Conchi build log: 8g microswitch ergo keyboard and trackball

[intrio]

WHAT
I'm building a custom ergonomic keyboard: its wireless, will work with any device ( android, ios, mac, pc), split yet no cable, low cost $250 ish, somewhat inspired by a myriad of inspiring alternative keyboards: touchstream, kineses, ergodex and DataHand, maltron.

It's named the 'Conchi' after it's conch shell like appearance + HCI (human computer interface).  initial pass is raw material, exposed wires, polished style as permitting.

WHY

More

The core motivation is to help combat my chronic tendonitis/tendonosis, acquired being a software developer consultant with long hours, fast typing speed, and enable me to continue working longer as still how I support myself

There are 2 phases of this project:

More

1) short term, the primary goal of this project is to replace the BTC 'mini multimedia keyboard', the model I've been using for about 6 years (going through one every couple years), and keeping the same 86 key (tenkeyless) layout that has been baked into my muscle memory, so I can keep working without so much effort learning a new layout, and simultaneous combating tendonitis/RSI. ASAP.   In a typical workweek I  spend a fair amount of time recovering from using even this relatively low force (60-65g) keyboard, and other switches are only marginally better, while 8g is in a league all it's own, being less than the typical force a finger puts at rest.

2) longer term, an experimental DataHand like approach, with thimble like fingerwells for minimum travel in varied directions. With 5+ keys per finger, likely using pressure sensitive rather than switches.

KEYs using alternative micro SWITCHES

More

This is one of most important parts.  Our fingers receive a tremendous amount of work (abuse?) from hammering these tiny nails into the keyboard over vast distances, probably the equivalent of thousands of miles of running over a decade.   I'll be using super low force subminiature microswitches, that have 1/8th the force at 8g of typical 65g+ force keystroke activation, and have it modelled after my hand and range of finger travel.

 I have done significant research on keys before selecting micro-switches (inspired by Tim Tyler's experiments http://mykeyboard.co.uk/microswitch/ though we differ in approaches).   

 About a decade ago having my first flareup of tendonitis that didn't go away after a weekend, I tried every keyboard at every computer stores in the area, before I decided on the BTC which was accessible at Fry's electronics.   Prior to that, I used an exotic Fingerworks Touchstream for years, which although is 'zero force' but like hitting the flat screen of an ipad, it really isn't ergonomic, particularly on the back of the hand having to keep the hand above the surface from touching it accidentally, as well they are finicky, break and now can't be repaired.  I have played with light keyboards (laser, lightio), virtual keyboards on ipad/android tablets, my fingers hurt using them for any prolonged use as well as the lack of feedback are deal killers.  4th generation haptics will eventually improve this, but there needs to be a full feedback loop of finger hitting the switch but not activating it, intital resistance pressing to start activation, a breakaway as it's activated with a progressive break to avoid hard bottoming out, and ideally a restorative force (like a trampoline or piano hammer weighted key) This isn't really possible without a physical mechanism.


To explore, I ordered a sampler of Cherry MX and some other keyboard switches,  to compare in a real world test setting.

• 0 force optical proximity sensors
• 8g, 10g, 16g, 20g Omron subminiature lever microswitches.
• 35g Qateron Clears,
• 45g+-65g Cherry Switches, linear and tactile
• 60-65g scissor action and the cheapo BTC mini 86 keyboard that has been my workhorse for years,

I measured their activation force using a kitchen scale a bolt loaded up with washers till it could no longer statically support the load.

Next I rotated through them hacking into a Kensiginton 4 button trackball,

primary goal replacement of my left and right mouse button, to figure out what worked the best, as the built trackball clicking is very heavy like 90g+ force, and that is one of the biggest pain points I have, the right thumb, click, double click, click to hold and drag, select constantly throughout the day, which sometimes wakes me up at night from the pain. This 1 minute tour  capturedvideo.MOV (6035.1 kB - downloaded 145 times.) shows the layout, all four buttons ended up being used, 3 on thumb (click, double click, middle click), one on pinky ( right).

That revealled I liked the 8g-16g Omron most, and found 20g too hard for ideal, and the 8g so far are working the best for me in the trackball in a side clicking manner, even though it's most prone to accidental clicks placing the hand in it.  The optic proximity sensor worked but had the issue with you can't rest your finger near it, creating a different form of work being done, and there is no feedback loop as to where/when you are/click.

The Omron microswitches are available at very different forces, in the same standard size, and a couple different ways to actuate it from a plunger, to a lever, to variations on the lever (roller).    These are considered subminature, are smaller and longer than a normal key switches, and the layout I'll do will end up being more compact likely than a traditional keyboard, but fatter columns, so that more macros can be accessed.  They run about $1.80 to $2.50 each depending on who you order from and what quantity.   I ordered 150 of them from Newark in the UK as they have significantly cheaper pricing than Mouser, Parts Express.

150 switches for:
86 keys for the first conventional one
50 for the datahand like one.
extras for things that go wrong or hacking.

So far I won't need key caps, the lever is out of steel on these microswitches is big and smooth enough to work well.  Though I am experimenting with a dab of silicon to glue the lever to the switch as it can ocassionally lift up which makes it feel ...wrong.

CONTROLLER

More

Initially I'll be using some cheap $29, Bluetooth keyboards off alibaba hacked apart as bas, if that doesn't work or for a more powerful version I'll be using two Teensy 2.0++ one for each hand as I want to be split handed, and direct wire to avoid ghosting/need for diodes, lots of unnecessary soldering, and matrix complexity, plus it has USB mouse and keyboard emulation and is trivial to program ...first key microswitch triggering os key or mousem example took about an hour, having  never used an arduino before.

I'm excited by the wireless and battery powered potential, to help increase ergonomics by easily repositioning it, vs being locked into generally arms in front which has lead to chronic pain aside from fingers, and taking it on the road, with whatever devices I have on hand or are to come ( unlike ps2)

I had considered using one of my existing defunct keyboards or getting a wireless keyboard just for the controller+wireless+battery, but the mini keyboards are almost monolithic in construction with dual sided ribbon tape and tight traces that seem overly complicated to hack ( though after trying several found an ok candidite for a basic version), as well there is flexibility in Teensy code to expand to mice/footswitches, chords etc. e.g. in a finger well with N, S, E, W buttons, pressing N+E in near the same time might translate to a single keystroke of NE. Or two fingers both pressing E at the same time might be copy, and 2 W might be paste, similar to the touchstream's cool pallette of these common operations.

WIRING

More

I'm going to use standard twisted wire hand soldered direct to the teensy 2.0++, and skip the PCB board as I'm going for curved, quick and dirty as I know these projects are highly iterative, you don't know what's needed exactly till actually using it.  I hope my soldering skills are up to the task   Thankfully as it's only 86 keys it's not that many solder joints actually.

I'll be using a breadboard for the first pass.

MOUNT

More

I'll be using Polycaprolactone (hot water moldable thermoplastic) aka Shapelock, Instamorph, I've build many grips/mounts out of this material and it's strong flexible easy to rework with a hot air gun, this latter comes in super handy as you find the thing needs to be just slightly different.  As a bonus, the switches seem to bond with the plastic when it's hot, without damaging them, so that is how I'll mount the microswitches in the curves sculpted after my finger/hand travel.

This will accompany a zero gravity recliner "throne" I'm making out of Kee Klamps, and memory foam. To accompany the existing  sit, stand, lie,  treadmill, whatever position, pipe desk with 4 monitors and LCD monitor arms.

---------------------------------------------DEV LOG SUMMARY ------------------------------------------------------
3/20/2015
At this stage i have ordered the parts, which should arrive in the next week or 2.

[Bucake]

cool project

[suicidal_orange]

This should probably be in Making stuff together! but this is a close second

Interesting project, those are some seriously light switches but sadly it sounds like you need them.  I'm very interested how the polycarbonate case turns out - hadn't thought about using it but could be ideal for custom builds.

[Huxley2500]

Nice project.  Keep us updated.

[intrio]

So got the polyprolactone today and started playing/sculpting some parts. Which are going to look organic, perhaps geigeresque.

One I'm calling the "conchi" (conch shell + CHI (computer human interface)), which is a conch looking, where the grooves are made from my fingers movement.

The other is a exploration on how many directions the thumbwell can reasonably hit without confusion. at least 6, likely 8-10 for me (which will also differentiate it from the datahand).  In using the trackball for awhile now, I've very happy with the new switches+layout. Very intuitive, and could easily handle more buttons for different things.

Not shown is the arm rest made out of the same material, that now has the previous shown trackball from earlier attached to it, to help align the hand/wrist, and is molded after my forearm.

I also got some mini joysticks to try out, as alternatives to the microswitch, but both the ones I got (sparkfun) are way to high in force, and not so easy to dissaemble.

[intrio]

2x teensy++ 2.0 arrived today, still waiting on wire and switches

[intrio]

This should probably be in Making stuff together! but this is a close second

Interesting project, those are some seriously light switches but sadly it sounds like you need them.  I'm very interested how the polycarbonate case turns out - hadn't thought about using it but could be ideal for custom builds.

thx, any idea on how to move the thread to that? i don't see the option.   It's actually polycaprolactone not polycarbonite, but is great for builds, though it does stain easily

[KHAANNN]

I was amazed by that thermopoly material, it's certainly seems fun to play around with, and it seems practical for many uses, for example I was going to 3d-print some caps for naked fence-wires a while ago, but it was too much trouble so I postponed it, the thermoplastic material seems ideal for the job, just apply a glob and lock etc.

However, for a keyboard, it doesn't seem logical

Since you're in US and likely have all access to all sorts of stuff easily, I would suggest getting a 3d-printer, and start using an easy tool like tinkercad.com - you could more easily build a more solid keyboard/trackball solution that's actually usable and visually pleasing (visually pleasing if you calibrate it right)

I decided not to pursue my 3d-printed keyboard project, as even relatively large 3d-printer don't have enough area for an 60% plate, yet after seeing a guy at reddit print a keyboard in 2 pieces, I have to say I'm tempted, a multi-piece/constructed keyboard makes sense, but the plate part is challenging, yet you don't have that worry, as 50% of an ergo keyboard is likely smaller

[intrio]

I had though about 3d print and cnc, but I don't exactly know what I'm doing yet... too much new/unknowns, so don't think I'd be very good modelling it.  The material is plenty strong similar to nylon, slightly flexy, and can be cut, sanded drilled and dyed, smoothed so on par with 3d printer plastics.

consider this phase a sketch/prototype with the ability to be easily reworked as I find out what's needed from actually using it, that ultimately can be redone properly to look better.  Having built other things before, I've learned that these sorts of human interface projects take several interations (3+ a day), and 3d printing is just too slow/expensive, by the time it takes to model a complex curve surface, let alone print, I can have a real one out of this plastic, actually in use, <20 minutues

also This is something I need to continue working, so am not focusing on looks this pass, and it will morph as additional parts come in, wires will dissapear etc.

[ctm]

Wow that's really cool! Looking forward to updates!

[intrio]

300 omron low force microswitches arrived today, so i am unblocked from proceedeing, when time permits

[intrio]

this is a shot of the zero gravity chair workstation that accompanies this project, the trackball on the right ...lol pic upside down for reasons i don't know

i've tried at least 4 different wrist forearm alignment molds,  to keep wrist angle neutral, made from plastic to foam, my conclusion is that you really need a adjustable one that changes throughout the day, so thinking of a ball joint, or heavy velcro/ magnet

as well i've tweaked the buttons on the trackball a couple times, by using a heat gun to move a millimeter and few degrees, that helps tweak far easier than if it was on a pcb board, which supports my theory that scultping thermoplastic is a better choice than 3d printing for first passes, when going so unconventional like this.

so far after a few weeks of use, the new microswitches and positioning, my mousing hand has substantially less stress on the thumb, its not waking me up at night from pain yay!

[intrio]

here's a experiment in user (aka me) centric design, if the palm was fixed, what possible range of motion could be reached by each finger, this might be packed into many more possible keys beyond a normal key, mapped to macros, which might further reduce effort from the many multiple key combos.

its not quite accurate together like this as finger range overlaps.

thumb in particular has a wide area, that might even have multiple layers, inside and outside

[KeypressGraphics]

Awesome project dude, I'm really looking forward to seeing what kind of device you end up with.

[infiniti]

This should probably be in Making stuff together! but this is a close second

Interesting project, those are some seriously light switches but sadly it sounds like you need them.  I'm very interested how the polycarbonate case turns out - hadn't thought about using it but could be ideal for custom builds.

thx, any idea on how to move the thread to that? i don't see the option.   It's actually polycaprolactone not polycarbonite, but is great for builds, though it does stain easily

I've moved this to the making stuff together! sub-forum.

Interesting project you have here.  Keep up the good work!

[intrio]

I've moved this to the making stuff together! sub-forum.

Interesting project you have here.  Keep up the good work!

double thanks infiniti


MOD NOTE: you're double welcome! #powah

[IdentityNon]

I really, really want to see where this goes.

[intrio]

so looks like using coat hangers and stringing on the switches like beads will be a viable approach. This is nice as the coat hanger wire is reasonable stiff, cheap, accessible, easy to bend to make the curve, easy to reposition the keys on it, add or remove them quickly nondestructively.

Conventional keyboards have 5 primary rows, 6 with function keys, 7 if you have multimedia buttons.

this could easily do 9 (2 more rows which would already be substantial more keys to press) to 12, depending on what goal is being aimed for density vs less likely to hit accidentally.

[kurplop]

Clever idea with the coat hangers.

I like to see alternative designs like yours and will be watching your thread.

[CPTBadAss]

Yeah....so I'm going to need like 45 more updates from you intrio. Could you make that happen so I can properly bask in the awesome that is this thread?

[intrio]

It's been awhile, so busy haven't had time. But getting back into it on weekends.

 This time working with clay to build arm trays, when temp is in 70F's standard art store plastilina clay is on par with rubber with hardness, when 100F you can easily press into it, for body impressions even molds.   

This existing keyboard and trackball is the 'control' I'll be using to contrast the Conchi and related experimental work. My first pass will be trying to replicate the size of that keyboard layour but split to keep the learning curve as low as possible.

I spent the last week using this clay foundation for fulltime work, ( covered with paper towels as clay is greasy) in middle of a work release and very clear benefits markedly less stress on trapezoids, and wrist flex tenons, and tendons on back hand, the ability to lift up the keyboard to keep the wrist neutral is awesome, and the elbow bumper and forearm wells act as a gentle reminder, when one is out optimal alignment, it doesn't feel right.   That said there are things I'll change, but has given me enough confidence to try a thermoplastic casting of it as clay does eventually bend (which is one reason it looks ugly, had to repair a few times).  Not shown in similar work with chair.


The massiveness and form fitting of the clay allows the body to relax, some internal cue of the brain is hyper-aware of "falling" so lighter weight supports (e.g. tv tray, art drawing board) so at least my brain tends to tense and hold muscles, leading to knots and tendonitus.  I first noticed this on my sit stand workstation when I switched out the surface from steel wire to a massive butcher block countertop from IKEA.  Massive surfaces, dull the vibrations during typing almost completely.

The malleability is also helpful, particularly on left side i found dents handy for my elbow when supporting my chin ( prone to have this position when thinking) and right side for the double duty of keyboard and trackball, and similar elbow dent

[intrio]

heres how it looks in use, keeps the arm neutral when normal typing and supporting the head.  . Support points are the palm of the hand and the back of the forearm, with no contact at he wrist so wrist tendons don't have pressure on them.

[intrio]

trying to figure out if microswitches should go verticle or horizontal, they are slightly bigger than the keys I've been usijng in the long axis

[intrio]

here playing switch distance and curve, seeing how close they can be so that a finger pressing straight down doesn't hit the others immediately by, but interestingly two buttons could be pressed at once of aimed at between the keys, opening up the possibility for chording, maybe things like cut copy paste, function keys.

[intrio]

i've been hacking the microswitches, the addition of a cut silicone tube as a ligament between the steel lever and orange plastic microswitch, GREATLY improves feel and responsivity, by eliminating play in every direction,  making it super responsive and feel stable, normally the metal lever is pretty loose, so jiggles, making fingers, and can even get caught on edges and lift up.

[intrio]

 here exploring again the microswitch orientation, and using the same silicone tubing as the key cap, both which i really like.

the verticle orientation makes the key column very thin, and potentially fan like expandible, useful for puttin in a travel bag

Normal membrane switches keyboards have a rubbery dome or layer, whose give, help absorb impact ( unlike an hard iPad screen).   so using the silicone in any fashion helps make a more comfortable keypress.

the connecting surface to the finger also matters, both as a cushion and tactile feedback so muslce memory can auto hone in.   so here by various cutting of the small silicone tube, making a few points to match a subtle arc of finger tip pad, thinking the spiky version good for home row, the end bubble for normal keys.  it has the other benefits that it extends the length making the pressable area bigger, and space between keys smaller, and more grippy, its almost impossible now to not hit a key.

[intrio]

in related work, the hacked microswitches are being used in a new lighweight wireless foot mouse/ sock/ sandal to help ofload work from overworked hands and fingers, particularly thumbs

  here a anker five button wireless optical mouse was removed from its housing, optical mouse disabled and buttons rewired to use microswitches, click, double click, middle click, right click work  and the sandal tabi boot inspired design allows them to reasonably work in any orientation regardless of foot position e.g off floor cross legged ( my first foot mouse was floor operated), and can almost be practically walked in ( useful going to kitchen to get a snack)

i first tried a gyroscopic mouse, which was great for a bit, but it had the deal killer no middle mouse button for scrolling, and the mouse couldn't be turned off, so moving feet around would compete with arms for driving mouse resulting in confusion, and worse no way to separate toe movement from foor movement so trying to click would force mouse outside of button in infuriating way.

[intrio]

really happy with the wireless mouse hack for the foot mouse, i ordered 2 mini wireless keyboards/ mouses, and will hack the switches and wireup my own, this particular style of mini membrane switches are very easy to hack vs most keyboards with impossible to solder on traces, simply peel back the film and connecting pad, and careful solder wires to each connector on the pcb board.

this approach has many advantages over the teensy, its cheaper, simpler solder and go...no diodes, matrix code etc, its RF wireless out of the box and can work a long distance, drivers included for pc, mac and android, its battery powered effeciently ( some of these are power efficent enought to last a year on a single set of batteries), its MUCH smaller, basically 2 chips do everything. 2 can be used at same time for independent hands at any orientation, no connecting cable.

[intrio]

turns out the silicone tube slices when beaded onto the wire, works well as a way to hold in place the switches on a coat hanger out, this is useful for reconfigurable figuring out spacing and arc,  particularly from the perspective of muscle memory, where each button for each finger thinks/ feels fingers should be, prior to locking down in plastic.

i will be doing these as rows to match the keyboard i'm used to, with the exception that each half the keyboard might have some overlap with the opposite side, so they can be hit from either hand

[intrio]

first row of keys created, had to scavenge coat hangers for support... most i have these days are entirely plastic 😜 .

A fine tip black Sharpie is being used to label the keys which works well on the silicone tube, those drafting courses in juinior high helped with lettermanship.

 The spacing of the keys allows the fingers to be put inbetween for rest as the switches are so lightweight they won't hold the fingers up at rest.

[intrio]

changed up silicone 'key cap' to be more rounded matches curve of finger pad nicely without much work

[intrio]

yay first half of keys done, exciting as it will be soon time to solder these puppies on and try actual use.  I'm guessing 15 hours of work prior to a first functional version, and should with the holiday break be useable before start of new year, months ahead of other exciting keyboards and inputs that are on the rounds.

I've been using the time waiting at the airport for the holiday travel to get this far.  Good way to pass time, though it does evoke some curiosity from bystanders, wondering what is being done.

To construct these, I am using a tiny pair of scissors meant for trimming beards, which both bypass airline security and offer easy access for creating the T shaped cut necessary for the top part of the keycap curve, and a pair of needle nose tweezers to get the oring and 'keycap' on to the steel microswitch lever.   When placing the oring which is being used as a tensioner and shock absorber, i have to tune it a bit...too far and the key might not fully reset, i partly do this via holding it up to the ear to hear the action.

So far I'm really excited to try these out in real world typing. They are FAST 10-15 clicks possible per second) with the tensioner and keycap, a comfortable fit, almost no play, slop, slide, chatter, and much crisper than the heavier sluggish other keys I have tried.  There is a delightful auditory and tactile feedback, from the landing of the finger on the keypad, tension warning of the impending click.  A clear down click and clear up click, and release, which can be felt even with headphones on which along with the subtle keycap shape, means the fingers and muscle memory can do the driving as it should be, versus ipad and touch stream inability to be sure whats being pressed.

[intrio]

115 keys modded ready for action.  I think this is enough for first pass of alpha numeric and core symbols but probably not all function keys, had 2 bad switches in the group so far.

[intrio]

starting to take shape, got right side (minus numeric and function key rows) in a rough form, its looking more like a maltron meets robotech black and white metal, than i had first imagined.

i'm matching the various compound curves the fingers form at the perceptual distance eyes closed of a keyboard row, and key position, one at a time, top to bottom, and adjusting the key cap shape and size, never really occurred to me how much pinkies get quite a bit of variety.

The wire foam and silicone are a decent 'sketching' medium for this highly iterative build only know right when you feel it process, its amazingly light, which might make it even wearible once finalized.

It's going to be interesting to see how it works, my goal of doing the holiday letter on it is a new years goal. .

[intrio]

coat hanger is turning out to be too stiff for the rather sharp breaks as fingers get fully extended which is tuning out very important for ergonomics minimal finger travel from neutral position, so switching to solid core 24 gauge wire, which is a pretty good as a scaffolding, a bit like tensegrity when used with solid parts in between.  It's actually amusing to me as i got this wire meant for the keys, not wiring them up.

first 3 fingers get 2 switches each on the top row, and 6 on the pinky.

next step will be 'setting it' locking in the shape, by filling the gaps with thermoplastic, which is very flexible, when heated to the hot point it is clear and pours like honey it actually doubles as hot glue, sticking to plastic in ways that can be rather permanent, and lower temps a tacky, reversible bond.

[kurplop]

I've got nothing to add, but keep up the reports. I'm watching your progress with interest

[intrio]

rigidized by the thermoplastic, which i added a bit like winding twine around the gaps, then pushing it into the gaps with a small flathead screwdriver, then i set in the freezer for 10 minutes to solidify.

its going to be a annoying to take these apart if a switch goes bad, but its not that hard if you have a hot air gun with small nozzle and adjustable heat.

[intrio]

trying a finger centric tapestry/weaving with wire, as doing rows at a time leads to impracticle to solve key collisons. 

this is for the right hand index finger, it has 11 keys vs the normal 10, which might be either key will hit a letter, or something else, maybe redundant with other side of keyboard.

it looks like the robotech cannons, and also like an organist foot pedals

[intrio]

middle right finger, 16 keys possible vs typical 9

[intrio]

turns out this wireless keyboard isn't a full keyboard, missing important keys like pause, break and backspace, which are a bit of a bummer, as they are heavily used in programming, though I might be able to software remap the function or media keys i rarely use. Not sure yet.

even with this limitation, i'm still going to try it first over the teensy, as the smaller self contained wireless multidevice is worth it.  a bigger worry is the mouse built into the device doesn't seem to be disableable so far, this might be cool,or might suck.

as it has less keys and i'm curious about how small of a minimal finger travel, i redid the right hand again to support the keys vs the maximum keys reachable by a finger. i drew a keymap, and color coded the sections then used similar colored wire to help keep track, the sharpie doesn't stick well on the silicone out side, which can be used to temporarily mark the key. some keys like arrows are going to get moved.

I have figured out 2 weavings with wire, one dense (right), one looser (left).  the dense one is being used for the first pass wireless version, and may end up wearable like the footmouse, a brace/ exoskeleton to hold it in place.  i have enough parts for a keyboard that has significantly more keys than a full sized keyboard which would be interesting to explore macros, and might get embedded into the next version of the chair which will have articulated arms

one reason i'm excited about this keyboard is in the past do to my special needs i needed to stockpile the scissor action keyboards as they were the only ones i found that worked for me and they stopped being made.   They wear down in year or two and fast forward, my stockpile is down to my last one.   Now woth this approach as the use easy to find and easy to hack parts, i shouldn't ever have to worry about a closed source monolithic keyboard that can't be repaired.  Even better in the future I expect these sorts of devices to get even easier to make, especially if 3d printed with conductive traces embedded, so keys just need to be snapped in, or even the whole switch 3d printed, keycap and all.  That would open up the way towards mass personalization, could easily see sending a kinect like scanner to people that can map finger travel, selectable keymap, that is then custom printed for them.

[intrio]

an alternate keycap using paper under the semi transparent silicone, seems viable

[intrio]

went back to thermoplastic, the wire was good for initial placement and getting rough feel, but couldn't get the depth of curve needed, or stiffness needed to type well.

this time i'm laying it out on foam ( the thermoplastic doesn't stick to) making sure the fully depressed key doesn't hit the raised key below it, and that the curve matches that of the finger arc with palm relatively anchored in one place.  It's a bit like laying bricks in a curved wall, bricks rpthen mortar between, then layer over all, then repeat.

[intrio]

since i've been focusing on fingers, but now looking at created the base, trying to figure out how they connect so using coat hanger to try and get spacing and angles correct, but i will likely do it again tomorrow after a visit to hardware store for longer wire, coat hangers aren't long enough

[Camineet]

Really hard-core awesome project.  With 8 g keys, I could finally start typing again instead of having to use speech to text

[intrio]

realized cardboard and sharpie was a better pattern than wire for getting finger arc and finger spacing right, so here using opposite hand to hold pen or trace around finger.


from here i will cut out the parts, then i can use a rolling pin to flatten into about 1/4" thick the thermoplastic, then use shears to cut around the carboard like a pattern to match

[intrio]

Really hard-core awesome project.  With 8 g keys, I could finally start typing again instead of having to use speech to text

thanks man,  yeah i hear ya, speech to text is harsh on technical things...tried programming with dragon speak and woah was that frustrating. it's not clear yet how well this will work, but I do have high hopes and tenacity whatever happens.   

[intrio]

I've got nothing to add, but keep up the reports. I'm watching your progress with interest

thanks for the vote of confidence and following the adventure along.

[intrio]

packing foam is easier to cut than cardboard, taped to connect, still using wire for holding the curves and here's though it's still fiddly getting all the parts to work together with the hand constantly, moving a bit here and there

[intrio]

foam and wire worked well as scaffolding for the thermo plastic, and doesn't care about getting wet, which is useful to set parts in place.   I first created the palm rest, and beginnins of finger wells, then did one finger at a tome match the previous foam pattern, solidified it under cold water except for the end, which was still gooey then rested whole thing overhanging a table edge, and adjusted position till it felt right.  my goal is to have fingers move minimally (avoiding wrist bending and raising forearm, necessary for longer finger travel) as keys will be raised, each top surface of the support should barely be touchable.

there will be additional rows for the first, thumb and right finger, but gotta start somewhere.

 i'm gonna have to play with it a bit using a heat gun and flat blade sculpting tool (aka screwdriver) before adding the switches

next challenge is to figure best way to attach the switches this way, the thermoplastic can be tapped and I could use screws, or using the super hot thermoplastic a small dab on the switch and

[intrio]

right hand middle, ring and pinkie mostly done and in right position, its really interesting the fingers create almost helical patterns, like a 45 degree diagonal.


i'll probably redo the plastic again, having learned a few tricks getting keys on in right place, and a few alignment issues not easy to fix with keys on, the hot air gun creates a nice soft patch to glue these in, which a screwdriver can help create flanges to keep it in place