Ultra mini wrist mechanical hack

[chimera15]

Using the same techniques I used to build my steampunk mini:
http://geekhack.org/showthread.php?t=6313&highlight=steampunk

I cobbled this together using the wrist keyboard I bought here:

http://geekhack.org/showthread.php?t=7598

The existing keyboard was very difficult to type on.  Now it's possible to touch type on it using one or two hands.  It's a bit difficult getting used to pressing numlock and caps lock before and at the end of the sentence still for a period, but otherwise it's not bad.  My speed is only slightly decreased.

[chimera15]

Oh I forgot to add, I stole all the switches and the board from a dell at101w, so it uses black alps switches.  I still need to find some better keycaps for it probably.

[chimera15]

lol Well the idea is that you can wear it and walk around and still use it.  This keyboard has a glove that you put on, then it attaches to the glove with the velcro.  Usually I have it attached to velcro to my tx2500z.  It's nice cause you can basically velcro it anywhere and not worry for room, but still touch type.  All the other keyboards this size have keys that are way too small.  I would guess this is the smallest full sized full format keyboard possibly ever made, and definitely mechanical.  It only has 40 keys.

I suppose you could velcro it to another body part, or just put it in a specially made pocket or something and it'd work so you could free up your other hand.

Since you can now touch type with my version which was very difficult with the original since it used rubberized keys that were prone to double stroking and the keys were sunk into a metal frame, I suppose it opens up putting it on different parts where you wouldn't need to see the keys, like maybe on a lower hip or something.

[chimera15]

This?
http://blog.russnelson.com/2006/Mar
http://www.chordite.com/

Yeah, there are a lot of one handed keyboards, but they all use shift keys and different things that slow down typing.  This is full format, full keysize, just like a normal keyboard.

I see what you mean though, that's pretty cool how he made it so you can hold it and type on it.

That's a pretty good idea, and looks like I could do better with the construction of it.  I wonder how he shifted the keys though, and what kind of software/hardware he used.  He said he got up to 33 wpm, not bad.

The main problem with all of those though is that 8 keys don't equal 40 or really more.  Even this one has a pretty bad layout since it doesn't have a shift key, you have to press caps lock as far as I can find, and you have to press function keys instead of shift to get the punctuation.  That's bothersome enough, but I couldn't imaging having to shift to multiple profiles to get to different letters.

[chimera15]

Ah cool, he tells you how to build it!!! Coooool.  That might be my next project.

[chimera15]

These seem to be the ones.

Looks interesting:
=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b]http://www.marquardt-switches.com/tact-and-key-switches.html?&tx_productkatalog_pi1[mode]=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b


http://switches.com/
http://blog.russnelson.com/2006/Mar

It's at the bottom of the second link.

The board looks more interesting as it's apparently very programmable.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=AT90USBKEY-ND

[chimera15]

Looks like the board is out of stock everywhere, at least a month long lead time if they can even get them, and for one 1.

[JBert]

This?
http://blog.russnelson.com/2006/Mar
http://www.chordite.com/

What, no portable Datahand to be found? :-)

Wonder what switches he used - doesn't seem to be on the website.  I'm always on the lookout for low force momentaries smaller than the Cherries.

This makes me think about Tim Tyler, he did some research about those little buggers - even to a point where a keyboard would become barely usable.

[chimera15]

That's true, datahand switches or something like them approximating their layout would solve the limited switch problem.  The datahand has 50 switches? 5 for each finger?  You could almost use a hhk chip for that.

[chimera15]

Really though, thinking about this, what's been done by them isn't much different from an n52, or maybe even an n50, which you would be able to approximate that same function with fairly easily.  You'd just have to pull the n50 apart and construct a holder for it and it'd do the same job.  The only real hard part would be coming up with a matrix that would work with the software.  I think the n50 only has 3 profiles, and basically 14 buttons, so that's 42-3, 39 possible key presses.    I guess it could be done, if it has 4 profiles like the n52, then it could have 53 combinations which then would be enough.

[chimera15]

Yeah I think it would be fairly easy to do this with an n50.  You could even do Ripster's mechanical hack to it too.  That might be pretty sweet.  

The only thing that might give me pause is that when I use an n50 intensely with windows xp I get an occasional bsod.

[JBert]

It is a very simple concept really. And great too, because there is no friction etc. but still a tactile feeling and at very low forces.



In the middle of the rectangle you see a tiny magnet. The opposing magnet is on the key cap (removed here).  On both sides of the rectangle are an optical receiver and transmitter, like you see in mice from that era. As you move the keycap in the direction of the arrow, separating the magnets gives the tactile feeling, and leaves a gap (the rectangle) so the receiving sensor gets a signal.

How does the push-button work though, is this also a "magnetic/optic" keyswitch?

[chimera15]

So I'm working on a new ultra mini based on the i-go bluetooth ultraslim portable keyboard.  I ruined the matrix on it though.  The matrix was only on a plastic sheet so it can't be soldered to.  Is there a rule of thumb or a logic to redoing the wiring on a matrix anyone?

[chimera15]

I tried to trace a couple of the circuits but it's like a mess to me.  I guess I'll try to look at Ripsters n52 hack again.

[chimera15]

N-key with diode matrixes tend to be pretty straightforward.  Others tend to be a mess.

hmm, so the n52 has a relatively easy matrix to this one?    I thought of a new way of thinking of it though.  If you trace the matrix from the switch, you can try to find the fastest way for the two contacts back to the board.  I think I may be able to trace them all that way.

As long as the length of the wire doesn't matter to the board I think I should be ok.....  How does a chip determine what position the key is in the matrix.  How does it receive a signal if there are other switches that are open in the matrix?  Is it like one wire is a ground that goes to them all,  so like this board has 17 contacts, it has 1 ground and 16 lines, or somehow else..  I don't understand the basics of how a matrix is layed out for a keyboard to work.

Oh I get it, so it's like 2 wires in parallel, with like 20 switches between 2 wires, but then how does the chip decide which switch has been closed?  If it's the length of the wire, or resistance it has I might be in trouble.

[chimera15]

I know it simply can't be resistance or length of wire or my steampunk mods wouldn't work cause I solder huge lengths of wire in some case to extend switches to different layouts..  It has to be when the key is pressed the chip receives signals on certain wires of the matrix that determine it right?  To me right now though it looks like every switch is wired to every other switch on the board practically... maybe I just need to trace it better

[chimera15]

Ah thanks.  Figure 2 explains it to me.  There are multiple ground lines, and then multiple tree lines with lots of branches.  So a combination of a tree and a ground line equals the position.  Thanks a lot.

I think I can figure it out based on that.  I just need to determine which lines are the grounds, then run those out to the switches, then find the single line that branches on one side of the switches and run one line with lots of branches off to the switches that way.  I think it should be relatively simple.  

The ground line will be on the opposite side of the first switch, so I should be able to figure it out just counting how many branches the ground line has off it vrs what looks to be a tree.

[itlnstln]

Yep - that's it. It's all in an attempt to reduce controller pins.

I'm glad they picked a good reason.

[chimera15]

lol  

It looks like every tree line has about 5 switches, so far so good. I'm waiting for more wire as I ran out, but I think I should be able to figure it out.  I'm using different colored pens to highlight the switches from each line.

[john_3000]

These seem to be the ones.

Looks interesting:
=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b]http://www.marquardt-switches.com/tact-and-key-switches.html?&tx_productkatalog_pi1[mode]=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b


http://switches.com/
http://blog.russnelson.com/2006/Mar

It's at the bottom of the second link.



Just for the record, Russ likes those Marquardt switches much more than I do.  For a Chordite I recommend microswitches with the little levers, not regular keyboard-type switches.  Some are discussed here and here
(circa 2005)

[chimera15]

These seem to be the ones.

Looks interesting:
=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b]http://www.marquardt-switches.com/tact-and-key-switches.html?&tx_productkatalog_pi1[mode]=overview&tx_productkatalog_pi1[modifier]=0&tx_productkatalog_pi1[value]=6450&tx_productkatalog_pi1[pointer]=0&cHash=c2a011b15b


http://switches.com/
http://blog.russnelson.com/2006/Mar

It's at the bottom of the second link.



Just for the record, Russ likes those Marquardt switches much more than I do.  For a Chordite I recommend microswitches with the little levers, not regular keyboard-type switches.  Some are discussed here and here
(circa 2005)

Wow cool, are you one of the guys that made those?  Do you know if this board would work for it?

http://pjrc.com/store/teensypp_pins.html

Would the software have to be rewritten or something?

A bluetooth board like I'm working on now would seem to be more ideal for a hand keyboard like that wouldn't it?

[john_3000]

Wow cool, are you one of the guys that made those?  Do you know if this board would work for it?

http://pjrc.com/store/teensypp_pins.html

Would the software have to be rewritten or something?

A bluetooth board like I'm working on now would seem to be more ideal for a hand keyboard like that wouldn't it?

When it comes to Chordite keyboards I am definitely the guy.  I didn't know about the teensy boards.  Holy smokes.  Thanks.  I will definitely check it out.  First guess: the firmware will run without modification on the ~646.  If not it should be easy for me to fix it so it does.

My problems with Bluetooth are (a) I can't find any specs and (b) I worry about security.  The link between keyboard and host is not something you want accessible to strangers.  But in general you are right: anything wearable is going to benefit physically from wirelessness.  In the field wires and connectors break

[chimera15]

I got tired of waiting for the good wire I ordered, so I went at it with some junky stuff from radioshack.  After 3 days of frustration and tracing paths I finally got it to work.  For a while I thought i had blown it because I couldn't get it to pair.  It turned out that one of the crimps I used didn't connect to one of the wires.  I had to track it down with a multimeter.  Anyway, I think it might be the first bluetooth ultra mini mechanical keyboard ever.  The base keyboard was hacked out from my zenith model f with green alps, that I didn't really use or like, then I used the remaining real clicky whites from a Focus board that I used to modify one of my siigs with.  Should be an awesome keyboard and will probably become my main driver I hope if I can get used to having to use a separate key to activate the numbers.

[roadblock2thesun]

@chimera,

What was your process for soldering the leads to each switch, your welds look nice and neat. I'm having a hard time holding the wire, the solder and the iron all at the same time...

[chimera15]

@chimera,

What was your process for soldering the leads to each switch, your welds look nice and neat. I'm having a hard time holding the wire, the solder and the iron all at the same time...

You tin the wire first, and the switch.  Easy with stranded stuff, never use solid wire.   You can prop the wire or the solder up, or bend it so it's easily accessible.  Once there's a good amount of solder on the wire, then all you hold is the wire and the soldering iron against the terminal.

[chimera15]

I got tired of waiting for the good wire I ordered, so I went at it with some junky stuff from radioshack.  After 3 days of frustration and tracing paths I finally got it to work.  For a while I thought i had blown it because I couldn't get it to pair.  It turned out that one of the crimps I used didn't connect to one of the wires.  I had to track it down with a multimeter.  Anyway, I think it might be the first bluetooth ultra mini mechanical keyboard ever.  The base keyboard was hacked out from my zenith model f with green alps, that I didn't really use or like, then I used the remaining real clicky whites from a Focus board that I used to modify one of my siigs with.  Should be an awesome keyboard and will probably become my main driver I hope if I can get used to having to use a separate key to activate the numbers.

Show Image

Oh I forgot to add, it's Christmasy too. roflol

[chimera15]

I just got one of these mini bluetooth keyboards:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=150386010210&ssPageName=STRK:MEWNX:IT



I was surprised to find out that they're conductive rubber over a pcb.

These are much better candidates for mechanical bluetooth keyboard hacks if anyone is interested in doing one in the future.  Especially as they're full layout unlike the igo.  They're also significantly cheaper.

[chimera15]

With one of those it should be easy enough to convert a siig minitouch to be wireless effectively, or make a new one.

[chimera15]

[john_3000]

For the record I have received 2 teensy boards, modified the USB Chordite code to run on the teensy's atmega32u4 processor and downloaded the hex file to a teensy board.  It enumerates as it should: a combined  keyboard and mouse (a USB compound device).  To go farther, i.e., to see if it actually works, I need to connect it to switches.  That will take a bit longer but I don't foresee any big problems.   I want to try a slightly different mechanical design.  I'm aiming at a DIY prototype that folds to go in a pocket.  The teensy board makes that possible.

[chimera15]

For the record I have received 2 teensy boards, modified the USB Chordite code to run on the teensy's atmega32u4 processor and downloaded the hex file to a teensy board.  It enumerates as it should: a combined  keyboard and mouse (a USB compound device).  To go farther, i.e., to see if it actually works, I need to connect it to switches.  That will take a bit longer but I don't foresee any big problems.   I want to try a slightly different mechanical design.  I'm aiming at a DIY prototype that folds to go in a pocket.  The teensy board makes that possible.

Oh wow. That sounds amazing.  Are you going to make the software/code available?  You should start your own Chordite thread here?

[john_3000]

Oh wow. That sounds amazing.  Are you going to make the software/code available?  You should start your own Chordite thread here?

Yes I'll publish everything.  Since my last note I've tried grounding some individual input pins and the teensy does produce the  appropriate characters on the screen so there's a reasonable chance that the code already works.  You could have the hex right now if you needed it but it's probably better use of time to wait for collected instructions and the source

[microsoft windows]

I just got one of these mini bluetooth keyboards:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=150386010210&ssPageName=STRK:MEWNX:IT

Show Image

I was surprised to find out that they're conductive rubber over a pcb.

These are much better candidates for mechanical bluetooth keyboard hacks if anyone is interested in doing one in the future.  Especially as they're full layout unlike the igo.  They're also significantly cheaper.

How bad do they feel?

[chimera15]

How bad do they feel?

Well, touch typing is basically impossible with it because of size and squishyness.   It's ok as a thumb board, but you're not going to get any speed out of it.  They're a lot like those flexible water proof keyboard's if you ever got one of those, there's nothing hard anchoring the key, so they're squishy, and fairly hard to press.  The board itself is a masterpiece of pcb making and should make an excellent full sized compact mechanical if I can find the right raw board to hack into it.  If the scorpius m10 that I have coming can't be hacked apart to make it, then I may end up just taking apart a siig.

[The Solutor]

It's ok as a thumb board, but you're not going to get any speed out of it.

I'm sure that clouds of teenagers will disagree with you...

The world record on the iphone, w/o any tactility is 82 wpm...