Author Topic: Re-Create the DataHand - Thumb cluster under development. Project 75% done.  (Read 182606 times)

0 Members and 1 Guest are viewing this topic.

Offline damorgue

  • Posts: 1176
  • Location: Sweden
    • Personal portfolio
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #200 on: Sat, 04 January 2014, 00:52:30 »
Have you used smoothed the surfaces with acetone or used any sort of lubrication?

Offline yasuo

  • Posts: 976
  • Location: ID
  • spanengan puyeng newbie
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #201 on: Sat, 04 January 2014, 01:37:32 »
Really niceBaby DH,interested with thumb cluster like DH or different
Logitech MK220 Colemak DH
SplitSyml by Moz BlacksMx fuk blacks

2/3 8.5pm                                          in de la my september month ya da all get my fukka "fake message"

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #202 on: Sat, 04 January 2014, 09:43:22 »
Have you used smoothed the surfaces with acetone or used any sort of lubrication?

No. Some of the parts directly from Shapeways
could be used without modification. On other prints
I trimmed or scraped with a hobby knife until there
was no interference, or sometimes just forced the
part to move until it self-machined into a smooth fit.
This only takes a few seconds if it is going to work.

Really niceBaby DH,interested with thumb cluster like DH or different

Not certain what you mean yasuo, but I intend to
make a thumb cluster very similar to the original one.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #203 on: Sun, 05 January 2014, 23:40:00 »
The PCA9655E I2C I/O expander is driving
me a bit crazy. It seems to insist upon an
I2C STOP before a read or it refuses to
actually follow the command for which
register to read out. (instead providing
one of its choosing!) Of course, the
datasheet makes no mention of this
behavior whatsoever.

While this isn't a huge problem, I have
wasted most of a day figuring this out,
and it may end up making the software
less efficient. What a pain.

Offline agodinhost

  • Posts: 767
  • Location: Brazil, RJ
  • Soylent green is people ...
    • Dr Ian O Xaman
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #204 on: Mon, 06 January 2014, 07:15:35 »
Have you used smoothed the surfaces with acetone or used any sort of lubrication?
I used acetone vapor for ABS.
It also works on SLS nylon?
Building one square I2C keyboard with those 1200 switches (thanks JDCarpe)
GH60 |GH60-Alps |GH60-BT |GHPad/GHPad Alps |GH60-Case |Alps TKL |EL Wire |OS Controller, Round 2 |My Custom Keyboard |WTT/WTB

Offline engicoder

  • Posts: 718
  • Location: North Carolina
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #205 on: Tue, 07 January 2014, 21:45:27 »
Just noticed there is an original Datahand up on ebay http://r.ebay.com/SDPf04....A bit spendy  :eek:
   

Offline csirac2

  • Posts: 3
  • Location: Canberra, Australia
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #206 on: Tue, 07 January 2014, 23:17:24 »
It'll probably go for even more than that. I've seen Amost-new-in-box models go for nearly $3,000...

Offline arisian

  • Posts: 13
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #207 on: Wed, 08 January 2014, 15:53:01 »
Yeah, they show up occasionally.  They generally sell for $1-3k, assuming they're in decent shape; this one is a Personal rather than a Pro, so it probably will sell for somewhere in the lower part of that range.  In any case, the prices are too high and availability too low for long-time users to be able to keep using them when our old ones fail, which is precisely why everyone is so excited by this project :)

In any case, this is bit off-topic; we should probably try to avoid derailing this thread every time one shows up on E-bay.

Offline csirac2

  • Posts: 3
  • Location: Canberra, Australia
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #208 on: Thu, 09 January 2014, 02:00:47 »
Back on-topic, OldDataHands kindly helped me get an order for the finger key switch parts from ShapeWays going, so they're on their way.

Just trying again to put a DigiKey order together, I seem to have constant interruptions this week :)

Very exciting times. Hope to be assembling something toward the end of the month and in doing so hopefully contribute in some way.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #209 on: Sun, 12 January 2014, 23:13:49 »
Spent a few hours tonight and now have
the left-hand matrix scanning as well as
the right-hand matrix. I don't have any
USB code running yet, but am scanning
at ~5.5ms intervals limited by the 100kHz
I2C bus frequency. So far so good!

Offline claussen

  • Posts: 6
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #210 on: Mon, 13 January 2014, 13:00:56 »
Wow, amazing progress.   Note to self:  completely forget about cool projects and good things happen :)

Curious what your current plans are for making larger quantities of the metal springs?
Seems like hand bending those is going to be a source of some quality control issues even just on a single unit -- getting the right profile dialed in and mass produced seems like it would be a key part of a group buy, since the tooling costs will be non-trivial but making several thousand would be otherwise quite cheap, yeah?

I spent November working abroad without my DH desk setup -- still recovering from the trauma.  Ugh.
Then we had a break-in at my office -- thank god the thief didn't know my DH rig was more valuable than the laptops he stole...

Offline regack

  • Posts: 655
  • Location: Thessia
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #211 on: Mon, 13 January 2014, 19:03:08 »

Ok, so I just got these today.


50809-0


I was able to bridge all of them - except for the one with the 10 mil gap - but some took more than one application of solder. The three that were the easiest were:


1) 5mil interlock
2) 4mil interlock3
3) 0.05mm spacing


Those three bridged as soon as I put the iron on them, and didn't wick away.  I'll try with a second board just to see what happens and if it changes.  I tested the connections with the fluke. 


Let me know if you want the .mod file or the whole project or something.
« Last Edit: Tue, 14 January 2014, 07:23:33 by regack »

Offline jalli

  • Posts: 101
  • Location: Toronto, Canada
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #212 on: Mon, 13 January 2014, 20:13:40 »
This is just awesome, rather than making a guess and design based on it you did a proper test, my hat's off to you!
Antonia

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #213 on: Mon, 13 January 2014, 22:16:38 »
Curious what your current plans are for making larger quantities of the metal springs?
Seems like hand bending those is going to be a source of some quality control issues even just on a single unit -- getting the right profile dialed in and mass produced seems like it would be a key part of a group buy, since the tooling costs will be non-trivial but making several thousand would be otherwise quite cheap, yeah?
I've been meaning to attend to this topic. There is a local
shop which seems to specialized in these types of things.
I'm going to have them quote it. My problem is that I'm an
electrical engineer and I don't know dimensioning & tolerancing
for mechanical drawings. You've spurred me into action though.
Here's a screenshot of my first attempt at capturing the design
with the important dimensions:

50822-0

Anyone care to help an EE out with some mechanical drawing
insight? Here's the matching .dxf file: I would welcome any
corrections. (drawn with LibreCAD)

* clip_mechanical.dxf (35.3 kB - downloaded 96 times.)
« Last Edit: Mon, 13 January 2014, 22:18:17 by OldDataHands »

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #214 on: Mon, 13 January 2014, 22:29:40 »

Ok, so I just got these today.


(Attachment Link)


I was able to bridge all of them - except for the one with the 40 mil gap - but some took more than one application of solder. [size=78%]The three that were the easiest were:[/size]


1) 5mil interlock
2) 4mil interlock3
3) 0.05mm spacing


Those three bridged as soon as I put the iron on them, and didn't wick away.  I'll try with a second board just to see what happens and if it changes.  I tested the connections with the fluke. 


Let me know if you want the .mod file or the whole project or something.

regack!!! Amazed and delighted is what I am! You originally said
that you wouldn't be the fastest. That seems to have been sandbagging.

That test board is a thing of pure beauty to mine eyes. I will enthusiastically
take the .mod, and/or the whole project. You Da Man!

I'll PM you.

Offline damorgue

  • Posts: 1176
  • Location: Sweden
    • Personal portfolio
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #215 on: Tue, 14 January 2014, 00:03:34 »
Curious what your current plans are for making larger quantities of the metal springs?
Seems like hand bending those is going to be a source of some quality control issues even just on a single unit -- getting the right profile dialed in and mass produced seems like it would be a key part of a group buy, since the tooling costs will be non-trivial but making several thousand would be otherwise quite cheap, yeah?
I've been meaning to attend to this topic. There is a local
shop which seems to specialized in these types of things.
I'm going to have them quote it. My problem is that I'm an
electrical engineer and I don't know dimensioning & tolerancing
for mechanical drawings. You've spurred me into action though.
Here's a screenshot of my first attempt at capturing the design
with the important dimensions:

(Attachment Link)

Anyone care to help an EE out with some mechanical drawing
insight? Here's the matching .dxf file: I would welcome any
corrections. (drawn with LibreCAD)

(Attachment Link)


A few pointers:

-Tighter tolerances = higher cost. Larger number of tolerances = more places they will have to measure and verify = higher cost.

-The thickness of the sheet is specified to +-0.02mm. This is very tight and close to impossible. The material will be thickened or flattened at the bend depending on how they draw or bend it. It is also very difficult to measure if that thickness is correct all the way so best case scenario is that they ignore it.

-The 0.04mm width of the spring is quite tight as well, does it need to be that exact?

-Geometric tolerances. There are signs which is used to specify flatness, circularity etc. I suggest you look through this link and use them rather than text.

Edit:
-Stacked tolerances:
http://i.imgur.com/qB0XNle.png
The dimension marked in red will be +-1.3mm if you add up the tolerances of the other dimensions. Be aware of where you put the dimensions and exactly which dimension is actually important. Just something to be aware of.
« Last Edit: Tue, 14 January 2014, 00:07:56 by damorgue »

Offline regack

  • Posts: 655
  • Location: Thessia
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #216 on: Tue, 14 January 2014, 09:31:32 »
Let's see if this works...


Offline Wildcard

  • * Esteemed Elder
  • Posts: 1042
  • Location: Fields of Columbia
  • When caffeine isn't enough
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #217 on: Tue, 14 January 2014, 09:41:54 »
Wow, great progress on this project. I wish the data-hand had become more commercially successful, I'm glad to see this revival making it so far.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #218 on: Tue, 14 January 2014, 23:15:55 »
A few pointers:

-Tighter tolerances = higher cost. Larger number of tolerances = more places they will have to measure and verify = higher cost.

-The thickness of the sheet is specified to +-0.02mm. This is very tight and close to impossible. The material will be thickened or flattened at the bend depending on how they draw or bend it. It is also very difficult to measure if that thickness is correct all the way so best case scenario is that they ignore it.

-The 0.04mm width of the spring is quite tight as well, does it need to be that exact?

-Geometric tolerances. There are signs which is used to specify flatness, circularity etc. I suggest you look through this link and use them rather than text.

Edit:
-Stacked tolerances:
http://i.imgur.com/qB0XNle.png
The dimension marked in red will be +-1.3mm if you add up the tolerances of the other dimensions. Be aware of where you put the dimensions and exactly which dimension is actually important. Just something to be aware of.

Thanks for the pointers and info damorgue!

The shim stock I used as a starting material claims a
thickness tolerance of 0.00075" which is almost exactly 0.02mm;
however you're correct that my thickness spec in the drawing is
ambiguous - I don't actually care about what happens to the
thickness in the bends, or really anywhere but along the flat bottom
where the magnet will contact. I'll have to see if I can find a way to
communicate that in the drawing.

Regarding the width: the holding force will be closely tied to volume
of material in the leg of the clip that the magnet contacts. +/- 2.5%
of the width is ~ 0.04mm. +/-5% of the thickness is 0.02mm.
Stacked together this is a 15% potential variation in holding force,
which I already think is a lot - I don't think I should open those specs.
Thoughts?

I agree that the dimension you noted ends up being wide open at
+/-1.3mm. That is intentional - There is some latitude in where the
top corner ends up relative to the right-side. It does not need to be
kept tight, but can't be left unspecified since it must fall within a window.
It serves as the retension mechanism so that the clip doesn't fall out...
just needs to dig into the plastic a tiny bit.

I will look up the symbols, take another swing at it, have an ME buddy
review/update, then post and send for quote. Any bets on the per-clip
cost at 1000 pieces?

Thanks again for the input!

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #219 on: Tue, 14 January 2014, 23:19:42 »
Let's see if this works...
Thanks regack!
I spent 2+ hours tonight trying to update my footprint in KiCAD
to match the "5-mil interlock" footprint you designed...
but something went very wrong after I was done... I saved the
module and the tried to pull it into my layout, but the footprint
didn't load anything like I had saved it - totally trashed. Oh well.
I'll try again in a day or two.

Offline jacobolus

  • Posts: 3634
  • Location: San Francisco, CA
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #220 on: Wed, 15 January 2014, 05:26:11 »
Itís maybe the wrong place to ask, but does it seem to you guys like the sideways motion of the fingers to activate the left/right switches is substantially more difficult than moving the finger down to activate the bottom switch (flexing the whole finger) or the up/down switches (flexing/extending the 1st finger joint), especially for the weakest couple of fingers?

I wonder if it would ultimately be better if each finger assembly had 4 switches (bottom, up, diagonal down-left, and diagonal down-right) instead of 5 [or possibly even just 3 switches per finger], and then some work was done on the letter assignments to each switch, possibly with some more chording, etc.?

[Iím speaking purely hypothetically here; I never had the cash to obtain a Data Hand, and wouldnít know where to try one.]

If youíre going to completely re-make something from scratch (esp. if the goal is to attract new users), it might be worth trying to optimize its function. Then again, maybe itís better to just get as close as possible to the original, to serve all the original data hand users who are up a creek with the originals out of production so long.

Offline gbjk

  • Posts: 7
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #221 on: Wed, 15 January 2014, 06:03:58 »
In short: No.

Once you're used to a datahand, there's nothing wrong at all with the layout.

I've developed a strange mannerism of pressing the "Normal" key in between typing,
and just generally, because nearly all issues are being in the wrong mode.

That's probably just a "typing tick" though.

I've used a datahand for 4 years now, about 60 hours a week as a heavy typist of everything from code to long emails.

Offline wejn

  • Posts: 12
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #222 on: Wed, 15 January 2014, 17:07:23 »
I've developed a strange mannerism of pressing the "Normal" key in between typing,
and just generally, because nearly all issues are being in the wrong mode.

That's probably just a "typing tick" though.
Hear, hear! I found myself nodding at this because that's exactly what I'm doing. After several years of using DH exclusively I also have this habit of pressing "N" because wrong mode accounts for over 90% of all the [few] errors I make when typing.

And I have to concur that changing the layout doesn't make sense (IMO). The current layout just works. My only gripe is price of a second unit (I'm carrying it to the office and back home on regular basis) and the relative bulkiness.

Offline Pacifist

  • Report me *again* if there are gifs in my sig
  • * Elevated Elder
  • Posts: 3599
  • Location: Cali
  • on hiatus
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #223 on: Wed, 15 January 2014, 17:16:25 »
Any estimate of a cost when this project is fully finished?

Hopefully as far away from the 1k on ebay

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #224 on: Wed, 15 January 2014, 20:02:24 »
Any estimate of a cost when this project is fully finished?

Hopefully as far away from the 1k on ebay

My current guess is that the parts will cost you ~$500-$600.
Perhaps a group buy could knock that down by $100 if you
get sufficient volume on the PCBs and electronics.

Here're the details:

In order to purchase the components of the finger
switches to make a single keyboard is ~$250.
(not counting the metal clips - no quote yet
and not counting shipping)

I can imagine the thumb clusters hitting $150 or
more depending upon the volume (cm^3, not quantity)
of 3D-printed parts.

That leaves the case as a big question mark. Could
easily land between $100 and $200, or a DIY thing
out of plywood!

I guess that a group buy should be able to knock
$100 off the total for electronics and PCBs.



Offline hillel369

  • Posts: 2
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #225 on: Thu, 16 January 2014, 15:28:23 »
Just wanted to say you've got another Datahand user here who'd be happy to go in on a group buy.

Thanks for all your work so far. I've been following the post for a while, it looks like you're making great progress.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #226 on: Fri, 17 January 2014, 00:05:29 »
Thanks to damorgue for the input. With
that, some serious redrawing, then redlining
by an M.E. friend, then more redrawing
I now have a drawing to get quoted.
I'll pass along the info once they get
back to me.

Here's a pic:
51270-0

and here's the .dxf for those interested.* clip_1_mech.dxf (31.54 kB - downloaded 100 times.)

Offline gbjk

  • Posts: 7
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #227 on: Fri, 17 January 2014, 04:16:46 »
OldDatahands,

Have you thought about sharing all the resources you've created thusfar on github or similar?

It would allow others to contribute more easily, and also be a launching platform to gain popularity and, I guess, orders.

Like you said, you're not thinking about the enclosures yet, but someone else could start to, and contribute the files to you for you to approve, etc.

( Sorry. Thinking about this like a distributed software developer :D )

G

Offline nclu

  • Posts: 3
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #228 on: Fri, 17 January 2014, 12:41:57 »
Been lurking this one since June too.  Only discovered the datahand just after it was no longer available, then stumbled across your project after a year and half of pining for one.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #229 on: Fri, 17 January 2014, 13:34:14 »
OldDatahands,

Have you thought about sharing all the resources you've created thusfar on github or similar?

It would allow others to contribute more easily, and also be a launching platform to gain popularity and, I guess, orders.

Like you said, you're not thinking about the enclosures yet, but someone else could start to, and contribute the files to you for you to approve, etc.

Already onboard with this approach, but am waiting to get
some feedback from a couple people before widely
advertising it. Expect a post here in a month or two.

Offline soliton

  • Posts: 3
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #230 on: Sun, 26 January 2014, 14:49:18 »
as stated before, i will buy one as soon as its availavle if you start a crowdfunding.
pls let us know!

Offline myetex

  • Posts: 1
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #231 on: Wed, 29 January 2014, 10:34:01 »
Just registered to say, please crowdfund this!  I'd absolutely buy a pair.

My only requirement is that the left and right devices must be able to be positioned separately.  Ideally, each device would be a separate Bluetooth keyboard.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #232 on: Wed, 29 January 2014, 15:38:37 »
My only requirement is that the left and right devices must be able to be positioned separately.  Ideally, each device would be a separate Bluetooth keyboard.

How often are you willing to replace batteries in a keyboard?
How big a battery are you willing to tolerate?
I ask because this switch design is power hungry.
Each IR LED will take > 5ma when active, each
IR photo transistor takes the same if key is pressed
When scanned. With 5 cols on each half you've
Got an average draw approaching 50ma when idle.
AA batteries might last a couple days. Because of this
I'm not convinced that this is a great design for wireless.
Mine will be wired.

There are people who've done wireless mods, but they're
relying on large, external battery packs. Not something
I imagine a lot of people being happy with.
« Last Edit: Wed, 29 January 2014, 15:42:26 by OldDataHands »

Offline claussen

  • Posts: 6
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #233 on: Thu, 30 January 2014, 12:28:51 »
The optical switch design is definitely a painful one for a battery approach.

For a battery-powered approach, I would consider a capacitive system -- it would take some tuning, but one could position sensors such that the movements of the leg springs or magnets would create a large enough signal to be sensed reliably and with a much smaller bias/sensing waveform than an optointerruptor.  A Cypress button-scanning touch IC should be able to run like that for a very, very long time.

One bonus would be ambient light tolerance -- direct sunlight borks my DH system.
 

Offline tricheboars

  • * Esteemed Elder
  • Posts: 939
  • Location: Denver
  • Keyboards are Important!
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #234 on: Thu, 30 January 2014, 19:44:23 »
holy ****. wow. this is incredible work OldDataHands. you should be really proud.
|  Fundamentalist ErgoDox Zealot  |

Offline dorkvader

  • Posts: 6282
  • Location: Boston area
  • all about the "hack" in "geekhack"
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #235 on: Thu, 30 January 2014, 21:40:45 »
The optical switch design is definitely a painful one for a battery approach.

For a battery-powered approach, I would consider a capacitive system -- it would take some tuning, but one could position sensors such that the movements of the leg springs or magnets would create a large enough signal to be sensed reliably and with a much smaller bias/sensing waveform than an optointerruptor.  A Cypress button-scanning touch IC should be able to run like that for a very, very long time.

One bonus would be ambient light tolerance -- direct sunlight borks my DH system.
 

Yes, I think hall effect / magnetic is probably a good idea, over optical (especially if it's affected by mere sunlight) and capacitive (which are a lot harder to actually read, due to all the interplay between capacitors and analog effects, etc. There is a reason old capacitive keybaords used several hundred volts to strobe the matrix!)

Offline claussen

  • Posts: 6
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #236 on: Fri, 31 January 2014, 16:11:13 »

Yes, I think hall effect / magnetic is probably a good idea, over optical (especially if it's affected by mere sunlight) and capacitive (which are a lot harder to actually read, due to all the interplay between capacitors and analog effects, etc. There is a reason old capacitive keybaords used several hundred volts to strobe the matrix!)

Most of that difficulty with capacitive buttons was engineered out over the past ten years by Cypress, Atmel, Synaptics and others.  They've got pretty impressively robust calibration and variable baseline management -- far, far more sophisticated than is required for simple button thresholding.
Especially when you sense a metal piece as the trigger instead of a squishy, variable-moisture-content finger.

But I'm curious if you know of a many-channel hall-effect system that operates at low power?  I haven't seen one before, and that would a great thing to know about!

Offline soliton

  • Posts: 3
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #237 on: Mon, 03 February 2014, 19:05:06 »
My only requirement is that the left and right devices must be able to be positioned separately.  Ideally, each device would be a separate Bluetooth keyboard.

How often are you willing to replace batteries in a keyboard?
How big a battery are you willing to tolerate?
I ask because this switch design is power hungry.
Each IR LED will take > 5ma when active, each
IR photo transistor takes the same if key is pressed
When scanned. With 5 cols on each half you've
Got an average draw approaching 50ma when idle.
AA batteries might last a couple days. Because of this
I'm not convinced that this is a great design for wireless.
Mine will be wired.

There are people who've done wireless mods, but they're
relying on large, external battery packs. Not something
I imagine a lot of people being happy with.

if battery life will be low go for a integrated battery charger. ppl will tollerate the fact that they have to hock the keyboard up to a powersupply every 2-3 days much more easy than switching the battery every week.
it would be even better than no wireless option.
logitechs first aproach to wireless optical mouses was like that and it was a big success.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #238 on: Tue, 04 February 2014, 15:03:27 »
if battery life will be low go for a integrated battery charger. ppl will tollerate the fact that they have to hock the keyboard up to a powersupply every 2-3 days much more easy than switching the battery every week.
it would be even better than no wireless option.
logitechs first aproach to wireless optical mouses was like that and it was a big success.

That's a nice idea. An integrated 2 cell LiPo battery or similar... still, first I need to get the thumbs working, and have a fully functional design before worrying about extras.

Offline arisian

  • Posts: 13
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #239 on: Wed, 05 February 2014, 16:59:35 »
if battery life will be low go for a integrated battery charger. ppl will tollerate the fact that they have to hock the keyboard up to a powersupply every 2-3 days much more easy than switching the battery every week.
it would be even better than no wireless option.
logitechs first aproach to wireless optical mouses was like that and it was a big success.

So long as it can be used when plugged in; this should be "you can unplug the USB cable for a while and it'll still work," not "it's a wireless device and when your batteries stop holding charge you can't use it for more than an hour at a time," the way a lot of those old mice were.  I'm all for wireless as an *option,* but frankly these things are bulky and heavy enough that worrying about the cord is kind of a non-issue for me.  Obviously, the units need to be separable; they always have been (and having the option for a longer/more flexible cord connecting them would be great, but I suspect we'll get that anyway), but I agree with OldDataHands that wireless is something to worry about *after* we get a working version of the wired model.

Offline Turbinia

  • Posts: 64
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #240 on: Thu, 06 February 2014, 12:16:50 »
I think someone mentioned this http://handheldsci.com/kb at some point. Would require there to be the TRRS wire between the units still, but would mean no additional work would be required to make it wireless. Just pop that under the hood.
| Dolch | KBT ONE | QFR w/PBT | Poker II |

Offline MiaKnecht

  • Posts: 1
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #241 on: Mon, 17 February 2014, 20:27:42 »
Hi,
First let me say.. you're awesome for all the work you've done... simply amazing.  I have DataHand lust.. I must find one, or even better have a new improved one.   Is there an update on this project?   I'd definitely be in if this goes to a group buy..  OR... if there is a way I can somehow get my hands on just the plastics I'd rock out my own processor and PCBs...

Thanks for all the work you're doing to keep this alive..

Mia K.

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #242 on: Wed, 19 February 2014, 21:01:01 »
Let's see if this works...

Hi regack,

So I've spent many hours now in trying to incorporate the best of your solder-bridge
layouts into my existing footprint. I keep running into the basic problem with kicad
that it doesn't consider adjacent or overlapping, but separate pads to be connected.
It then turns my otherwise clean design into a mess of a ratsnest of unconnected pads
I've filed a bug report with kicad: https://bugs.launchpad.net/kicad/+bug/1280645
in the hopes that they'll address this.

In the meantime I have moved on to working on the thumb switch mechanical design. 
I don't have anything to show there yet, but work has begun.

If nothing changes in kicad, I'll eventually adjust to one of your less-optimised, but
still working layouts... but thumb switches first.

Thanks again for your work there.

Offline regack

  • Posts: 655
  • Location: Thessia
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #243 on: Wed, 19 February 2014, 21:28:26 »
Let's see if this works...

Hi regack,

So I've spent many hours now in trying to incorporate the best of your solder-bridge
layouts into my existing footprint. I keep running into the basic problem with kicad
that it doesn't consider adjacent or overlapping, but separate pads to be connected.
It then turns my otherwise clean design into a mess of a ratsnest of unconnected pads
I've filed a bug report with kicad: https://bugs.launchpad.net/kicad/+bug/1280645
in the hopes that they'll address this.

In the meantime I have moved on to working on the thumb switch mechanical design. 
I don't have anything to show there yet, but work has begun.

If nothing changes in kicad, I'll eventually adjust to one of your less-optimised, but
still working layouts... but thumb switches first.

Thanks again for your work there.

Interesting... I'm not sure I noticed that happening, but then my board layout was very very simple, maybe it didn't manifest in that case. 

In case it helps, I'm running build 2013-02-13 BZR 3947 (pcbnew)

Offline dorkvader

  • Posts: 6282
  • Location: Boston area
  • all about the "hack" in "geekhack"
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #244 on: Wed, 19 February 2014, 23:05:07 »
would it be possible to just have a trace running between the pads and then have kicad to the routing and whatnot, then before you send it off to the fabs, "cut" each bridge?

Offline Turbinia

  • Posts: 64
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #245 on: Fri, 21 February 2014, 09:07:22 »
Sounds like you may have built your component/layout in a way that causes this. Components like resistors have unconnected pads and do just fine, I would check to make sure the layout isn't saying the pads should be connected. That should fix the rats nest.
| Dolch | KBT ONE | QFR w/PBT | Poker II |

Offline OldDataHands

  • Thread Starter
  • Posts: 277
  • Location: Michigan
Re: Re-Create the DataHand - work on finger switches nearly complete
« Reply #246 on: Sun, 23 February 2014, 17:09:26 »
Some good news: Work on the thumb switches has begun.
Here is a screenshot of the current state:

55368-0

Reconizable, but still quite a bit of work to do.

In more good news: The KiCAD guys have responded to my
bug report with a possible workaround to the
pads-not-being-connected-when-overlapped bug. Perhaps
I can now succeed in incorporating regack's suave solder-bridge footprint?

Offline Stone

  • Posts: 22
Re: Re-Create the DataHand - Thumb switches under development. Project 70% done.
« Reply #247 on: Wed, 26 February 2014, 13:13:20 »
As a 5-year Datahand veteran can I just say that this is totally awesome. More people deserve to know true typing happiness :)

Barring a slightly sticky 'a/1' key mine's been near-flawless ever since I got it out of the box. I should probably clean it more often that I do, though (it's so fiddly...and I'm terrified of breaking a key - Lynn promised to put some spares into my last order but now they're fully closed that door's slammed shut). Other than the housing being bulky and the feet not working very well, my only real issue has been that I didn't buy the DataChair arms to fix it to my chair before they ran out of stock.

Keep the updates coming, it's looking great! :thumb:

Stone

Offline Stone

  • Posts: 22
Re: Re-Create the DataHand - Thumb switches under development. Project 70% done.
« Reply #248 on: Wed, 26 February 2014, 14:13:52 »
OK, so I've read the whole thread now. Few thoughts:

I would really love to be able to mouse without leaving the DH. Right now I'm using it with an Evoluent VerticalMouse placed very close to the right side and reaching for it - not great fun, but usable. For me the best two options to put a trackpoint-style 'nubbin' would probably be here:


or here:


The first would be more instantly-accessible, but is a slightly less comfortable position to maintain and may restrict how close you can get the finger units to the palm for users with smaller hands - although mine are already on the small side. The second means rotating your hand by one finger's worth (i.e. the middle finger ends up resting in the index finger's keywell) but would IMO be more comfortable for prolonged mousing; it feels like I could use it for longer, anyhow.

In either case you could use the mouse-mode thumb-switch to remap space and Alt as the left and right buttons respectively - this would let you statically assign the right index-finger keywell as the up/down/left/right arrow keys in mouse mode since nobody uses the DH mouse function ;) I'd quite like to have the left index finger keys as home/end/PgUp/PgDn rather than the default map which I find slightly weird.

I've never been a massive fan of the 'down' keyswitches as they've always been a bit mushy - perhaps you could experiment with a Cherry brown in the middle and the DH-style NSEW keys?

It would be amazing to have an alternative - or a mashup of old and new pieces - as the obsolence problem has been worrying (clearly for all of we DH users!) for a while now. I really struggle with all other keyboards now; I've ruptured all of my sagittal bands, had tendon grafts to repair eight of them and tore three of those repairs in an RTA last month so it'd be a major blow to have it die on me - it'd probably mean giving up significant typing altogether. Even getting to the point where there's the chance of replacing broken parts is truly awesome. Never felt the need to roll my own controller but mine does crash occasionally (plus it's PS/2 and it's getting harder and harder to find PCs that support it!) so it's really heartening to see so much good work going on. I'm an electronics guy and have occasionally access to some very high-res printers, do let me know if there's anything I can do to help!  :thumb:

Stone

Offline Hashable

  • Posts: 4
  • Location: Ottawa, Canada
  • KAFP PHE TP U K
    • Ted Morin
Re: Re-Create the DataHand - Thumb switches under development. Project 70% done.
« Reply #249 on: Wed, 26 February 2014, 19:26:00 »
Oh wow, I can't believe that I found such a cool project just a couple weeks after getting my ErgoDox :P

I was wondering if this design still uses optical switches (which have the disadvantage of potentially not working in direct sunlight)

If it is so, why? Is there a viable alternative?