Mini Keyboard with Trackpoint

[iMav]

Thanks for keeping us updated lowpoly.  When your vision finally goes production, you can count on several purchases from me and all the free advertising on geekhack you want.  

[bhtooefr]

Quick question, will the switches be PCB or plate mounted?

[lowpoly]

Thanks for keeping us updated lowpoly.  When your vision finally goes production, you can count on several purchases from me and all the free advertising on geekhack you want.  

Thanks, iMav.

Quick question, will the switches be PCB or plate mounted?

I'd prefer without plate for easier modding. Might not be an option with an OEM.

[lowpoly]

I'm currently evaluating zinc diecasting as the material for the case. Machined Aluminum will be too expensive. I don't think I'm even going to look into that. And plastic is not a good base for painting.

[wellington1869]

lowpoly I'm looking forward to putting your new keyboard between two slices of rye bread and eating it.
(You know, cuz its good enough to eat)... so hurry up and manufacture the damn thing already... I'm hungry!

[lowpoly]

I work on your dinner as fast as I can.

[xsphat]

Imight want seconds ...

[lowpoly]

This came in the mail today:



One of three pointing stick modules. These are for notebooks. The stick needs more length and I have to figure out how. Needs to be cheap and stable.

Also got some answers from the Zinc Diecast company. Due to the tooling that is neccessary a production run of 1000 won't make much sense. 10000 would be a number where the tooling cost divides much better.

Cost wise it probably wouldn't matter if the case is made from Zinc, Aluminum or Magnesium. Still waiting for an answer from the Aluminum guys. Zinc is much heavier though. This might be a good thing? I have a feeling that Zinc diecast has a cheap connotation but I might be totally wrong here.

And I have decided that I want this 'board to have nKey rollover, even if it is not aimed a gamers.

I also played a little with the emachineshop software. This is basically a 2D software. I tried to import the case which wouldn't work at all. I don't think you can build this case from scratch because of the many oblique angles. So, emachineshop is a great thing, but not for me right now.

[wellington1869]

emachineshop intrigues me. Its like manufacturing for the masses.
the trackpoint looks neat. I think your prototypes are going to look *awesome*.
Zinc vs aluminum/magnesium - if the price difference isnt much, which one do you prefer? I dont know if I've ever handled a zinc keyboard. Heavy wouldnt be bad if the price is same, but think of the shipping cost you (and the manufacturer) would save on, when you ship 10,000 keyboards, if they're made out of aluminum rather than zinc

[lowpoly]

The pointing stick module even contains the controller. Compare that to the huge controller in the pic above (it's the turquoise pcb). That's the result of 10 years of miniaturization.

I think I'd prefer the "quality weight" of Zinc and the good reputation Aluminum has for cases. I have to figure out how much the weight difference really is. I finished the 3d case but it has some obscure errors so that the volume can't be computed. Will do it again tonight.

[lowpoly]

I ran into something which might be a serious problem with the controller I use. I'm not sure.

When the FN key is pressed the whole matrix is re-scanned and all pressed keys are reported to USB. Also when the FN key is released.

This leads to double key strokes when the FN key is released earlier than the actual key it modifies. Let me try some examples:

Shift, intended mode of operation:

1. Press Shift
2. Press a
3. Release a
4. Release Shift

The modified key is 'bracketed' timewise by the modifier key. This is where the FN key works perfectly.

Now if you type in a rollover style:

1. Press Shift
2. Press a
3. Release Shift
4. Release a

This works for Shift, Ctrl, etc. unless you hold phase 4 so long that the key repeats, in which case you get

Aaaaaaaa

With the FN key however:

1. Press FN
2. Press a (creates 2nd layer key code, for ex. CrsrLeft)
3. Release FN
<------ Here the matrix is re-scanned, a is found as still pressed and forwarded to the buffer
4. Release a

This design is intentional. I see the logic behind it but can't imagine how to work around this in an easy way. It looks like thath with regular modifier keys (Shift, Ctrl, ...) the operating system translates the buffer content to keyup and keydn events by comparing buffer content. With the Layer tables this won't work so easily because the modified key is a different code with FN pressed or not.

I have to talk to the mfgr on the phone I guess.

[lowpoly]

This was my first try. I'm sure I can bring the tooling costs down. Well, somewhat.

[lowpoly]

I'd still be interested in comments on post 220. Am I the only one to 'roll' the modifier keys?

I talked to the controller chip manufacturer and they might change the bahavior. If I buy enough chips, of course. This raises the question how correct rollover behavior is supposed to work with function key layers.

I also have to work on the construction. The prototype is lacking in that department. What I want is a plastic endoskeleton. For one this will support the switch pcb so that no mounting plate is neccessary. And, all parts can be bolted to the skeleton which will ease assembly and disassembly. For ex. if the pointing stick is bolted to the skeleton instead of the switch pcb it will not only have a more stable support it will also make it much easier to exchange the switch pcb.

This skeleton will be a complex part. I don't think I can do this with Rhino. There is no way I can test if everything will fit afterwards. And with the cost of tooling I don't see this as a trial-and-error process. Have to take a look at CAD software.

[lal]

Observing my own typing habits I note that I'm quite a disciplined modifier user, always releasing the modifier *after* the modified key. Maybe I'm not doing it when I'm not looking. But IMHO the sudden appearance of additional characters after using some "rolled" Fn-combo is a no-go and I vote strongly against it.

[lowpoly]

But IMHO the sudden appearance of additional characters after using some "rolled" Fn-combo is a no-go and I vote strongly against it.

This is one of the things that might lead to a situation the Das III has right now. I'm going to avoid it. I don't know what the possibilities of the controller programming are so I have to get back to this later to define the correct rollover behavior.

The 3D parts are done for now except for the famous bezel which was mentioned a few pages ago. Have to put all parts in place first. There is still a lot of work left in the order of weeks. And I'm currently working fulltime on this. Here's a small parts preview:

[lowpoly]

I might have discovered another problem.

When I use the mini on my Thinkpad the Trackpoint requires considerably more force than the built-in one. I played with the settings and the Thinkpad Ultranav settings don't affect the external Trackpoint. I had tried that before but didn't think much, it's different technology and a different manufacturer.

The regular mouse settings however do work for the external TP.

Currently I suspect that the ps/2 to usb conversion translates the TP to a regular mouse. Which would mean no driver could detect it ever.

That would be bad.

[lowpoly]

In the device manager there are two HID devices identified as "ps/2 to usb". That's because the ps/2 to usb converter has two inputs, one for mouse, one for keyboard. Maybe a custom converter might be able to solve that and report the correct device.

[lowpoly]

I probably would have to write a usb driver for the trackpoint to work correctly. And have a custom ps/2 to usb converter for correct identification.

Anyway, got some Trackpoint caps:



The green one as a blue would be my favorite.

Also, I took some pics with the new Ziotec cable which was a gift from xsphat (thanks again, man ):





Stupid cables on the floor...

[bigpook]

way to go lowpoly. your work is awesome. I hope you are able to mass produce these, if so I am in for one.

[wellington1869]

me too. I still want to put it in a sandwich and eat it.

[eugenius]

This keyboard would interest me (enough to consider buing) if:

1. The trackpoint was flawless and had dedicated buttons.
2. See 1.
3. It had at least a FN key.
4. Was completely flat (easy carry, packing etc)
5. It had no metal bottom (that's silly in a portable keyboard).
6. It had quiet, light touch tactile switches (cherry browns, strogmans with cut springs).

PS: Don't get me started ...

[bigpook]

What is the status with this keyboard? Any idea if/when it may be available?

[lowpoly]

This keyboard would interest me (enough to consider buing) if:

1. The trackpoint was flawless and had dedicated buttons.
2. See 1.
3. It had at least a FN key.
4. Was completely flat (easy carry, packing etc)
5. It had no metal bottom (that's silly in a portable keyboard).
6. It had quiet, light touch tactile switches (cherry browns, strogmans with cut springs).

PS: Don't get me started ...

1. What do you mean by dedictated buttons? The pointing stick will have a simplified driver for the first release. Only what's important like speed and scrolling.

2. see 1

3. The swirl key is the FN key. There are two and you can re-program the controller anytime.

4. Completely flat is not possible, you need space for the controller. I think I will drop the additional power supply for the hub and put an additional ps/2 port there. So people can have full NKRO if neccessary. There will be a minimum height for the connectors.

5. The metal case is cool. In aluminum it's not heavier than plastic because the material is thinner. Zinc would be heavier. I prefer Zinc because weight feels like quality (we are talking about a couple of grams here). You want to pick it up and say "wow that feels solid". Its main purpose is not being a travel keyboard although it will travel easily. The main purpose is that you don't have to leave the home row. And the small footprint has several advantages beyond travelling.

6. I want to have a couple of switch variants from the start.

What is the status with this keyboard? Any idea if/when it may be available?

Not before 2010. I will need seed financing which is kind of difficult to get here. Designing the case and skeleton for ex. will require a professional CAD operator.

I'm still working on the 3D. A first informal funding meeting was supposed to happen two days ago but it was postponed.

As soon as the 3D pics are ready I will do a preview (aka "vaporware") site.

[bigpook]

Thanks for the information lowpoly, I am sure I am not the only one that is interested here.

[Mikecase00]

Just stumbled across this.  While my personal tastes would include an editing block and cursor keys, your work is bad-ass and should it become available for purchase in the future, I'd snap one up.  Nice work...

[eugenius]

1. The mouse buttons from that thinkpad below are dedicated buttons.

2. See 1.

3. I though those were the windows / command keys. I suppose I could reprogram any key, but a FN key is a FN key because when you press the combination, the keyboard and not the drivers generate the correct key code, like for arrows, function keys etc.

4. If it's not completely flat, at least not so angled that putting it in a laptop bag beside the laptop scratches both in time. My point is, the more angle, the less portability. A little angle doesn't hurt at all, if the edge is rounded.

5. If it's almost the same weight, the metal it is. I think it will take some kind of alloy, not simple aluminium to keep it light though.

6. The more switches, the better. Light BS, Browns, Reds, Blues, Strongmans - all good.

--------------------------

7. Maybe you can think of a system to let the enterprising portable user stitch some velcro patches on the bottom.  Like some holes in the case, and an easy way to diassemble and put it back. Screws!

8. HARDWARE support for Colemak right from the start, without drivers. Dip-switches, etc. It's easy to move the keys around, especially with Cherry switches. And it would be a great marketing tool. Behold, the bearded woman, err, the first Colemak keyboard. It's QWERTY too, just switch the keys around and flip a switch, no drivers. My bad if there was a Colemak keyboard already.

9 Make your own layout for a ****load of money, with hardware support. Don't make it programmable, just a ask a lot of money for a reprogramming and a keycap replacement.

[lowpoly]

1. The mouse buttons from that thinkpad below are dedicated buttons.

2. See 1.

3. I though those were the windows / command keys. I suppose I could reprogram any key, but a FN key is a FN key because when you press the combination, the keyboard and not the drivers generate the correct key code, like for arrows, function keys etc.

4. If it's not completely flat, at least not so angled that putting it in a laptop bag beside the laptop scratches both in time. My point is, the more angle, the less portability. A little angle doesn't hurt at all, if the edge is rounded.

5. If it's almost the same weight, the metal it is. I think it will take some kind of alloy, not simple aluminium to keep it light though.

6. The more switches, the better. Light BS, Browns, Reds, Blues, Strongmans - all good.

--------------------------

7. Maybe you can think of a system to let the enterprising portable user stitch some velcro patches on the bottom.  Like some holes in the case, and an easy way to diassemble and put it back. Screws!

8. HARDWARE support for Colemak right from the start, without drivers. Dip-switches, etc. It's easy to move the keys around, especially with Cherry switches. And it would be a great marketing tool. Behold, the bearded woman, err, the first Colemak keyboard. It's QWERTY too, just switch the keys around and flip a switch, no drivers. My bad if there was a Colemak keyboard already.

9 Make your own layout for a ****load of money, with hardware support. Don't make it programmable, just a ask a lot of money for a reprogramming and a keycap replacement.

1 On the prototype the trackpoint buttons are directly connected to the trackpoint pcb. Maybe you can see it in the pic. It's the same with the pointing stick samples I have. Should be a non-issue.

3 You can re-program the FN key as well. I think the final version will have three additional layers instead of one so this may come handy.

5 The prototype has a wall thickness of 0.8mm.

6 Not sure about Alps. I haven't looked into it closely but I think there will be no room for plate mounting. Have to check that.

7 There is self adhesive velcro. Why do you want it?

8 As long as you use blank keys, no problem. I have yet to figure out my keycap options. What I found so far doesn't look too good and I'm admittedly procrastinating here. Switching caps around doesn't work with current MX caps.

9

[itlnstln]

7 There is self adhesive velcro. Why do you want it?

If you use the Industrial Velcro, that adhesive does not come off.  I have used in some car audio installs where I had to mount some components under seats on carpet.  I used the Industrial Velcro (I got it from Home Depot) and you are more likely to rip the carpet off the floor than pull the component off.

[eugenius]

1 On the prototype the trackpoint buttons are directly connected to the trackpoint pcb. Maybe you can see it in the pic. It's the same with the pointing stick samples I have. Should be a non-issue.

3 You can re-program the FN key as well. I think the final version will have three additional layers instead of one so this may come handy.

5 The prototype has a wall thickness of 0.8mm.

6 Not sure about Alps. I haven't looked into it closely but I think there will be no room for plate mounting. Have to check that.

7 There is self adhesive velcro. Why do you want it?

8 As long as you use blank keys, no problem. I have yet to figure out my keycap options. What I found so far doesn't look too good and I'm admittedly procrastinating here. Switching caps around doesn't work with current MX caps.

9

3. I hope you get the hardware versus software thing I am saying here, because I don't get what you want to say .. please explain.

6. Ok, I don't think Alps are such a demanded option anyway ... but as a price issue, to me the fake alps sound cheaper, hence bigger profit, or lower loss.

7. I want to use the keyboard directly from a bag for a tablet pc, like this:




The glue is hard to remove, hard to customize the patches of velcro, and I would rather have the option to remove them easily. Plus, the velcro would make the keyboard slide all over the table, If I get it out of the case.

8. I guess you're referring to the different shape for differrent rows. To get some keys that would be swappable hence flat, please investigate the keycaps from flat programmable keyboards (with or without relegendable keycaps) that use the MX switches - Tipro makes some for example. Just a suggestion ...

[lowpoly]

3. I hope you get the hardware versus software thing I am saying here, because I don't get what you want to say .. please explain.

6. Ok, I don't think Alps are such a demanded option anyway ... but as a price issue, to me the fake alps sound cheaper, hence bigger profit, or lower loss.

7. I want to use the keyboard directly from a bag for a tablet pc, like this:

Show Image

Show Image

The glue is hard to remove, hard to customize the patches of velcro, and I would rather have the option to remove them easily. Plus, the velcro would make the keyboard slide all over the table, If I get it out of the case.

8. I guess you're referring to the different shape for differrent rows. To get some keys that would be swappable hence flat, please investigate the keycaps from flat programmable keyboards (with or without relegendable keycaps) that use the MX switches - Tipro makes some for example. Just a suggestion ...

3 The FN key doesn't have a special pin on the controller or something. It is just another key in the matrix. You can use any key you want when re-programming the controller. It works without a software driver then.

7 Got it now. Where would the Velcro have to be? I'll probably leave the two inner rubber feet away in the final version.

8 Cherry has these as well. I just didn't have them on my list yet. There are already a lot of keycap variants I'd like to have (white, black, with and without letters, ANSI, ISO).

[eugenius]

3. Got it.

7. Actually, I thought of a lot of attachment systems, I don't know what would actually be the best ..

So far the best  option is still (almost:)) permanently attached velcro patches with matched rubber velcro feet.

I'm trying  to come up with a more elegant and easy to use solution that leaves nothing permanently sticking to the bottom ... but from an usability standpoint, that's the best  so far.

[itlnstln]

The "Industrial Strength" Velcro will not unfasten.  Great stuff.  Regular-a*s Velcro sucks.

[lowpoly]

I already used Velcro to attach one of the smaller pcbs. :smile:

The feet are 1mm high. That's the limit. Using screws is probably the best solution.

Just stumbled across this. While my personal tastes would include an editing block and cursor keys, your work is bad-ass and should it become available for purchase in the future, I'd snap one up. Nice work..

I missed your post. Thanks.

Another difficulty came up which is EMC testing (electromagnetic compatibility). To get that nice little CE sign. Looks like the test is expensive and there's no way around it. There's even the possibility that you fail it again and again. So commercial pcb design is not a garage operation anymore.

[keyb_gr]

Actually you can slap on a CE sign if you're confident that radiation will be within the allowed limits. It's just that you're liable if someone complains about EMC issues and the device is found to be non-compliant.

It can't be that strict really. There are many CE-marked devices on the market where you really wonder whether they ever saw an EMC test. Some cell phone chargers seem to be pretty nasty (even by the standards of small switching power supplies), apparently still being able to generate heavy interference in the FM broadcast band.

A keyboard should be a relatively uncritical component, no huge currents or anything - in my experience, it's the data transmission that's most critical (oldschool edgy 5Vpp signals, no surprise there). If you keep half an eye on potential EMC issues during development and do some testing with a decently sensitive LW/MW/SW/FM receiver (DE1103, E5, ICF-SW7600GR or such) without any indications of excessive radiation, you should generally be fine. Of course it would be advantageous if you could afford a 2-layer PCB, close-by bypassing of the controller (and following other guidelines for RF-level design) would also seem like a good idea.

[lowpoly]

Actually you can slap on a CE sign if you're confident that radiation will be within the allowed limits. It's just that you're liable if someone complains about EMC issues and the device is found to be non-compliant.

A recall of the whole series might be neccessary/forced then meaning $$$.

It can't be that strict really. There are many CE-marked devices on the market where you really wonder whether they ever saw an EMC test. Some cell phone chargers seem to be pretty nasty (even by the standards of small switching power supplies), apparently still being able to generate heavy interference in the FM broadcast band.

Some non-european manufacturers just print it on the label because it is almost impossible to sue them. Inside the EU that's different.

A keyboard should be a relatively uncritical component, no huge currents or anything - in my experience, it's the data transmission that's most critical (oldschool edgy 5Vpp signals, no surprise there). If you keep half an eye on potential EMC issues during development and do some testing with a decently sensitive LW/MW/SW/FM receiver (DE1103, E5, ICF-SW7600GR or such) without any indications of excessive radiation, you should generally be fine. Of course it would be advantageous if you could afford a 2-layer PCB, close-by bypassing of the controller (and following other guidelines for RF-level design) would also seem like a good idea.

Thanks for the additional info.

I guess the EMC issue explains why the Cherry switch pcbs have a 2nd copper layer on the top side?

[keyb_gr]

I guess the EMC issue explains why the Cherry switch pcbs have a 2nd copper layer on the top side?

They do? Looking at my pics, yep, that seems to be a dual layer PCB "upside down". Such a ground plane is quite handy as you always have a low-impedance ground nearby. (Larger arrangements should implement star grounding in addition.)

Indeed, the nice large keyboard matrix would probably make a fairly decent antenna without any precautions.

[chimera15]

Beautiful thing. You don't have any arrow keys though?

[lowpoly]

:smile: They are on a function layer. You press one of the function keys with your thumb and a cursor key on/near the home row.

[lowpoly]

Preview of the 3D model I'm working on for some time now:

[pic removed, final 3d on page 20]

The glossy keycaps are not going to happen but I thought it would be nice.

[D-EJ915]

I like the blue colour of the switch things, should get some keycaps that colour...

[o2dazone]

is that ambient occlusion I see? Looks pro man, love it

[mtl]

I like the concept.  Have you considered adding the middle button?  It makes scrolling with the trackpoint so much easier!

[pmyshkin]

This is a great design, and it's amazing that you're actually working on getting these mass produced.

My only concern is that you completely removed the grave/tilde key, and this key would be important to *nix users. I've seen some keyboard manufacturers try to put this key in the place of the left logo key, and that's just clunky. The only good solution I've seen is on the Sun Type 6 keyboards that split the long backspace into grave/tilde and backslash/pipe keys and put the backspace in the place of the original backslash/pipe position right above enter.

Also, I'm sure people have brought this up, but you could replace capslock with control and both control keys with FN keys.

[bigpook]

The white one looks incredible. When can I get one : o

[wellington1869]

The white one looks incredible. When can I get one : o

seriously low-po, send it to market already!

[lowpoly]

Thanks for the comments.

is that ambient occlusion I see? Looks pro man, love it

It's a skydome. Didn't work as well on the keys as in the car scene I re-used for this. Diffuse illumination is baked so a 2048x1024 image will render in under a minute. This will finally allow me to render 360 degree animations for various color combinations.

And it better look good, 3D was my job for several years. :smile: Lighting and materials are still temporary.

My only concern is that you completely removed the grave/tilde key, and this key would be important to *nix users. I've seen some keyboard manufacturers try to put this key in the place of the left logo key, and that's just clunky. The only good solution I've seen is on the Sun Type 6 keyboards that split the long backspace into grave/tilde and backslash/pipe keys and put the backspace in the place of the original backslash/pipe position right above enter.

Also, I'm sure people have brought this up, but you could replace capslock with control and both control keys with FN keys.

Grave/tilde is on Fn/Esc. I'll probably keep the large Backspace key to keep the layout standard. There are enough keys for reprogramming. I personally do not like Backspace above Enter because I sometimes hit Enter instead of Backspace. But everything is re-progammable.

I already have Ctrl on CapsLock. Didn't want to do any non-professional keycap text changes though.

The FN keys are left and right of Space.

The white one looks incredible. When can I get one : o

While the initial idea for the 3D visualization was to try out the bezel, my main motivation to keep it going was the white board with the blue pointing stick.

As for a time frame, I have no idea. When the 3D model is semi-finished I'll continue with a preview/blog web site.

[lowpoly]

I like the concept.  Have you considered adding the middle button?  It makes scrolling with the trackpoint so much easier!

Missed your post, sorry. I'm undecided on the third button. ATM I'd rather use one of the keyboard keys for that (or for right click) or both buttons together. Have to think about it.

Got some trackball samples:



Unfortunately they do not roll smoothly but rather click-click-click. This does not work for cursor movement on a big screen. I'm talking to sales if they can do something about it. Pressing the ball down is like a button click.

[bigpook]

seriously low-po, send it to market already!

Ditto. Can I get mine in white?

[lal]

Missed your post, sorry. I'm undecided on the third button. ATM I'd rather use one of the keyboard keys for that (or for right click) or both buttons together. Have to think about it.

Three mouse buttons is a MUST.  I didn't realize until now it has only two.  Middle button is the standard way of pasting the clipboard content in most Unix systems, and it's used for many other kinds of useful things.  It wouldn't be quite right if it had only two.  Just my 2¢.

[lowpoly]

Wait..what..a trackball in place of the trackpoint? Hmm I guess it indeed would get a little bit too small. Or is my space bar splitting trackball dream coming true?

I'd like a third button/scroll wheel combo. My only concern is that adding things deflate the minimal design. Which is really stunning! Wow (I'd prefer matte black though).

I talked to a supplier about pointing sticks. Turned out he sells these trackballs as well. They're small enough to integrate them into the the space bar. The way they are it wouldn't work right now though.

When I try to be honest with myself, I think that three of these buttons will clutter the design. Somebody else here said lately function shouldn't follow form on a device like this. Which is true. Maybe I'll split one button. Or add another much smaller one. It's not possible to space three mouse buttons evenly because of the MX switch. I have to postpone this for now and will pick it up when the preview website is finished.

The 3D design is very iPod-esc, that is, the graphical marketing efforts. I know this isn't possible (small series and all), but I'd love to see that in a wide rainbow of colors. They would look great, side by side. Then even girls would buy one. The pink and such.

Not sure if different colors for the case will be a problem. One of the reasons I spend so much time with the 3D is that I can use it to generate plenty of images for a web configurator.

Keycaps is another thing. Cherrys will be available in black and white only. The ipod similarity was coincidence caused by keycap choice and unfinished lighting/materials.

Ditto. Can I get mine in white?

I'm working on marketing issues right now because I realized that I will need third parties to do the CAD and electronics works. Which requires funding. Which will be easier with the materials I'm working on right now.