busy with RL, I would have loved to be a part of this.

Best of luck to make this a success

I would highly recommend editing the switches holes slightly to allow for using alps switches. That would be awesome.

[samwisekoi]

I'm interested in the beta/production round because it looks like it will be feasible to use it with ALPS switches, thanks to the mirrored holes.

Interesting project, had I not been so busy with RL, I would have loved to be a part of this.

Best of luck to make this a success

I would highly recommend editing the switches holes slightly to allow for using alps switches. That would be awesome.

Someone send me a dimensioned drawing and I will take a look for the next version, which is already in design.

 - Ron | samwisekoi

[MOZ]

Check OP: http://geekhack.org/index.php?topic=47744

[samwisekoi]

Sorry; won't work with Cherry, flipped Cherry, and Alps.

The flipped Cherry pads -- and their connections -- will short out the Alps switch.


Pad positions for Cherry, flipped Cherry, and Alps switches.
Cherry: 1&2, Alps 3&4, Cherry LEDs 5&6

I can do MX+Alps or flipped MX, but not both.

 - Ron | samwisekoi

[samwisekoi]

Also, the second pass design is a bit cleaner.  The image below only shows the row and column traces, but you can see they are better than in PROTO_1.  Also, I have added SMD resistors for the LED lighting, although I have also retained through-hole resistors for WASD, CAPSLOCK, and NUMLOCK.  Possibly I will be able to run the LED traces so that if you install all SMD resistors, you get a single light circuit, but if you install the through-hole resistors that will enable WASD, NUMLOCK, or CAPSLOCK LEDs at those positions.  No promises, however.

I have not decided where to place the multi-board extension pads, or in fact what cable to use.  DB-25 (VGA) at the top might work, or just a ribbon.  Whatever gets put there, it needs to be simple enough for average folk to buy, install and use.

More

GH36 PCB DRAFT v14080813b2 by samwisekoi 2014

 - Ron | samwisekoi

[MOZ]

Sorry; won't work with Cherry, flipped Cherry, and Alps.

The flipped Cherry pads -- and their connections -- will short out the Alps switch.

(Attachment Link)
Pad positions for Cherry, flipped Cherry, and Alps switches.
Cherry: 1&2, Alps 3&4, Cherry LEDs 5&6

I can do MX+Alps or flipped MX, but not both.

 - Ron | samwisekoi

Why not have the switch holes on one side (Left or right) connected to one side of the matrix and the other side to the other side of the matrix.

[samwisekoi]

MOZ; I will take a look at that in a bit.

Attention LED-knowledgeable folks!

I am not a backlit keyboard guy.  I have one, a TG3, and I use it to game in the darkness.  And I want to add LED support to the GH36 for gamers and backlight fans.  However, I want to enable builders to choose full backlighting or WASD/NUMLOCK/CAPSLOCK LEDs, or both.

To accomplish this feat, I have come up with a circuit that enables both on the same switch.  If I have done this properly, installing the through-hole resistor (labeled "WASD") will enable the WASD LED circuit, but not the LED1 circuit.

Contrariwise, installing the SMD resistor (labeled "R14") will connect the LED to the LED1 circuit, and not to the WASD circuit.

Those are the two important items I would like validated by someone whose expertise extends to that space.

A third "nice to have" feature of this circuit that I believe will also work, is that installing BOTH resistors will illuminate the LED when either circuit is powered.  This is less a feature than an error-accommodation side-effect.

So, does the circuit below enable selection of the desired LED circuit by installing one resistor versus the other?


GH36 Proposed LED Circuit

 Thanks in advance for any learned responses!

 - Ron | samwisekoi

[MOZ]

To accomplish this feat, I have come up with a circuit that enables both on the same switch.  If I have done this properly, installing the through-hole resistor (labeled "WASD") will enable the WASD LED circuit, but not the LED1 circuit.

Contrariwise, installing the SMD resistor (labeled "R14") will connect the LED to the LED1 circuit, and not to the WASD circuit.

Those are the two important items I would like validated by someone whose expertise extends to that space.

A third "nice to have" feature of this circuit that I believe will also work, is that installing BOTH resistors will illuminate the LED when either circuit is powered.  This is less a feature than an error-accommodation side-effect.

Installing the resistors in either one of the resistor location will enable that circuit given the transistor state (Which I am assuming is connected to the uC?) however, if both are installed and both circuits are enabled then the resistance will be halved and you cold potentially damage the LED.

[samwisekoi]

Installing the resistors in either one of the resistor location will enable that circuit given the transistor state (Which I am assuming is connected to the uC?) however, if both are installed and both circuits are enabled then the resistance will be halved and you cold potentially damage the LED.

Thank you very much, MOZ!  So my plan will work so long as the builder installs one and only one resistor per position?  (And yes, those traces are connected to pins 12/D7, 13/B4, VCC, and GROUND  on the Teensy.)

Moving on to LED placement, I want to enable:

1) Full backlighting with PWM control (B6:LED1).
2) Top-row backlighting with  PWM control (B5:LED2).
3) WASD/ESDF backlighting with PWM control (D7:WASD).
4) NUMLOCK LED on NUMLOCK key (B4:NUM).
5) CAPSLOCK LED on CAPSLOCK key when flipped (secondary use of B5:LED2)
6) Top-right stand-alone LEDs (no switches) for WASD, NUMLOCK, and CAPSLOCK.

This is how that will be on the PCB:


[iGH36 Proposed LED Placement[/i]

Comments?

 - Ron | samwisekoi

[MOZ]

Regarding the LED circuit, I don't see why not, it is a simple matter of closed circuits, if only one resistor is used, then there is only one path for the current to flow, if both are used, then since the resistances are in parallel then assuming the resistance are same, it will be halved (Parallel resistors rule).

Regarding LED setups, how many available pins do you have for LEDs? Why not go for individual LED control using spare pins (If available) or a secondary I2C device.

[James35]

Is it possible to make a qazpad layout with this project?  I'd like a gap between the numeric pad and the extra rows.

[samwisekoi]

Regarding the LED circuit, I don't see why not, it is a simple matter of closed circuits, if only one resistor is used, then there is only one path for the current to flow, if both are used, then since the resistances are in parallel then assuming the resistance are same, it will be halved (Parallel resistors rule).

Regarding LED setups, how many available pins do you have for LEDs? Why not go for individual LED control using spare pins (If available) or a secondary I2C device.

Thanks.  Four pins are available; three of which are PWM.

Is it possible to make a qazpad layout with this project?  I'd like a gap between the numeric pad and the extra rows.

Not the prototype, no.  But I have been wondering if future versions  should allow a 0.375" gap between the top row and the lower five rows, as found on a "normal" keyboard.  So, possibly that will be supported, but probably not the gap to the left two rows, as that could limit the potential uses of the general-purpose keypad.

I love the Qazpad, however, so I am open to modifications that don't impact general usefulness.

[edit] I did a quick layout check, and adding an optional 0.375" gap under the top row is a simple change if I rotate the new switches 180.  Can do, easy.
[/edit]

Taking the Qazpad design and scaling it way the heck up, would anyone be interested in an add-on configuration to turn a 104-key keyboard into a GH-122?


GH-122 by samwisekoi 2014. GH-36 modules highlighted with white borders.

 - Ron | samwisekoi

[SpAmRaY]

Maybe it's time to revisit the idea of a modular keyboard with multiple inter-connectable PCB's to make the different layouts. There's already an interest check for a full size plate that can be broken apart to smaller plates just need to expand it a little. Not sure what to do for cases though.

[samwisekoi]

Maybe it's time to revisit the idea of a modular keyboard with multiple inter-connectable PCB's to make the different layouts. There's already an interest check for a full size plate that can be broken apart to smaller plates just need to expand it a little. Not sure what to do for cases though.

Giant hunks of CNC-cut aluminum!  Gotta have the top 24 up at an angle anyhow!

 - Ron | samwisekoi

[margo baggins]

is it too late to get involved in the proto stage?

Always down for prototyping electronics

[AKmalamute]

is it too late to get involved in the proto stage?

Always down for prototyping electronics

Pretty sure the prototype order has already been sent off for production. I thought about asking how much experience he wanted for prototypers but I'm not completely sure I have an extra teensy plus some other concerns about being fast enough with test results, etc.

 I will however want in on the beta-round.

[samwisekoi]

Proto PCBs have shipped!

If you are on the prototypers list (see the OP), make sure you PM me your current shipping address!

Also, Mr. Mettaliqaz, sir, how and which Easy AVR version should we use at this stage?

 - Ron | samwisekoi

[Grendel]

Any chance to get onto that list ? I can throw together some basic firmware w/in an hour or so if need be.

[samwisekoi]

Any chance to get onto that list ? I can throw together some basic firmware w/in an hour or so if need be.

Let me see what is left after shipping and doing my own testing.

 - Ron | samwisekoi

p.s . There WILL be a second round, and that PCB design already looks better than the prototype.

[infiniti]

Count me in for the next round!

[metalliqaz]

Proto PCBs have shipped!

If you are on the prototypers list (see the OP), make sure you PM me your current shipping address!

Also, Mr. Mettaliqaz, sir, how and which Easy AVR version should we use at this stage?

 - Ron | samwisekoi

Sorry I've been trying to buy a house.  Between that and work... just so little free time.  I'll post it today.

[metalliqaz]

Okay that didn't take long...

Windows beta download
Multiplatform beta download

Edit: Use new release at the Easy AVR USB thread

Test and tell me if:

• Keys don't get mapped to the location you intended
• LEDs are inverted
• You do a debug build and get stray '*'

I fully expect backlights to take a few tries to get configured right.  It's no big deal.

[samwisekoi]

Wow! Thanks very much!

 - Ron | samwisekoi

[samwisekoi]

Christmas came early this week!

 - Ron | samwisekoi

[SpAmRaY]

Those are gorgeous!

[jdcarpe]

Woo hoo!

[Vibex]

Christmas came early this week!

 - Ron | samwisekoi

Looks awesome. Would love to get my hands on one when they are out of the testing phase.

[metalliqaz]

Wow, they look good.  They look real good.

[James35]

I'm interested in the Number Pad version as long as it has LED backlighting.

[domoaligato]

Can't wait!

[samwisekoi]

Doing some unit testing before I send the prototypes out.

 - Ron | samwisekoi

p.s.  Plan on using PCB-mount switches!  I didn't, and so they are not crisply lined up. 

[samwisekoi]

Ok, I have completed unit testing enough to know that the only thing preventing shipping these out to the prototypers is fear of unknowns.

Switches work.  Diodes work.  (Three of them are intentionally placed upside-down, and have little caution symbols next to them.)  WASD LEDs work.  Transistors are a PITA to solder, but they go in fine.  Stabilizers work.

One problem found: There is an extra trace that shouldn't be there, and probably was unintentionally placed. Fixing that requires two tiny cuts in a trace. I may do that to all of the PCBs, and I will do it for anyone who is not comfortable with that task.

As I mentioned above, I strongly recommend the use of PCB-mount switches.  They work WAY better, and snap right in with a satisfying thunk.  it makes for much easier soldering, and the result is much better as well.

FYI, these prototype PCBs cost me $14.70 each, and I'll send them to you via Priority Mail (another $5.80 per shipment).  So if you want to toss some of that my way via PayPal, I won't object.  paypal AT samwisekoi DOT com

Anyhow, once I find a good source for PCB-mount switches, I will be quite happy with this.  But there are many things I will change for the Beta run.

Please remember to PM me your ship-to address!

 - Ron | samwisekoi

[AKmalamute]

As I mentioned above, I strongly recommend the use of PCB-mount switches.  They work WAY better, and snap right in with a satisfying thunk.  it makes for much easier soldering, and the result is much better as well.
[...]
 - Ron | samwisekoi

MK has clears in PCB-mount, although I think that's the only one. Other switches would require buying clears and swapping the sliders -- I commented on this in their "name anything" thread but haven't seen a response.

Still, you can get PCB mount clears right now, if you want.

[samwisekoi]

Ok, I pulled all the switches off, and removed the diodes from a bunch of Browns I had.

So here it is.  Tonight's beauty shot of the GH-36 Matrix Keypad.  Just waiting for its Teensy.

Until tomorrow,

 - Ron | samwisekoi

p.s.  It is Sunset here; the greenish tint is a reflection.

[worldspawn]

I'm in on the beta run     will monitor the thread.  I have a laser cutter, so if peeps need some cheap acrylic cases/plates, get in touch.

[metalliqaz]

I have a laser cutter, so if peeps need some cheap acrylic cases/plates, get in touch.

RIP your inbox

[samwisekoi]

HUGE THANK YOU  to metalliqaz for his efforts in tuning the firmware last night.  Even though we have an untested board, and it was far away from him, and I have very limited Windows skills (Linux-dude, me) I present to you the first line of text produced by a GH36!

aaaaaaaaaaaaaa111zzzzzzzzzzzzzzzzzzvvv vza111111aaaaaaa111aaazzvvvvvvvvvvzzzzaaaaaazzzzzzzzvvvvzzaa1azvvvvvvvvvvza11111

It might not seem like much, but it is the first cry from a newborn keypad that I hope will be a great addition to the keyboard arena.

Thanks again to metalliqaz, and watch your mailbox for the PROTO_1 boards!

 - Ron | samwisekoi

[samwisekoi]

Prototype PCBs have been put in the mail.  If you sent me your address, I sent you a pair of PCBs.

All PCBs have had the errant trace cut and then tested.

Some of the pins (the ones at the Teensy for Col7-11) were mis-labeled, so I blacked those out.

I can confirm that metalliqaz' latest version (version number unknown to me at this time) of the Easy Keymap firmware-generator works perfectly.  The resulting firmware has been loaded onto the Teensy, and the keypad then tested on both Windows and Linux.

Go forth and prototype!

 - Ron | samwisekoi

p.s.  I advise against mounting the Teensy upside-down on top of the keypad.  However, Retro DSA works wonderfully and has all the 1x keycaps you could want.