(Model MF) Remodeling the Model M (aka.. the Mara)

User avatar
lot_lizard

12 Sep 2016, 22:23

E TwentyNine wrote: Would some kind of hi-res halftone dot approach give more control?
It's hard to say yet. The software that drives the laser is a print driver that turns greyscale into dithering (along with variable wattage and travel speed). I would have to play with it more, but custom dithering is certainly an idea if it doesn't behave like I am wanting. All and all, I'm impressed with their (Epilog's) approach though. It certainly provides options to your needs... either "just go did this" if engraving something hard, or a large amount of variable "knobs" for softer materials. The color scale is 16-bits. These 16-bits get converted to tones of grey, and each tone of grey can be manipulated for speed and wattage.


If I can't get the crayon dialed in pretty close, it is time to blame it on the crayon and move to another medium :)

User avatar
lot_lizard

13 Sep 2016, 02:26

Decided to get rid of the F and work on the hard part (extended the M backdrop to even make it more problematic for testing). Getting decent at this, but think I've pushed the crayon to its limit.

Time to do some actual research about dye-subbing plastics with a laser (vs. just making $hit up). Unfortunately, this is pretty uncharted territory. We know enough about the properties of everything at this point to make some educated experiments going forward, but I would plan on Hajime being our go-to for MF caps unless a discovery is made (pretty sure we are the first to dye-sub PBT with Crayola with reasonable results... so it could happen), but who knows at this point. Orange turned out really nice btw.

At 1200 dpi with vector graphics, I know this is doable... Just have to find the right medium for the dye. The crayon, if I get the dye color right, distorts the resolution. I can get higher resolution, but bubble the wax causing color changes. If nothing else, a fun experiment
image.jpeg
image.jpeg (665.21 KiB) Viewed 5692 times

User avatar
wcass

13 Sep 2016, 02:32

i have some caps that are dyed. The dye only penetrates a few microns deep but is very long wearing. What would happen if you hit that with a laser? I would like to send you some purples and blacks (pearl and pebble underneath) to play with.
caps.jpg
caps.jpg (572.41 KiB) Viewed 5665 times

User avatar
phosphorglow

13 Sep 2016, 03:18

lot_lizard wrote: Haven't you been a busy little beaver. That's awesome progress, and great ideas on the case finishing options. Just curious, how many of these could be made in a week without sacrificing quality of the product, or (more importantly) your quality of life?.
Thanks! Well, once all the hard parts are out of the way, painting the mold and injecting the resin doesn't take all that long. Most of the wait involves letting them cure, so with whatever time constraints I have I'll make a cautious guess of about 5 a week. Possibly more if de-mold time is reasonable and I can clean up the flashing fairly easily.. I'll cast 'em until the mold wears out... and then make another mold. Heck, I may even make another mold while casting from the other one. Haven't had time yet today to make the new support shell, but hope to before zonking out for the night.
lot_lizard wrote: Do you mind if I send you one of each of the remaining daughter boards (USB-C/mini/micro) along with the controller PCBs? Once I have a finished product there, I will finish the port covers for the other 3. Let me know what you have in mind for merging. I would love to minimize variations where it made sense (saves everyone time and cost)
Sure, send 'em along! Hopefully the USB-C port is the hybrid through-hole/SMT style...? As far as merging, I suppose the best contributions to merge are the dimensions I know to physically work. We'll figure it out!

Anyhoo! Super exciting to hear that the engraver is so versatile. Definitely opens up some neat possibilities then.

The keycaps are looking REALLY promising considering you used a crayon *and* it stayed after industrial solvent. Crayola me impressed.

User avatar
E TwentyNine

13 Sep 2016, 04:03

Ellipse posted some dye-sub work going on over in the F77 thread, might be worth checking out:

group-buys-f50/brand-new-f62-kishsaver- ... ml#p328761

User avatar
lot_lizard

13 Sep 2016, 17:04

Added the SVG for the MF logo (key cap) to the GitHub repository. The document is saved as the size I need for cap testing currently, but it's SVG, so you can scale up without loss. Included a png representation in the repository as well (shown below).

https://github.com/lot-lizard/Model-MF/ ... ter/mf.svg

The more I look into our laser options, and the properties of PBT, I think we might actually be able to pull off pretty much anything we want to do with practice. Standard marking (black or close to) and foaming (frothy white) should be slam dunks, but even the color "sublimation" I was trying earlier looks very promising with proper additives. Pigmenting with various colors is a relatively new concept, so would be on our own for research... but the additive productive manufacturers should be able to give us a guess at where to start. Along with speed and power, wavelength looks like it will play a big part there (similarly to if you wanted to etch inside of glass). The problem is that we are not a thermoplastic with good properties for lasers. They have concocted polymers that make it easy to work with lasers, but PBT is almost on the opposite end of that scale. We'll get it figured out, but the window for tweaking here is much less forgiving.

WCass, I did try some cutting/engraving on a couple of caps last night, and was able to get through what I believe would be the depth of your dyed keys without melting the PBT. If we melt, it will just mix your dye with the underlying PBT ivory you are wanting to show through. There is a really fine line there it seems, but the results I believe were good.

Image

User avatar
lot_lizard

13 Sep 2016, 22:02

Got to play for a few minutes over lunch again today. This is about as deep as we can go before structural issues in the cap would become a consideration. @WCass... I believe about 15-20% of this depth is all it would take to reliably eat cleanly through the dyes you had pictured, but I wanted to see how much of a relief we could really give it. It feels lovely to the touch if wondering.

EDIT: updated pics
Correct color, but blurry
Correct color, but blurry
image.jpeg (335.56 KiB) Viewed 5468 times
Correct color, but blurry
Correct color, but blurry
image.jpeg (405.85 KiB) Viewed 5468 times
Horrible color, but detail
Horrible color, but detail
image.jpeg (416.09 KiB) Viewed 5446 times
Horrible color, but detail
Horrible color, but detail
image.jpeg (457.58 KiB) Viewed 5446 times
Last edited by lot_lizard on 13 Sep 2016, 22:48, edited 2 times in total.

User avatar
Techno Trousers
100,000,000 actuations

13 Sep 2016, 22:25

Woah, that looks really cool! Is that the thinnest you can make the lines? Like, could you actually do an infilled Shift key? I'm picturing infilled white on black, though I'm sure the cost and effort would be prohibitive.

User avatar
E TwentyNine

13 Sep 2016, 22:42

That is flipping amazing. Seriously a new frontier for BS caps.

What I'm picturing is keys dyed dark grey then laser etched. Filling in the etching with white wax ala my old D&D dice or some other more permanent substance. Selectric looking keys, white on grey. :mrgreen:

User avatar
lot_lizard

13 Sep 2016, 22:57

Techno Trousers wrote: Is that the thinnest you can make the lines?
About ~1/2 the size of the F is probably realistic for a clean cut (attached a pic below for scale). 1200 dpi
Techno Trousers wrote: Like, could you actually do an infilled Shift key? I'm picturing infilled white on black, though I'm sure the cost and effort would be prohibitive.
E TwentyNine wrote: Filling in the etching with white wax ala my old D&D dice or some other more permanent substance. Selectric looking keys, white on grey. :mrgreen:
I believe this was where WCass was heading when he was thinking about dyeing the caps some color, and then etching just enough off to have the original color underneath. I got a little carried away just wondering how deep I could go. It would be about 15% of that depth to get rid of the dye in his previous pics I think (and that might be a big number... won't know till I try on the caps he is sending). Wouldn't take much. If nothing else, a fun experiment.

I think it would look nice even the same color all the way through on blue, purple, green, red, etc. in relief. Or as you say, if you come up with a clever way to fill, but that might get tricky because of the concaved surface of the key. Some brilliant white black caps dyed to the desired color, and then relief-ed just enough with the laser to bring the white back would be very doable. Could line up an entire set and pop at once with a trivial jig. As you were asking before, designs in relief would be pretty easy here too (now that I understand the settings).

The laser has the MASSIVE upside over dye-sub in that alignment will never be off (no sheet to position). So no wonky Unicomp alignment issues. Good times ahead I think.

EDIT: and with 2-piece caps, there is really no reason we couldn't cut all the way through in some pattern or font as well. Couldn't do with 1-piece, but might be a fun effect on 2-piece.
Screen Shot 2016-09-13 at 3.35.10 PM.png
Screen Shot 2016-09-13 at 3.35.10 PM.png (34.74 KiB) Viewed 5441 times
Last edited by lot_lizard on 14 Sep 2016, 09:24, edited 2 times in total.

User avatar
E TwentyNine

13 Sep 2016, 23:00

How does it handle curves?

User avatar
lot_lizard

13 Sep 2016, 23:13

E TwentyNine wrote: How does it handle curves?
Really well... Better than I expected. You have to have quite a variance in depth before the laser becomes out of focus (think like 1/4 inch). The cap looks good in person. It would be higher detail if not so deep (just the nature of anything that burns/heats).


But personally... I think it is as crisp as any dye-sub technique. There you have bleed (especially with a rough surface), and would explain why Unicomp struggles to not be blurry

EDIT: 1/2 inch should have been 1/4 inch above (correcting). I have to think at 1/2 inch it would get blurry, but didn't try

User avatar
lot_lizard

14 Sep 2016, 04:17

E TwentyNine wrote: How does it handle curves?
It dawned on me that you might be referring to curves in the 2D and not the 3D (my last response was about how forgiving the laser was on the z-axis for differences in key cap height). Here is an Inkscape vector I threw together to test 2-D circles (smallest inners are 1mm). It would get tighter, but anything less is a little silly. We will only get better with experience (this is day 2 mind you).


The texture feels great for something like the Esc (wouldn't want out in the middle of the board myself). Moral... I think we can dream up some fun stuff here. Do patterns, dye the entire thing afterwards, then Selectric font the white back (preferably we started with brilliant white instead of ivory)... sky is the limit really.

EDIT: I should mention that my phone is crap at taking pictures of these things. They look much better in real life. Might should try videos to see the texture. For some reason, the pictures look much less defined and uneven. Not sure why... but an optical illusion is occurring.
IMG_8553.jpg
IMG_8553.jpg (1.4 MiB) Viewed 5486 times
Last edited by lot_lizard on 14 Sep 2016, 04:43, edited 1 time in total.

User avatar
E TwentyNine

14 Sep 2016, 04:37

Below is a colorway I've been thinking of for a long time now, as a throwback to the old selectrics with that scheme (plus the dark green accents. Ignore the nonstandard layout, I'd want classic SSK, no winkeys). With selectric fonts could be even better. Never thought it would be anywhere near possible.

Image

And if that laser could be successfully trained on an SSK case or maybe a Phosphor resin creation, I could get my logo too... :mrgreen:

User avatar
lot_lizard

14 Sep 2016, 04:55

E TwentyNine wrote: Below is a colorway I've been thinking of for a long time now, as a throwback to the old selectrics with that scheme (plus the dark green accents. Ignore the nonstandard layout, I'd want classic SSK, no winkeys). With selectric fonts could be even better. Never thought it would be anywhere near possible.
...
And if that laser could be successfully trained on an SSK case or maybe a Phosphor resin creation, I could get my logo too... :mrgreen:
It's a nice look for sure. I think we are getting there. Have to prove the dye and then "bring back the white font" concept, but on paper it is very sound. I did try dye-subbing with a crayon the entire cap, and then brought back an ivory letter with very good results (without going very deep). I would be taking more pictures of all of this, but they are just turning out "less than reality"... so hesitant to upload.


If we do prove all of this, I will put together a little key jig for the laser bed with spacing, and put together a SVG template that would let you set your own fonts I think. If you dyed the caps ahead of time, told me what color was underneath, and filled in the SVG template for me... I would be happy to load it and click "run". A couple of test caps with what you send me is probably ideal to make sure you like the prototype results. Seems like a reasonable model that wouldn't take any of my time really (once all the patterns were made... 10-15 minutes max per run). Would be happy to. I doubt I would want to do some "production run", but a personal board would be fun.

Your SSK logo would be pretty easy as well. Getting black into plastic is pretty easy at this point. Multiple attempts have turned out favorable (with an without the crayon ;) ). We are onto something I think

User avatar
Techno Trousers
100,000,000 actuations

14 Sep 2016, 05:17

lot_lizard wrote: If we do prove all of this, I will put together a little key jig for the laser bed with spacing, and put together a SVG template that would let you set your own fonts I think. If you dyed the caps ahead of time, told me what color was underneath, and filled in the SVG template for me... I would be happy to load it and click "run".
** drooling **

Maybe I'll be able to do a set of white on blue mods. That's been my personal dream to combine with the Unicomp blue on white alphas. I guess I'd need a set of brilliant white modifiers, then dye them an appropriate color of blue for lasering. A real buckling spring "red alert" set should be possible too. I can't wait to see how this all comes out.

User avatar
E TwentyNine

14 Sep 2016, 13:40

lot_lizard wrote: I did try dye-subbing with a crayon the entire cap, and then brought back an ivory letter with very good results (without going very deep). I would be taking more pictures of all of this, but they are just turning out "less than reality"... so hesitant to upload.
Upload them anyway. And I've been meaning to ask - how do you dyesub with a crayon? How did you do the early MF key with crayon creation?

User avatar
lot_lizard

14 Sep 2016, 16:38

E TwentyNine wrote:
lot_lizard wrote: I did try dye-subbing with a crayon the entire cap, and then brought back an ivory letter with very good results (without going very deep). I would be taking more pictures of all of this, but they are just turning out "less than reality"... so hesitant to upload.
Upload them anyway. And I've been meaning to ask - how do you dyesub with a crayon? How did you do the early MF key with crayon creation?
I'll dig the camera out and see if I can do a little better job. All photos so far have been shot with a phone.


Verbose "nerd speak" to follow... proceed with caution.

I'm using the term "dye-sub" VERY loosely just because that is the process we are familiar with. This uses a very different transfer process, but yields similar results. In dye-sublimination we use pressure to direct the "dye gas" after it leaves the transfer sheet. The heat required both turns the dye from solid (on the transfer sheet) to gas, but it also de-bonds the polymer enough to allow absorption without being so hot that thermal degradation occurs. When the polymer cools, the bonds are reformed. The only real downside to the process is that some of the gas is diffused to the surrounding polymer (no way to create truly hard lines).

Keep in mind this next part is with a crayon (not an ideal pigment additive).

A laser (when engraving/cutting) is actually vaporizing (boiling) the solid through a tiny hole (the resolution of the beam). In an ideal world, you want to completely cut/engrave everything on a single pass because the composition of the material might change when cooled and cause reflectivity on the next pass (this is ALWAYS minor, EXCEPT IN OUR CASE). For my little silly experiment with the crayon, I am changing frequency/power of the beam in such a way that the underlying PBT is never vaporized, but becomes a liquid crystalline. The crayon has a much wider temperature range where it remains a liquid (lower melting point / higher boiling point). The beam is able to shoot through the liquid crayon, and heat the PBT without impacting the chemical properties of the crayon melt. Eventually the PBT melts enough to absorb the pigment and we move the laser to the next "keyhole". Back to doing in one pass, since we are mixing materials, we now have a different makeup that we would need to melt since the crayon is now physically mixed with the polymer (bonding is very different than with the PBT alone). So all of the settings would need to be adjusted for the second pass. This is different than engraving say metal where you would just keep pounding away. If we hit the PBT too hard, it vaporizes, but the vapor would have no where to go with the crayon melt above (creating a discolored mess).

So we are similar to the "dye-sub" technique, with a couple distinctions...
  • pigment is transferred as a gas in dye-sub, liquid in our case
  • dye-sub uses pressure to direct the gas, where we essentially just rely on gravity (think water seeping into dirt on a molecular level)
  • dye-sub diffuses pigment into unwanted areas because the entire polymer is heated at once, where we only heat the size of the beam
Again... we will find proper additives for pigment that were made for this process. I doubt the crayon will yield the best results possible. Just a fun experiment.

User avatar
lot_lizard

14 Sep 2016, 22:08

Couple of bits. I just acquired 8 XT's for a song from a local. So they will be joining the switch pool, and dropping the per switch price a couple of cents.

FINALLY got around to getting the split spacebar sizes over to WCass and i$ for the additional PCB traces/pads (sorry guys). As part of testing and finalizing key spacing, I decided to test fit another shell and stole a couple of more keys from the Wheelwriter (note that are not inserted into barrel cutouts, and are just resting to get the idea if they looked misaligned). I like the look, and not sure if I've ever seen the industrial blues on a white shell... so thought I would share. Phone potatoed the blues a bit, but the idea was captured. The black badge is the WCass PCB badge from the Seebart buy. We still need to have the smaller SSK version made.
FullSizeRender-6.jpg
FullSizeRender-6.jpg (1.14 MiB) Viewed 5271 times
FullSizeRender-7.jpg
FullSizeRender-7.jpg (1.17 MiB) Viewed 5271 times
IMG_8593.JPG
IMG_8593.JPG (2.3 MiB) Viewed 5271 times
IMG_8598.JPG
IMG_8598.JPG (2.22 MiB) Viewed 5271 times

andrewjoy

14 Sep 2016, 22:10

lot_lizard wrote: Couple of bits. I just acquired 8 XT's for a song from a local. So they will be joining the switch pool, and dropping the per switch price a couple of cents.

I will be willing to buy as many XT key-sets as you are willing to sell.

I love me them 1 part caps .

User avatar
E TwentyNine

14 Sep 2016, 22:29

lot_lizard wrote: I decided to test fit another shell and stole a couple of more keys from the Wheelwriter
Boy, you threw the whole gamut there, Mopar Blues*, APL, split spacebar with the Code key even, that Reset is off a terminal isn't it?

Looks good.

(* a little terminology nit on the blue keys, which keep getting called "Industrial Blues" by multiple people. They were seen on industrial keyboards, but they were in a very specific setting, an auto shop/dealership. They're from a Mopar Diagnostic machine that used an industrial keyboard with some custom caps. I prefer to call them Mopar Blues.)

User avatar
seebart
Offtopicthority Instigator

14 Sep 2016, 22:35

lot_lizard wrote: The black badge is the WCass PCB badge from the Seebart buy. We still need to have the smaller SSK version made.
That's right because that badge looks silly on that otherwise awesome SSK. That bottom row. :shock:
E TwentyNine wrote: but they were in a very specific setting, an auto shop/dealership. They're from a Mopar Diagnostic machine that used an industrial keyboard with some custom caps. I prefer to call them Mopar Blues.)
You know that for sure? I'm just asking. I've heard several people say that. IBM made them specifically for Mopar shops?

User avatar
E TwentyNine

14 Sep 2016, 22:55

seebart wrote:
E TwentyNine wrote: but they were in a very specific setting, an auto shop/dealership. They're from a Mopar Diagnostic machine that used an industrial keyboard with some custom caps. I prefer to call them Mopar Blues.)
You know that for sure? I'm just asking. I've heard several people say that. IBM made them specifically for Mopar shops?
Chrysler/Mopar had a deal with IBM to provide these things, with full industrial components.

Here's a page with one: http://auction.repocast.com/details.cfm ... 791#photos

I believe I had seen the announcement letter for it ages ago but no luck finding it today.

User avatar
seebart
Offtopicthority Instigator

14 Sep 2016, 22:59

Good to know, thanks for that link. I can update our SSK wiki page a bit then.

wiki/IBM_Space_Saving_Keyboard

User avatar
lot_lizard

14 Sep 2016, 23:16

seebart wrote: That bottom row. :shock:
I would prefer if the shift/tab were blanks, but didn't have on hand. I do like the idea of them matching the original space bar vs being a pebble color. Would be nice if it had the same rounded profile as the Code key, but those are first world problems. In the end I would likely order pearl blanks for the right shift and 1.5u from Unicomp (or just sandblast text off ala-Wodan Style).
E TwentyNine wrote: that Reset is off a terminal isn't it?
It's part of the same Model M APL set as the backspace and other keys. It's a fun set.


Ultimate irony... mixed keys make me a little nauseous. Not sure why, but never cared for it. Figured if there was ever a time, it would be these pics... but to be honest... its growing on me. Just someone please stop me before it turns into this
E TwentyNine wrote: Here's a page with one: http://auction.repocast.com/details.cfm ... 791#photos
I think Great/Interesting finds dropped the ball on that one. The machine scrap weight is worth 4x the final bid.

User avatar
lot_lizard

15 Sep 2016, 18:17

Been on a call this morning where I am just a warm body, so threw together the start of a project task list to help coordinate this a bit better going forward. The following data feed is about the only thing that is visible publicly to the world without you having a freedcamp account, and it is basically worthless lacking any sorts or filters. If you would like to be invited as a "contributor" where it is much easier to see status, PM me. If nothing else, the data feed spreadsheet gives an idea of scope if anyone ever debates about taking something like this on.

Freedcamp is a great little tool for silly stuff if anyone is ever looking to learn a "free" tasks tool. I am very pessimistic about when something is marked complete if you are concerned on progress, and pay no attention to the dates or owners. Just a nice checklist

Live Data Feed: https://freedcamp.com/api/v1/csv/f8fc95 ... b127a3f36b

The following are portions of the screen that I see to give you a better idea:
Spoiler:
Kanban perspective (also has a tasks list perspective if that is more comforting)
Kanban perspective (also has a tasks list perspective if that is more comforting)
Screen Shot 2016-09-15 at 10.47.54 AM.png (134.59 KiB) Viewed 5195 times
Details of a task
Details of a task
Screen Shot 2016-09-15 at 10.57.17 AM.png (281.7 KiB) Viewed 5195 times

User avatar
elecplus

15 Sep 2016, 19:10

Freedcamp looks interesting. I use Dolibarr, which is a full blown open source ERP that runs on any OS, or you can install on a web site. Tasks can be assigned to people or groups, BOM for parts listing for builds, etc. How many complete units of X can you build from the parts on hand? How many more of Y parts are needed? Contacts can have assignments. Almost anything can have custom fields added. Tracks sales, orders, quotes, purchases, etc. Tracks vendors and samples and all contacts and correspondence. Just a thought.

User avatar
lot_lizard

15 Sep 2016, 20:51

elecplus wrote: I use Dolibarr, which is a full blown open source ERP that runs on any OS, or you can install on a web site.
Yeah... you have something way more high-test than this. Think of these "task tools" like glorified Post-It note boards representing queues. I take off my board, put it in the queue on someone else's, etc. Really focused on "what's next?!?" for teams, and that's about it. Trello is very similar, but a little less feature rich.


They are really just virtual Kanban boards, which is really the heart of "lean". Would be tough to run an entire business from (unless maybe the business was selling Girl Scout cookies). Good for small projects though

User avatar
lot_lizard

16 Sep 2016, 23:08

Foam samples from our supplier came in, and had the opportunity to try the cutter on something serious. This is 3/32 (2.35mm) silicon rubber. It took 3 passes on a single side but we eventually got through (each pass cuts less deep and less clean because of debris). 2.35mm of 50 durometer A shore hardness silicon is at least 1/3 too thick, but wanted to see what this thing could handle. I THINK we could get through 1/16 on a clean single first pass (the thickness of our actual target), but worst case we would flip and cut clean through from the back on a second pass. I can crank it up higher, but want to be prudent of machine wear and fire hazards

It's a great material (same o-rings your used to), but on a nice scale and should work out nicely. Hitting once from each side will produce sharp corners. I did try already, and confirmed it is too thick. 1/16 will be our sweet spot.
image.jpeg
image.jpeg (376.44 KiB) Viewed 5053 times
Last edited by lot_lizard on 17 Sep 2016, 00:04, edited 1 time in total.

User avatar
Techno Trousers
100,000,000 actuations

16 Sep 2016, 23:42

Always great to see validation of a theory. It looks like you even got some charring on that test piece. Burn, baby, burn!

Post Reply

Return to “Workshop”