Author Topic: fixing the ergodox thumb section  (Read 151420 times)

0 Members and 1 Guest are viewing this topic.

Offline PieterGen

  • Posts: 135
Re: fixing the ergodox thumb section
« Reply #100 on: Sat, 25 October 2014, 16:28:45 »
I really like your Mini version. So much, that I'm volunteering to be your guinea pig :-)  In other words, I like to build it and share my thoughts. What are the keycaps in your pics? Are those the profile you recommend?

How reachable are the upper thumb keys (ESC and = in Reply #97) ? What use do you envision for the thumb keys: modifiers that need to be held down (like ctrl, alt, or layer shifts) or rather "normal keys" like the ESC and = keys  [yes I know this is in the numbers layer] ? 

Related: what would be the best switches? 'hold down' keys may be better served with linear switches, right? reds? blacks?  How about reds or blacks for all thumb keys and blues for the rest ?

Slightly off topic: My ideal would be a completely wireless setup. The easiest & safest may be: each half having it's own controller plus a (low energy) connection, be it wifi or Bluetooth (LE). This combined with a receiver that contains the "brains". Preferably small, the size of a USB-drive). This means 3 things to transport (2 halves + USB receiver) . But: no cords. Plus it's safer, because eavesdroppers can only intercept the individual keys, that do not have much meaning in a layered keyboard. They cannot hear the actual key codes that are sent to the computer, because those are calculated in the USB-receiver.


Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #101 on: Sat, 25 October 2014, 22:56:03 »
I really like your Mini version. So much, that I'm volunteering to be your guinea pig :-)  In other words, I like to build it and share my thoughts. What are the keycaps in your pics? Are those the profile you recommend?
The (Alps mount) keycaps in my pics come from a Canon typewriter. They are maybe halfway between Signature Plastics’ SS and DSS profiles.

I like the profile reasonably well, but it’s not quite ideal. I wouldn’t recommend it because it’s very difficult to find a copy of.

I would highly recommend using something sculptured. DCS is fine, or SS, or Cherry profile, or “OEM” profile, or whatever. I wouldn’t recommend using SA or DSA.

Quote
How reachable are the upper thumb keys (ESC and = in Reply #97)?
Very easy to reach, provided that the keycaps are sufficiently tall. They need to be tall enough compared to the other thumb keys so that when you press one all the way down your thumb doesn’t touch the key below. One thing that would work would be to use DSA profile for the lower thumb keys and SA (not sure which row is best) for those two keys. Alternately DCS row 5 works pretty well for these.

Quote
What use do you envision for the thumb keys: modifiers that need to be held down (like ctrl, alt, or layer shifts) or rather "normal keys" like the ESC and = keys  [yes I know this is in the numbers layer] ? 
I can imagine this keyboard being used in several different ways, depending on personal preference. I’ll try to give some specific layout ideas in the near future.

Quote
Related: what would be the best switches? 'hold down' keys may be better served with linear switches, right? reds? blacks?  How about reds or blacks for all thumb keys and blues for the rest ?
Personally I’m not a big fan of Cherry MX, but I’d use whatever switches you like. My little prototype is Matias clicky switches, but I just used those because I had a handy bag of them and they were pretty cheap so I don’t mind if I break some.

Quote
Slightly off topic: My ideal would be a completely wireless setup. The easiest & safest may be: each half having it's own controller plus a (low energy) connection, be it wifi or Bluetooth (LE). This combined with a receiver that contains the "brains". Preferably small, the size of a USB-drive). This means 3 things to transport (2 halves + USB receiver).
I wouldn’t recommend this for a prototype, but sounds like a great idea for someone to play with. Maybe HaaTa’s firmware could be tweaked to support such a setup.

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #102 on: Sun, 26 October 2014, 04:01:09 »
Jacobolus, when you want to compensate for contoured keywell with keycaps maybe you can get access to a cheapest 3dPrinter ever - some reprap. It is actually possible to print usable keycaps with a reprap: http://geekhack.org/index.php?topic=43362.0
It will not look very pretty but you can achieve funny shapes. Though it may not be practical since you may want to use only off-the-shelf keycaps.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #103 on: Sun, 26 October 2014, 15:17:31 »
For just experimenting, piling some kind of polymer clay or silicone on top of low-profile keycaps is sufficient to make various shapes, in just as effective a way as a cheap 3d printer will manage.

But yeah, for a flat PCB/plate keyboard, the goal is to figure out which easily available keycaps work best.

Offline Findecanor

  • Posts: 5035
  • Location: Koriko
Re: fixing the ergodox thumb section
« Reply #104 on: Sun, 26 October 2014, 17:28:51 »
I think you are changing too much. Do the thumb clusters, fix the bugs and add some options. Changing the numeric row is unnecessary.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #105 on: Sun, 26 October 2014, 17:38:08 »
The way I personally plan to proceed is to advocate that Massdrop:
(1) Continue to offer the Ergodox kits in their standard layout, but with the PCB and plate/case redesigned to save a bit of bezel space and move an ARM chip directly onto the PCB instead of using a Teensy 2.0, and of course to be compatible with Alps/Matias switches. Possibly use a different connector like a 4P4C telephone plug instead of TRRS. Possibly offer different case construction methods, for instance something like the Infinity keyboard’s bent metal construction, instead of layered acrylic. Even if it’s layered acrylic, slightly improve the thicknesses of the various layers.
(2) Additionally offer alternate-layout kits, in maybe a couple of designs. I want to figure out the best layout I can that is spiritually similar to the Ergodox, but I don’t feel any strong responsibility to preserve particular layout features in a new design.

I think you are changing too much. Do the thumb clusters, fix the bugs and add some options.
Well, for example, I think the current column stagger on the Ergodox is a significant design flaw. Increasing the amount of column stagger allows the hand to be positioned at a much straighter angle to the columns, and brings the thumb keys much closer to the natural resting position of the thumb.

The current Ergodox is basically an as-close-as-possible copy of the Kinesis Advantage layout, but a flat design needs some layout tweaks to continue to be effective, compared to a sculpted design, since the shape of human hands doesn’t change.

* * *

With that said, if you have a particular layout that you want to see happen, go ahead and make some mockups (or better yet, build a prototype!). Or are you entirely satisfied with your modded Ergodox design?
« Last Edit: Sun, 26 October 2014, 17:46:27 by jacobolus »

Offline sordna

  • Posts: 2248
Re: fixing the ergodox thumb section
« Reply #106 on: Sun, 26 October 2014, 21:04:40 »
Well, for example, I think the current column stagger on the Ergodox is a significant design flaw.

Actually the stagger is not too bad. If anything I would lower the pinky columns a bit. I disagree with designs showing a very pronounced stagger because while the middle finger is longer when extended, when you curl your fingers to hit the bottom rows, the fingers all come to very similar positions. You need to take into account the finger position for all rows of the keyboard and find a column stagger that is the best compromise.
Kinesis Contoured Advantage & Advantage2 LF with Cherry MX Red switches / Extra keys mod / O-ring dampening mod / Dvorak layout. ErgoDox with buzzer and LED mod.
Also: Kinesis Advantage Classic, Kinesis Advantage2, Data911 TG3, Fingerworks Touchstream LP, IBM SSK (Buckling spring), Goldtouch GTU-0077 keyboard

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #107 on: Sun, 26 October 2014, 21:34:18 »
Actually the [Ergodox] stagger is not too bad. If anything I would lower the pinky columns a bit. I disagree with designs showing a very pronounced stagger because while the middle finger is longer when extended, when you curl your fingers to hit the bottom rows, the fingers all come to very similar positions. You need to take into account the finger position for all rows of the keyboard and find a column stagger that is the best compromise.
Have you tried it?

Having just tested this (a couple weeks ago) by directly comparing a couple of prototypes, I can safely say the for my hands at least, 1/4u stagger between the middle finger column and the index/ring finger columns makes a substantial improvement when compared to 1/8u stagger. For myself, I would make the stagger even a bit more than that (perhaps 3/8u or even 1/2u of stagger), but I think 1/4u is a fair compromise.

I don’t plan to use any keys which are two rows below the home row, or if I did, I would build a sculpted board which orients those switches at a steep angle. For just one row below the home row, I don’t imagine anyone is going to have a problem with 1/4u stagger of the middle finger. I’ve tested the 1/4u stagger on friends with tiny hands and with enormous hands, and it seems to be reasonably effective for everyone.

The stagger on the Ergodox is much too limited, in my personal experience/opinion, and the experience of several other people I’ve talked to about it. Prototype keyboards with more stagger are substantially more comfortable.

As for the pinky column: when you say lower, you mean move it closer to the body? Yes, I agree. For myself, I would put the pinky column 3/4u closer to the body than the ring finger column. However, in testing, several people found that to be too much stagger for that column. As a compromise, 1/2u of stagger for the pinky column seems to work pretty well for everyone I’ve tested it on, even though it now requires a bit of reaching to get the pinky to reach the top/number row.

Obviously, testing on even more people would be helpful.

One thing to note: the angle of the hand relative to the keyboard makes a large difference. Most people are used to angling their hands quite a bit relative to the keyboard, as well as tilting their wrists slightly upward on the thumb side, because they are used to a standard QWERY keyboard. If you switch to a completely straight grid/matrix board, then it’s still natural to angle the hands quite substantially relative to the direction of the grid (maybe 20–30°) and tilt the wrists. However, on a column-staggered board with sufficient stagger between columns, it suddenly becomes possible/comfortable to orient the hands quite straight relative to the columns, and re-orient the whole half keyboard to best align with the wrist, turning, tilting, and moving it until it is maximally comfortable.

It’s obviously a bit better to add some vertical height difference between columns, which reduces the need for a horizontal stagger between columns (the Maltron and Kinesis both do this, the Maltron more effectively than the Kinesis in my opinion), but that gets much more difficult and expensive to manufacture.
« Last Edit: Sun, 26 October 2014, 21:45:15 by jacobolus »

Offline kurplop

  • THE HERO WE DON'T DESERVE
  • Posts: 992
Re: fixing the ergodox thumb section
« Reply #108 on: Sun, 26 October 2014, 23:27:48 »
I found that too much of a stagger complicates the profiles of the keycaps. That is, an over aggressive stagger was interfering with lateral finger movements on one of my prototypes.  Like Sordna, I found that with curled fingers a minimal stagger was adequate.

The problem is that not everyone has the same size hands or the same hand positioning on the keyboard. I think you're on the right track getting as many people as possible to test your layout and then find a compromise that will accommodate most people. I do think that you have a very good design already.


Offline tufty

  • Posts: 347
  • Location: French Alps
Re: fixing the ergodox thumb section
« Reply #109 on: Tue, 28 October 2014, 06:48:44 »
but that gets much more difficult and expensive to manufacture.
I've been thinking in terms of origami recently, particularly curved folds.

Offline PieterGen

  • Posts: 135
Re: fixing the ergodox thumb section
« Reply #110 on: Fri, 31 October 2014, 11:55:56 »
@jacobolus and others:

- any hints on where to find some DCS row 1,2,4,5 profile thick PTB blank keys? (Yes, I know the All Keycaps Site List)

- or would an 'OEM profile' do as well?

- lastly: if you would mount the switches stepped and at an angle, you could use uniform keys, right? Like, DSA. The construction of the board itself would be more difficult, but it would be much easier to find printed key caps, because now uniform key caps can be put on any key. I have no idea if this is doable?

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #111 on: Fri, 31 October 2014, 14:33:11 »
- any hints on where to find some DCS row 1,2,4,5 profile thick PTB blank keys? (Yes, I know the All Keycaps Site List)
DCS keycaps can be ordered from signature plastics. This is expensive in small batches, but becomes still-expensive-but-not-world-ending at larger order sizes.

Quote
- or would an 'OEM profile' do as well?
It might, try it out! I don’t have any OEM profile keycaps with an extra-tall function row, but I do believe they exist. (Though at that point those caps are getting really tall and you’ll probably get some bonus wobble.)

Just the standard rows of OEM, Cherry, DSA, SS, or similar profile work okay for the main keyboard section. I just think skipping a row and using extra-tall-f-row caps for the number row works even better.

It’s pretty important on the Ergodox (or on my designs) for the two further-away thumb keys to be substantially taller than the closer ones, so that when you press the further ones the closer ones don’t actuate. The number row and home row of Cherry profile or DSA or whatever might have enough height step for this, but the extra-tall f-row caps are recommended. (Or if using spherical caps, I think SA row 2 for the far keys and DSA for the close thumb keys would work pretty well.)

Quote
- lastly: if you would mount the switches stepped and at an angle, you could use uniform keys, right? Like, DSA.
I would mount them stepped but not at an angle (or a bit of angle might help for the bottom row). The angled tops are not crucial if you can add your own arbitrary height step. But yeah, that’s much more difficult and expensive to produce, especially at scale.

Offline steve.v

  • Posts: 171
fixing the ergodox thumb section
« Reply #112 on: Sat, 01 November 2014, 23:39:03 »
Have you considered making the pinky column a few cm(s) lower to compensate for its shorter length? I understand that each hand/fingers sizes are different among people; however I believe our pinkies are shorter than the other fingers and may benefit with a slightly lowered pinky column keys.

-update
After reading some recent comments, I am surprised to see the mentioning of a lowered pinky column; I thought I was the only one. The ergodox is truly a stepping stone in the right direction for a fine ergonomic keyboard, I think your layout may finish it and bring it very close to perfection. For me the two biggest changes that I would make to my ergodox is, have a closer thumb cluster of buttons, which you've implemented; and second, have a lowered pinky column cluster than the ring fingers.
« Last Edit: Sun, 02 November 2014, 00:45:43 by steve.v »

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #113 on: Sun, 02 November 2014, 00:34:23 »
Have you considered making the pinky column a few cm(s) lower [...] For me the two biggest changes that I would make to my ergodox is, have a closer thumb cluster of buttons, which you've implemented; and second, have a lowered pinky column cluster than the ring fingers.
This prototype had a very aggressively shifted pinky column, which I quite liked, but got feedback from 3 other people who tried it that the pinky column was shifted too much:


So in the next prototype I shifted it back a little bit:

Offline steve.v

  • Posts: 171
fixing the ergodox thumb section
« Reply #114 on: Sun, 02 November 2014, 00:45:57 »
Looking at your mockups I noticed the columns next to the index fingers are of same stagger positions, such as the T G B keys on the left hand; do you think a very small adjustment to have it lowered might be beneficial? I think that due to the natural lateral movements of the index fingers to reach other keys outside of its column it should be slightly lowered.
« Last Edit: Sun, 02 November 2014, 00:51:21 by steve.v »

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #115 on: Sun, 02 November 2014, 01:17:44 »
Looking at your mockups I noticed the columns next to the index fingers are of same stagger positions, such as the T G B keys on the left hand; do you think a very small adjustment to have it lowered might be beneficial? I think that due to the natural lateral movements of the index fingers to reach other keys outside of its column it should be slightly lowered.
Personally I find reaching diagonally sideways-and-upward to be a much easier motion than diagonally sideways-and-downward. Note that with the more aggressive stagger for the middle finger and pinky columns (compared to the ergodox), the hand will naturally fit at a straighter angle relative to the columns, making the direction of index finger motion a bit different than it would be on the ergodox.

It would probably be possible to guarantee every key was ideally reachable with a layout that drops some keys, like this, but then it would be basically impossible to put a QWERTY (or Dvorak, Colemak, etc.) layout on it:
« Last Edit: Sun, 02 November 2014, 01:20:09 by jacobolus »

Offline plainbriny

  • Posts: 192
  • Location: Taiwan
Re: fixing the ergodox thumb section
« Reply #116 on: Sun, 02 November 2014, 01:51:50 »
QWERTY or even Dvoark are not very ergonomic.

I think layout is secondary to the physical arrangement.

I have been considering custom layout on my ergodox seriously,
but haven't got the time and determination to do it.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #117 on: Sun, 02 November 2014, 02:03:29 »
QWERTY or even Dvoark are not very ergonomic.

I think layout is secondary to the physical arrangement.
Absolutely, it makes sense to design a new letter map to go along with a new physical layout, but if one of my design constraints here is to make something that’s a spiritual successor to the Ergodox, I think it’s a bad idea to make it impossible to put QWERTY on it, since it’s entirely reasonable for people to want to move the physical keys but not totally invalidate their muscle memory.

Offline daerid

  • Posts: 4276
  • Location: Denver, CO
    • Rossipedia
Re: fixing the ergodox thumb section
« Reply #118 on: Sat, 22 November 2014, 23:19:43 »
I'm really, really liking the direction this is going in , but I still think I'm going to need a wide spacebar. Lemme see if I can draw something up

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #119 on: Sat, 22 November 2014, 23:53:55 »
If this works out, I would love to make a custom layout Model F with split halves along the lines of:

or

Where the spacebars are Wheelwriter “code” keys. (And with two sets of capacitive pads under any 1x2 keys, in case someone preferred to have two 1x1 keys instead.)

(Since a Model F plate is curved, everything has to be aligned to a grid.)
« Last Edit: Sun, 23 November 2014, 00:01:35 by jacobolus »

Offline steve.v

  • Posts: 171
Re: fixing the ergodox thumb section
« Reply #120 on: Mon, 24 November 2014, 17:30:46 »
I like this very much.

Offline Zustiur

  • Posts: 235
Re: fixing the ergodox thumb section
« Reply #121 on: Tue, 25 November 2014, 00:14:28 »
I have to say that of jacobolus's sketches in the OP, the last of these is the most immediately appealing to me.
I'm still trying to digest the rest of the thread images because change in the bottom row for arrow keys doesn't sit well with me (my arrows are on layer 2, not on the bottom row)

I also have to echo the suggestion to increase the tilt of the thumb cluster relative to the matrix. On my Ergodox, I can only just reach inwards enough (with out leaving home row obviously) to brush the bottom-inward most 1x1 key. The other 1x1s in the thumb cluster are useless to me when touch typing, but I can reach further back towards my stomach quite comfortably.
I would also be in favour of the additional top row for F# keys, but that's a lesser concern, and something I should really fix with my layers.

Perhaps, for my hand and body size, the solution should be to remove the inner most column of the thumb clusters (the 1x1s) and add in the F# row.

I'm also in favour of a slight increase of the stagger, middle finger up a bit, pinky down a bit. Probably only by .025 in each case.

And a request I have not seen mentioned, I'd like the outer most bottom keys (ctrl on a standard keyboard) to be widened as I still insist on pressing these with my palm/knuckle rather than any fingers.

Offline Data

  • Posts: 2608
  • Location: Orlando, FL
Re: fixing the ergodox thumb section
« Reply #122 on: Wed, 26 November 2014, 12:17:48 »
I've been watching this thread for a while.

My only input is that the curved/arced/fanned thumb cluster would work much better, aesthetically, with arc-segmented (think cheese wedge) key caps.  When you use traditional rectangle key caps on them they lose a lot of visual appeal (which is important for attracting people to your new way of doing something), mostly because of the empty space between keys.  I also believe that, by filling in the gaps between keys better you create more surface area with which the thumb can actuate the switch.

Overall I applaud the effort and hope that it continues.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #123 on: Wed, 26 November 2014, 13:16:32 »
My only input is that the curved/arced/fanned thumb cluster would work much better, aesthetically, with arc-segmented (think cheese wedge) key caps.
I totally agree. Know anyone who sells those? :P

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #124 on: Wed, 26 November 2014, 13:29:33 »
mkawa can probably print them if you provide a model
... and shapeaways can definitely print them

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #125 on: Wed, 26 November 2014, 14:09:53 »
mkawa can probably print them if you provide a model
... and shapeaways can definitely print them
Both of these are going to suck compared to injection molded parts.

https://twitter.com/keyboardio/status/535882466934808576

Maybe fine for a one-off personal project where you don’t care too much about quality, but totally unacceptable for a commercial or semi-commercial product. (Same goes for 3d-printed anything to be honest.)

It would be possible to CNC mill acceptable-quality keycaps out of wood or plastic, but that’s way too expensive and time consuming to scale to more than a handful of units.
« Last Edit: Wed, 26 November 2014, 14:13:46 by jacobolus »

Offline Data

  • Posts: 2608
  • Location: Orlando, FL
Re: fixing the ergodox thumb section
« Reply #126 on: Wed, 26 November 2014, 18:14:05 »
Are custom molds an option? How are the guys making artisan caps doing it?

I realize artisan key caps are considered "low volume", but when you get up to a point where you need high volume production any of the injection molding guys should be willing to work with you. I'm pretty sure SP does this kind of work all the time. Up front costs for tooling would be understandably high but the cost per cap should be reasonable, especially if you're doing blanks.

Just a thought.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #127 on: Wed, 26 November 2014, 18:17:47 »
Up front costs for tooling would be understandably high
I’m guessing at least $500, maybe substantially more. But it probably depends on who’s making the tooling.

Offline Data

  • Posts: 2608
  • Location: Orlando, FL
Re: fixing the ergodox thumb section
« Reply #128 on: Wed, 26 November 2014, 18:21:19 »
Up front costs for tooling would be understandably high
I’m guessing at least $500, maybe substantially more. But it probably depends on who’s making the tooling.
I think it's $45 for a new legend, so your estimate is probably close. You'd also need a fairly precise 3D CAD drawing of your cap.

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #129 on: Thu, 27 November 2014, 09:02:56 »
Up front costs for tooling would be understandably high
I’m guessing at least $500, maybe substantially more. But it probably depends on who’s making the tooling.
If it is only in the range of $500 - $1000 then it does not sound bad at all. A group buy should handle that.

A 3dPrinted part does not look nice if one is not wiling to sand and polish it. As for as strength, it is in the range of 60 - 80% of an injection molded part. Since strength grows with square of part thickness, it is easy to compensate for in most cases. 3d printing is not good for higher volume stuff since injection molding gets cheaper with volume.

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #130 on: Thu, 27 November 2014, 18:31:33 »
It would be possible to CNC mill acceptable-quality keycaps out of wood or plastic, but that’s way too expensive and time consuming to scale to more than a handful of units.
Hmm, would CNC milled wood be stronger than 3d printed plastic for small things like keycaps? It looks to me that such a small pieces would like to break along growth rings. I have seen some keycpas made from wood but I did not notice any information about their longevity.

CNC milled plastic should do OK though.

Offline kurplop

  • THE HERO WE DON'T DESERVE
  • Posts: 992
Re: fixing the ergodox thumb section
« Reply #131 on: Thu, 27 November 2014, 22:26:08 »
I agree with Data's concern with the aesthetics of the gaps between the radial keys. Unfortunately the entire layout is riddled with gaps and holes that would require many custom shaped caps to fill in. This defeats the original design concept of using readily available caps to build an ergo-superior board.  If the design shows promise of being mainstream, then it would almost have to tool-up for custom shapes but for now I think it is more important to press ahead with the stock caps.
 
My own experience making wedged shaped caps for radial thumb keys was good. I milled an aluminum prototype to make a silicone mold, then cast the caps in urethane. I was surprised to find that the caps have proved to be not only attractive but quite durable, which was something I had concerns about. The bigger issue might be color matching and texture matching to stock caps.

If enough people sign up for the design, having injection molds is definitely the way to go, although the $500-$1000 estimate to make a mold is probably way too low.  In the meantime, urethane casting might be a good option on a small scale, that would be less expensive than milling and more durable and attractive than 3d printing.

It would be possible to CNC mill acceptable-quality keycaps out of wood or plastic, but that�s way too expensive and time consuming to scale to more than a handful of units.
Hmm, would CNC milled wood be stronger than 3d printed plastic for small things like keycaps? It looks to me that such a small pieces would like to break along growth rings. I have seen some keycpas made from wood but I did not notice any information about their longevity.

CNC milled plastic should do OK though.

I would not normally recommend using wood cross grain for keycap stems but not all woods are created equal. This was demonstrated to me today when my brother in law brought over a 5/8" fine threaded wood dowel today that came out of a 50+ year old motor from the shipyard. The rod was broken to remove it but the threads were intact, super hard and in flawless condition. Probably some variety of ironwood. 

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #132 on: Fri, 28 November 2014, 01:37:27 »
Hmm, would CNC milled wood be stronger than 3d printed plastic for small things like keycaps? It looks to me that such a small pieces would like to break along growth rings. I have seen some keycpas made from wood but I did not notice any information about their longevity.
Depends on the type of wood. Woods have a huge variety in material properties.

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #133 on: Fri, 28 November 2014, 08:20:23 »
The woods I have experience with must be some crappy cheap types.

It is probably better to forget about special keycap shapes for this keyboard if tooling can go above $1000 and more.
Or even better, you can do a poll how much more people are willing to pay for thumb keycaps which do not leave gaps, i.e. keycaps which do not increase ergonomic but improve visual appeal.

I was only surprised the price estimate is the range of low thousands since there was some documentary about LEGO on Discovery and they claimed few millions for tooling for a new brick shape.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #134 on: Fri, 28 November 2014, 16:10:19 »
I was only surprised the price estimate is the range of low thousands since there was some documentary about LEGO on Discovery and they claimed few millions for tooling for a new brick shape.
Tooling costs have a pretty big variation and I’m no expert. The kind of tooling I’m thinking of is a small number of molds for making just one keycap shape, milled out of aluminum, for use on a small-scale injection molding machine. Still, I’m really not sure what tooling would cost. I was putting $500 as a very very low lower bound, in an attempt to point out that even at that price it wouldn’t be worth it without like 100 people who care more enough about keycap aesthetics to spend extra on it (and yet would still not mind having a keyboard with various other odd alignments and gaps). It’s quite plausible that even the cheapest tooling would cost several thousand dollars. You’d have to make a CAD file and shop it around to shops in China to figure out what the actual lower-bound price is.

Lego needs to make gajillions of bricks to unbelievably tight tolerances, so they need to mill molds out of steel for use on super-fancy enormous injection molding machines, and make a whole bunch of them.
« Last Edit: Fri, 28 November 2014, 16:15:21 by jacobolus »

Offline Data

  • Posts: 2608
  • Location: Orlando, FL
Re: fixing the ergodox thumb section
« Reply #135 on: Fri, 28 November 2014, 16:13:17 »
The woods I have experience with must be some crappy cheap types.

It is probably better to forget about special keycap shapes for this keyboard if tooling can go above $1000 and more.
Or even better, you can do a poll how much more people are willing to pay for thumb keycaps which do not leave gaps, i.e. keycaps which do not increase ergonomic but improve visual appeal.

I was only surprised the price estimate is the range of low thousands since there was some documentary about LEGO on Discovery and they claimed few millions for tooling for a new brick shape.
At the tolerances LEGO is working with, and the fact that they replace the molds after a certain number of production runs, that figure doesn't surprise me. This could be a few steps less precise and still fit neatly on an MX stem. But we're really just guessing, honestly.

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #136 on: Fri, 28 November 2014, 16:20:12 »
The woods I have experience with must be some crappy cheap types.
Just because a wood doesn’t have extremely fine grain, or isn’t absurdly hard, doesn’t mean it’s “crappy”. Different woods are good for different things. I wouldn’t try to make a model airplane out of ebony, I wouldn’t try to make a baseball bat out of balsa wood, and I wouldn’t try to make a keycap out of pine.

Check out e.g. this chart: http://www.nrs.fs.fed.us/pubs/rn/rn_nrs38.pdf

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #137 on: Fri, 28 November 2014, 18:18:30 »
Jacobolus, that document is interesting but unfortunately it is geared towards lumbering since it contains things like densities, bark ratio, and moisture content. Hardly relevant for cap making.

The interesting stuff is young modulus across and along the grain. ABS plastics has it at about 2GPa and it is uniform. Oak (probably one of the stronger woods) has it at 11GPa along grain. That looks great. But the problem is you cannot CNC keycap all along grain. So the number which is actually interesting is young modulus across grain. And I did not find that number on-line quickly. And I'm afraid it will be significantly lower. At least that is what my experience hammering nails into wood indicates. One should be careful not to drive a nail along grain in thin pieces. Otherwise bad things happen :)

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #138 on: Fri, 28 November 2014, 21:05:13 »
Jacobolus, that document is interesting but unfortunately it is geared towards lumbering
Sure, that chart is only really talking about specific gravity. My point is just that there’s a huge variation from one type of wood to another. There’s similar variation in many other properties, including tensile and compressive strength with and against the grain, etc.

Here’s a better set of charts:
http://www.fpl.fs.fed.us/documnts/fplgtr/fplgtr190/chapter_05.pdf

The best woods for keycaps (especially the ability to machine small details accurately) are probably various imported hardwoods like rosewood, bubinga, teak, ebony, purpleheart, paduak, bocote, etc. I’m not really an expert though. It would take some experimenting to figure out which woods worked best.
« Last Edit: Fri, 28 November 2014, 21:16:51 by jacobolus »

Offline vvp

  • Posts: 886
Re: fixing the ergodox thumb section
« Reply #139 on: Sat, 29 November 2014, 11:52:03 »
Ok, that is a better document. Thanks. On page 5-26 we get Overcup Oak tensile strength parallel to grain of 77.9 MPa, on page 5-6 we get Overcup Oak tensile strength perpendicular to grain of 5 MPa. ABS plastics has tensile strength of 40 MPa. So ABS keycaps are about 8 times stronger than Overcup Oak keycaps (provided the dimensions are the same).

The best wood I noticed in that document was American Beech with tensile strength perpendicular to grain of 7 MPa. Still significantly worse than ABS. This also indicates that ABS 3d printed keycpas (strength at least of 24MPa) would be significantly stronger than the woods in the document which have tensile strength specified.

As jacobolus indicated, some exotic woods may be better and some of the woods in his list are even mentioned in the document but none of them has tensile strength specified.

Anyway, CNC machined wood keycaps will probably have as thick stem as my 3d printed examples so they should do OK compared to off-the-shelf ABS keycaps. We need only √8 ≅ 2.8 times thicker stem which is easily doable.

What is more interesting for me is that wood is about 16 times less strong perpendicular to grain compared to parallel direction. That is the number I wanted to find. And my experience driving nails into wood is now theoreticaly supported thanks to jacobolus' data :D

Offline berserkfan

  • Posts: 2135
  • Location: Not CONUS Not CONUS Not CONUS Not CONUS
  • changing diapers is more fun than model f assembly
Re: fixing the ergodox thumb section
« Reply #140 on: Mon, 01 December 2014, 20:20:25 »
just how the heck did I miss this?

YES, I am one of those people who went and bought an ergodox (because in my case it was cheap) and found that the thumb cluster distinctly not to my liking!!!!
Most of the modding can be done on your own once you break through the psychological barriers.

Offline teshdor

  • Posts: 8
Re: fixing the ergodox thumb section
« Reply #141 on: Wed, 10 December 2014, 14:03:08 »
jacobolus,

What do you think of this layout. This adds a space for a directional keypad, with the loss of only one non alphanumeric key. If you type in QWERTY, this would be the right hand "/ ?", if you type in Dvorak it would the left hand "; :". If you didn't want to add directional, it would still work without it. I wasn't a big fan of the staggered number keys, so I straightened those out. The thumb section is perfect with your new layout. The two 1.5U left most thumb bars, could be 2U (The 2U felt slightly more comfortable against my thumb), with the two 1.5U bars to the left of the home row moved up to make room for the 1U on top of the thumb bars. The pinky 1.5U bars are replaced with 1U, but they would probably be just as good with 1.5, as  before. If it is possible, the teensy/ microcontroller could be moved from the top, to a vertical position in the top left corner. Then, the PCB could be made smaller, and a potential to plug the USB cable directly into the teensy. Comments?

83845-0
« Last Edit: Wed, 10 December 2014, 14:11:42 by teshdor »

Offline luisbg

  • Posts: 248
  • Location: London
Re: fixing the ergodox thumb section
« Reply #142 on: Wed, 10 December 2014, 15:45:46 »
jacobolus,

What do you think of this layout. This adds a space for a directional keypad, with the loss of only one non alphanumeric key. If you type in QWERTY, this would be the right hand "/ ?", if you type in Dvorak it would the left hand "; :". If you didn't want to add directional, it would still work without it. I wasn't a big fan of the staggered number keys, so I straightened those out. The thumb section is perfect with your new layout. The two 1.5U left most thumb bars, could be 2U (The 2U felt slightly more comfortable against my thumb), with the two 1.5U bars to the left of the home row moved up to make room for the 1U on top of the thumb bars. The pinky 1.5U bars are replaced with 1U, but they would probably be just as good with 1.5, as  before. If it is possible, the teensy/ microcontroller could be moved from the top, to a vertical position in the top left corner. Then, the PCB could be made smaller, and a potential to plug the USB cable directly into the teensy. Comments?

(Attachment Link)

I like this very much. Considering building this from scratch.

I would probably remove the arrow keys but I understand a lot of people love them down there.
Leopold FC660M - Brown mx switches - black case - white blank keys :: ErgoDox - Blue mx switches - classic case - black blank keys

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #143 on: Wed, 10 December 2014, 16:26:46 »
What do you think of this layout.

Let’s directly compare and note changes you made:

- Made the top left index finger key further to reach. I’m -1 on that
- Reduced the column stagger on the pinky column. I’m -1 on that
- Shifted the far pinky column up quite a bit. Now the top two pinky keys in that column are very hard to reach IMO. I’m -1 on that.
- Added dedicated arrows. Personally I don’t have need for dedicated arrows, so I’m +0 on that.
- Switched the corner pinky key from 1.5u to 1u. I’m -0 on that. I think the previous one is pretty good for a shift or other modifier key.
- Put the number row into alignment with the columns. Personally I think this is worse functionally, but possibly better aesthetically (though there are so many gaps and funny shapes on this keyboard that I don’t think it’s a big deal. I’m -0.5 on that change.

I’m guessing I personally would prefer my version, but I haven’t physically tried yours, so it’s hard to be sure.

I personally think the best way to best meet the needs/preferences of different folks is to have one very compact version without dedicated arrows, and one larger version with an extra pinky column, an extra thumb column, a number row that doesn’t have gaps in it, dedicated arrows, and possibly F keys. On that one, it should be possible to basically copy most of the layout over from an existing TKL board.

Anyway though, don’t take my criticism personally: I’m just telling you my own personal preferences, and a lot of this is pretty subjective. Thanks very much for proposing something, and I hope you and other folks keep tweaking and offering alternative suggestions.

Quote
If it is possible, the teensy/ microcontroller could be moved from the top, to a vertical position in the top left corner. Then, the PCB could be made smaller, and a potential to plug the USB cable directly into the teensy.
If this makes it to production it would be an ARM chip directly on the board, and they’re so small they can fit essentially anywhere. No need to actively worry about where to fit it until we actually start designing a PCB.

Offline luisbg

  • Posts: 248
  • Location: London
Re: fixing the ergodox thumb section
« Reply #144 on: Wed, 10 December 2014, 16:41:21 »
If this makes it to production it would be an ARM chip directly on the board, and they’re so small they can fit essentially anywhere. No need to actively worry about where to fit it until we actually start designing a PCB.

ARM? I thought it would be Teensy/Arduino like most customs. That way we can reuse most of the firmware source code (and forks) we already have.
Leopold FC660M - Brown mx switches - black case - white blank keys :: ErgoDox - Blue mx switches - classic case - black blank keys

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #145 on: Wed, 10 December 2014, 17:20:38 »
Oh also, teshdor, welcome to geekhack!

Offline luisbg

  • Posts: 248
  • Location: London
Re: fixing the ergodox thumb section
« Reply #146 on: Wed, 10 December 2014, 17:40:26 »
Welcome teshdor! First post :)
Leopold FC660M - Brown mx switches - black case - white blank keys :: ErgoDox - Blue mx switches - classic case - black blank keys

Offline jacobolus

  • Thread Starter
  • Posts: 3661
  • Location: San Francisco, CA
Re: fixing the ergodox thumb section
« Reply #147 on: Wed, 10 December 2014, 18:52:22 »
ARM? I thought it would be Teensy/Arduino like most customs. That way we can reuse most of the firmware source code (and forks) we already have.
HaaTa’s firmware works on ARM. ARM chips are just as cheap (or even slightly cheaper), just as small, and much more powerful, and there’s a lot more work going into the ARM world these days. I think it makes more sense as a way forward in the future.

Offline teshdor

  • Posts: 8
Re: fixing the ergodox thumb section
« Reply #148 on: Wed, 10 December 2014, 21:05:35 »
jacobolus,

Thank you for your through comparison, between these layouts. I printed out a paper version of your layout and tried it out. I used DSA keycaps, as that is all I have. I'm not sure what percentage of Ergodox users use DSA, but due to the fact they were easier to obtain, cheaper in cost, and ability to freely customize their locations, that is what I have. Note,  I have medium/small hands. This is what I found,

Quote
Made the top left index finger key further to reach. I’m -1 on that
I agree with you here, I think reach in your original version, outweighs having bulkier keys.

Quote
Reduced the column stagger on the pinky column. I’m -1 on that
I also agree here, the pinky column stagger was ok, and put noticeably less strain on my pinky overall throughout hand movements.

Quote
Added dedicated arrows. Personally I don’t have need for dedicated arrows, so I’m +0 on that.
Arrow keys definitely arn't for everyone. From your previous designs, it looks like you tried to incorporate a directional pad similiar to the one on the Truly Ergonomic Keyboard. I think that style of directional pad is too close to where the wrist sits, and could be uncomfortable for some users. My version was there to be optional (as it used one of the primary pink keys), but integrated enough, so it would be out of the way of the wrist.

Quote
Switched the corner pinky key from 1.5u to 1u. I’m -0 on that. I think the previous one is pretty good for a shift or other modifier key.
I agree, bring back the 1.5U, in fact bring back a whole row of 1.5U. Going back to the original ergodox design, one part I really like is the chunky 1.5U rightmost keys (especially on my left hand, for ctrl, shift, and escape). Although, they are not as easy to reach due to them being in pinky range, they are large enough to side palm, and I like to move my entire hand and take my index or middle finger and hold them down for some shortcuts. This is a personal preference, and though these keys could be mapped to the thumb area. I really enjoy them. I would like to see how otherwise view this row of keys.

Quote
Put the number row into alignment with the columns. Personally I think this is worse functionally, but possibly better aesthetically (though there are so many gaps and funny shapes on this keyboard that I don’t think it’s a big deal. I’m -0.5 on that change.
Here is a picture of the columns in the staggered layout, as you have them now.
83894-0
Testing each key for comfort, 7, 8 and 9 all felt quite comfortable. Looking at the natural curve of my hand, I instantly could tell why. As my fingers extend they naturally spread apart, and land almost perfectly with the the locations of the 7, 8 and 9 staggered keys. However, when it came to the 6 and 0, each was somewhat painful (The 6 more than the 0). The 7 key is in a perfect spot for the extension of the index finger, but the 6 key is a stretch. In a non-staggered layout, my fingers were more cramped in the middle keys, but do not have to stretch as far for the outer numbers. Personally, I like the numbers, non staggered, as close to home row as possible. I urge everyone in this forum to try it for themselves, and report their hand size as well to see how they like it.

I am sure that staggered layout is good for some people, so I made a few changes to your original layout to come up with this,
83896-1

Since I prefer the non staggered layout, I just alligned the number keys and came up with this (adding right side 1.5U keys, with a 1U for delete,
83898-2
To fill it out, I added an additional top left 1U key and extended the 1U top right to a 1.5U. These buttons are not easily pressed, but makes it look nice, and allow sparingly used keys, that are annoying when you can't find them, like ~.

Thank you, and let me know what you think.
« Last Edit: Wed, 10 December 2014, 21:55:52 by teshdor »

Offline teshdor

  • Posts: 8
Re: fixing the ergodox thumb section
« Reply #149 on: Wed, 10 December 2014, 21:30:53 »
Also, just because I could...

...bam.....Numpad! and 3 1U to the top left.