geekhack

geekhack Community => Keyboard Keycaps => Topic started by: xiphmont on Mon, 10 September 2012, 04:59:40

Title: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Mon, 10 September 2012, 04:59:40
Partly to augment my own faulty memory, partly for others to find...

I've been playing with dyeing Model M keycaps.  There have been plenty of threads on the subject, but really oddly little methodical experimentation or hard recipes.  I'm working on some dye-jobs, so here's a few notes so far I've not seen mentioned elsewhere...

1) PBT is a polyester family plastic, it requires a dispersion dye.
     1a) Dispersion dyes require high temperatures; even boiling is barely high enough to do anything
     1b) Vinegar (acetic acid) and salt are useless for dyeing PBT
     1c) Some brands of dyes seem to include a sort of dispersion catalyst that allows them to work faster at low temp (eg, iPoly)
     1d) Model M keycaps can take boiling temperatures just fine.  I've boiled keycaps from '86 through brand-new from Unicomp.  I've not warped one yet (not even a spacebar) but I also don't let them touch the bottom of the pan.
     1e) Predissolve dyes into their own concentrate, then decant into the container used for dyeing.  Some of the dye components take *forever* to dissolve, and you don't want them settling on your keycaps.

2) Yes, RIT dye does work, at least, some colors do.
     2a) Black [powder dye] comes out a rich brown.  About 5 minutes produces a tan color, about an hour gives a deep chocolate brown (at 200F).
     2b) Gray is utterly useless.  After a full hour, it will turn a white keycap the slightest beige.  I suspect all the light colors have the same problem.
     2c) RIT blues seem to barely work.
     2d) Of all the RIT colors, reds and yellows seem to work best.

3) iPoly (a Jacquard brand) dyes work faster than RIT but aren't perfect either.
     3a) Black iPoly produces a hunter green color in 3-5min.  It also seems to leave a nasty residue on the top of the water.
     3b) iPoly blues and greens seem to work best-- useful!  RIT and iPoly 'good colors' can complement.
     3c) Royal Blue iPoly and black RIT can be combined to produce a good, deep black. 
     3d) Currently experimenting with a mix of Green and Blue iPoly and Black RIT to produce the slightly greenish gray more typical of the model M.

4) Old keycaps will not dye evenly, period.
     4a) results can still be acceptable, but they simply will not look like new keycaps that happen to be made in that color.
     4b) With good cleaning prep, the tops can be dyed evenly, but the skirt will always end up darker.... changes in the plastic due to exposure to sunlight?
     4c) New keycaps from Unicomp *do* dye perfectly evenly (assuming they're not dirty)

5) All of the above will stain the Bejeezus into a Corian countertop

Cheers!
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: fohat.digs on Mon, 10 September 2012, 06:49:19
I concur with a lot of that.

Dissolving in one pot and straining out the particles is a good idea. Mix at much stronger concentrations than what the package recommends for fabric.

I have had spectacular results with RIT dye powder with colors like Sunshine Orange and Golden Yellow in 20 minutes at temperatures well under boiling (75C/170F - 85C/190F). Scarlet Red worked well enough but was a lot more finicky. 2 applications helped the red.

Dark Green was horrible and blotchy after quite a long time, but looks fantastic on my son's faux camo board.

Cocoa Brown was similar but worse.

I have tried 2 different blues that failed utterly, and Royal Blue, after over an hour boiling, left the keys with a faint sickly pink/gray/brown tint.

To disagree, I used black RIT dye powder and got a breath-taking pure rich shiny black in 20-30 minutes at well below boiling. Try a 2nd application?

All the keys I dyed were old, but I cleaned them very well.

Now I really want to try those other blues. I guess I will hit up ebay.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xJaPx on Mon, 10 September 2012, 09:19:05
i used dark green as well, it was the worst experience.  Within 5 minutes they were brown, like a dark khaki color.  I was very confused, as i would wipe to dye off the spoon and it was a beautiful dark green, but they keys were brown.  I eventually boiled them for over and hour so theyd just be black =/.  At not a single point within the hour were those caps EVER green :'(

I also used powder blue, which was relatively uneffecitve.  After 1/2 hour, there was almost no color change.  After an hour, it was a light lavender
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: fohat.digs on Mon, 10 September 2012, 10:53:10
Sorry for the pictures. Bad light in computer room and I did not feel like unplugging it, etc.

But the "Dark Green" did "work" depending on your definition of the word "work" .....

You get the idea.

Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Mon, 10 September 2012, 13:27:31
Dissolving in one pot and straining out the particles is a good idea. Mix at much stronger concentrations than what the package recommends for fabric.

Ah yes.  Forgot to mention that completely.  I was not being super-mehtodical with concentration (I will in the next round when I'm mixing), but this was all at levels of ~ 1 package of dye to 3 cups of water.  The iPoly appears to be much more concentrated than the RIT.

Cocoa Brown was similar but worse.
Really, huh.  I'd have expected that one to work well given black coming out as rich brown.  "I guess not".

To disagree, I used black RIT dye powder and got a breath-taking pure rich shiny black in 20-30 minutes at well below boiling. Try a 2nd application?

I'm sure it would have gotten to a good deep 'red black' eventually.  Perhpas I'm misremembering how long it took to get to deep brown.  Need to take better notes :-)

Now I really want to try those other blues. I guess I will hit up ebay.

The blue iPoly was very effective and could take a key nearly to black. I have a nice navy-blue test key.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Sun, 16 September 2012, 04:41:05
I've read in several places "the temperature is the only thing that really matters-- keep it at or near boiling."  I wanted to do a controlled experiment on how much the temperature of the dye matters... This is only two colors, but they're from different manufacturers and they agree reasonably well. 

Summary: The hotter the water, the faster the dye takes.  Almost nothing happens at all below 50C.  The pic doesn't quite show the strong difference between even 95C and 100C; in person it looks like the uptake almost doubled.  I wonder how many splotchy/uneven dye jobs are from the temperature variations in the dye bath and no agitation/stirring.

[edit: updated with better pic]
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: fohat.digs on Sun, 16 September 2012, 08:36:02
Amazing!

My cocoa brown attempt was an ugly failure, and yours looks great!

And I considered black to be one of my most stunning successes, and you were unhappy with yours.

Thank you for this additional research!
 
I have wanted some sort of brown-on-brown keyboard look because all my desks are wooden, and, while "wooden" keyboards are cool to look at, I figure that they can't function well.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Sun, 16 September 2012, 13:43:53
Amazing!

My cocoa brown attempt was an ugly failure, and yours looks great!
I was only dyeing one at a time, keeping the temperature constant, and stirring. Also using carefully measured, double filtered dye with some extra detergent in lab-grade glassware. So... ideal conditions :-)  God knows how much it will suck when I try to do a whole bunch at once.

And I considered black to be one of my most stunning successes, and you were unhappy with yours.
Not really unhappy, it was my first attempt and I was surprised at how far off of a neutral black the color really was.  I'd been hoping I'd be able to dilute or shorten the dying time to fudge a gray color, that's the part I was disappointed would be a no-go. All the in-between shades were a rich brown, and it was hard to get them even to boot.

I haven't run a test on this specifically, but it looks like fast/high-concentration dye-jobs are prone to splotchiness and spots that don't take, especially with the Rit.  The iDye appears to be more forgiving.

Right now I'm trying out low concentration, high-temp (holding things right at boiling), longer jobs.  I just did a batch of green test keycaps with the iDye at a concentration of 1 packet : 64 L of water (yes, that dilute! I could see through the dyebath) and they came out looking like they'd been shot that way.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: rayuki on Sun, 16 September 2012, 15:02:14
I've read in several places "the temperature is the only thing that really matters-- keep it at or near boiling."  I wanted to do a controlled experiment on how much the temperature of the dye matters... This is only two colors, but they're from different manufacturers and they agree reasonably well. 

Summary: The hotter the water, the faster the dye takes.  Almost nothing happens at all below 50C.  The pic doesn't quite show the strong difference between even 95C and 100C; in person it looks like the uptake almost doubled.  I wonder how many splotchy/uneven dye jobs are from the temperature variations in the dye bath and no agitation/stirring.

wow, this actually gives me faith to give dyeing ago when my new model m arrives.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: WhiteFireDragon on Sun, 16 September 2012, 15:48:42
I've tried to dye my caps many times, all were failures. I wanted a dark grey color, so I thought I would buy RIT black dye, but it turned my caps to some ugly brown after about an hour. It looked blotchy and had uneven scatch-like marks all over. I was pretty disappointed in this, so I left it in there for another 3 hours to get black. And the black I got was also a little off. It didn't look like the black from stock ABS caps, but instead had a different shade. The PBT caps I used were KBC caps. Not sure why they took hours to dye and not minutes.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: SmallFry on Sun, 16 September 2012, 15:51:47
Mine turned out well....Click on the Model M link in my signature. Not sure what I did differently than you guys.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: Alessandro on Sun, 16 September 2012, 16:42:49
Mine turned out well....Click on the Model M link in my signature. Not sure what I did differently than you guys.

Awesome! I love all the different shades of red on the esc row!
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: SmallFry on Sun, 16 September 2012, 17:09:07
The different colors were originally beige and a darker almost grey color, hence why the modifiers and the different colors in the ESC row appear.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Mon, 17 September 2012, 02:57:43
Some more dye experiments from the past day...

Pic 1: a quick color chart of the dyes I got that worked well.  I made most of these before I realized that small differences in dye bath temperature make such a big difference in how fast and how hard the dye takes.  Most of these were made by taking the dye to a boil, taking it off the heat, and dropping in the keycaps, but I wasn't being particularly careful to keep the temps and timing strict.  Still... they're not *too* far off.

I also tried out the iDye Poly Black; I don't have any keycaps from that run because I did it last weekend, it went badly, and I didn't save the results.  The iDye colors really need to be filtered before use, else the results end up swirled and speckled due to a nasty, greasy residue, I think left over from the self-dissolving dye pouch.  In any case, once filtered, these dyes work great.  The black came out a sort of hunter green before darkening to black with longer soaking time.

Other colors that don't appear either sucked (Rit Pearl Gray-- did nothing) or I didn't have it to try.

Pic 2: experimentation with dye concentration using the iDye Poly green.  I made several dye runs, each time cutting the dye concentration to 1/4 or 1/2 the previous run (mixture concentration listed to the left).  These tests account for evaporation; I measured volumes and replaced lost water each run, but used the same dye.  I also did this one before realizing temperature was so important; notice the 1:4 L line is actually darker than the higher concentration 1:1 line above it.  Noticing that is what spurred me to run the temperature experiment (and then start being much more careful about temperature).

The conventional wisdom I've read is 'make your dye mixture as concentrated as possible'.  I've begun to think this is wrong.  Notice that the last row on this pic is equivalent to one packet of dye in 64 liters of water.  The dye bath at that point was a translucent middling green.  There was no problem with dye uptake, and the light green made by soaking in a lower concentration for longer is more even than the green made by soaking in a high concentration for a few seconds.

My best guess is that the dyeing is a simple exponential diffusion.  Eventually the concentration of dye in the plastic will reach equilibrium with the dye in the water.  Riding the very steep part of that diffusion curve will give a less even dye job than letting things run longer at a slower pace--- this last part is a guess.  I've not run a controlled experiment of that lovely little theory, so it could be completely bogus.

Pic 3: same experiment as above but using the Rit Black.  The Rit dye is either much less concentrated than the iDye, or it doesn't take up as well (or both).  The Rit Black is also prone to unnevenness before getting to a solid black, and intermediate stages are beige/brown/dark brown.

One thing I didn't expect with the black: as I kept cutting the concetration, the color appeared to get warmer (more reddish).  Perhaps this is just an illusion caused by differential uptake of the different dye colors, eg, the red takes longer to diffuse such that the longer a cap soaks, the redder it gets and you can't tell in the earlier runs because the resulting colors from longer soaks are too dark to see it.  Or, maybe the dye hue really was shifting after being at near boiling for several hours.  Dunno, maybe I should run another test... ;-)

Pic 4:  a newer shot of the temperature test keycaps from earlier-- I wanted to retake it paying more attention to exposure and white balance, and so it would match the rest of the pics
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: rayuki on Mon, 17 September 2012, 03:10:53
man if i could get all the keys to look as good as you got that rit black at 20mins, with a black case.... would look godly.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: xiphmont on Mon, 17 September 2012, 03:15:07
I've tried to dye my caps many times, all were failures. I wanted a dark grey color, so I thought I would buy RIT black dye, but it turned my caps to some ugly brown after about an hour. It looked blotchy and had uneven scatch-like marks all over. I was pretty disappointed in this, so I left it in there for another 3 hours to get black. And the black I got was also a little off. It didn't look like the black from stock ABS caps, but instead had a different shade. The PBT caps I used were KBC caps. Not sure why they took hours to dye and not minutes.

I've noticed light colored scratch marks and nicks on many keycaps I've tried to dye.  They seem to be worse with the Rit colors. Several colors just don't take evenly.

My guess: Any stress or damage to the keycap that causes the smallest crack or separation in the plastic, even just microscopic stress cracks like crazing, will block the diffusion of the dye.  It has to move through the plastic, and can't jump the divide.

Many of the old Model M caps I've dyed have marks and scuffs that stay lighter colored.  All of the new Unicomp keys I've tried have dyed ~ blemish-free.
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: hluo87 on Sun, 16 December 2012, 23:38:53
nice work
i spent the whole Saturday to dye me caps.....
it does not work with rit yellow.... :-Xß
Title: Re: New notes on dyeing Model M (PBT) keycaps
Post by: dorkvader on Mon, 17 December 2012, 00:26:48
nice work
i spent the whole Saturday to dye me caps.....
it does not work with rit yellow.... :-Xß
What happened? I would love to see the pictures.