[GB] The Smallfry Keyboard Kit (Closed for Orders, 97% shipped. SEE LAST POSTS)

[tahuds]

ah, so i'm going to have something more concrete to say about the coloring soon. the basic idea though is that my interactions post-quote from the anodizer have not been good, so i've pivoted on the coloring. this is going to be a bit experimental, but instead of anodizing, i'm going with something that i think is actually a bit cooler and may be one of the first times it's actually been tried. yes, because anodizing titanium to a fixed color scheme was not bleeding edge enough, i've opted for a process that is even more bleeding edge.

no, seriously. that's what i'm doing.

i'm going to nail this down with some experiments with the shop and then i'll be able to say more about it.

So long as the boards are shipped by mid December I'm cool with this.

[Vibex]

Spill the beans~!!

Uh uh, no spillers.

[mkawa]

also, the gasket folks proved empirically that they actually can't urethane coat this part. it's going to have to be PTFE. I'm going to see if there's anything i can do to prevent the kind of wear and defect that vibex experienced with his sample based on a couple of ptfe coated samples (which look great)

[PandaSPUR]

also, the gasket folks proved empirically that they actually can't urethane coat this part. it's going to have to be PTFE. I'm going to see if there's anything i can do to prevent the kind of wear and defect that vibex experienced with his sample based on a couple of ptfe coated samples (which look great)

Aw damn. Well, given the DIY nature of this kit, I'm sure individuals will probably find their own solutions if necessary

I'm already thinking about ordering a vinyl wrap to surround the gasket.

[mkawa]

now that's an interesting idea..

[Firebolt1914]

Sorry to bug you mkawa, but do we have a rough ETA of when this group buy will be done?

[mkawa]

blerg erg. i wish i could give a hard date, but every step in this production process has been an exercise in bug-fixing at a process level with multiple rounds of attempts at solutions. in retrospect i almost certainly should not have added the coloring step for this first try. unfortunately, i also think that my latest try at the gasket formulation is not working out well. i'll check in with the producer next week. in retrospect i got a tad too ambitious. once the results of this LAST round of experiments comes in, i will announce the final production details.

in detail:

- soft gasket squeezes out more and is more prone to being damaged, so i asked for a harder formulation with a ptfe coating. if this feels like it won't wear well to me, i'll have vinyl gasket protectors cut. i suspect this is less dire than it has been made out to be though, the original sample that vibex found the bug in was pretty far from the gaskets we're going to get.

- the plates are going to be 'painted' instead of being anodized. the paint is a ceramic coating commonly used on guns. we should be able to match the anodize colors with more translucency. i'm going to give the painter some latitude on this one, as i trust her completely to deliver.

- fastener and pcb lead-time is short. i can decrease gasket lead time by paying for more molds, especially after we nail the formulation. thankfully, i made slightly fewer molds than they originally wanted me to make just in case, so this won't be a deal killer. the material is finally cut (long story short, spaceX had a launch deadline..), so working out coloring is next.

- i may have boilermaker sort a few hundred thousand diodes while we're waiting for all of this to happen.

huge apologies to everyone for how long this has taken. when i started this, i had no idea what was going to happen in my life (3 moves in 3 months, the last across the country!) and had to drop off the face of the earth for a couple of those months to keep my sanity.

in retrospect the ultimate size of this ended up being the issue. i expected to move maybe 20 units, and it ended up being nearly 80. it's a lot easier to do experimental production of 20 units of product than 80 . just ask spacex (we share a pretty crucial shop for me :O)

[byker]

So the plates are cut right? Are we moving onto colouring them now? It seems like we could get everything else done, and just wait for the gaskets?

[mkawa]

it's a big fat dance, because nothing can ship until the longest pole is worked out, and every pole takes time and attention and has the possibility of becoming the longest pole. this is why i abandoned the anodizer. no personal contact, no trust and no timely responses from them. that was a pole that was short and suddenly became very long. same with the gaskets. i've been through _three_ sample batches now. i have to direct time and attention to the longest pole at any one time, but while i'm doing that, another blocker will pop up.. it's been a lot like whack a mole.

the upside is that this coloring option only became an option very recently. the downside is that it's experimental squared. but i'd rather work with a short-cycle experimenter that is a proven partner than a shop that won't give me the time of day.

there's also the fact that i currently own one fork, one spoon and two bowls. i literally had to take a day of work to set up my desktop in a place that wasn't next to my mattress (which doesn't have anything under it, in case you're wondering). i wouldn't live a life that wasn't exciting, but juggling lots of knives means dropping and stabbing yourself sometimes.

[mkawa]

also, there are still 8 units of product left that i'm holding the bag on. can't blame people for being cautious at this point, but juggling while holding a bag is difficult

[HoffmanMyster]

there's also the fact that i currently own one fork, one spoon and two bowls.

I found the non-metaphorical statement!  Do I win a prize? 

[byker]

Well IIRC the gaskets weren't part of the original group buy, but were brought up afterwards as a cool addition. We could always just finish the gb without the gaskets if it is too much to handle, then maybe look at dealing with them afterwards as a potential additional gb?

[mkawa]

i will fulfill anything i'm currently promising on the product page (with modifications if i have to, but you get the idea). that said, some kits will go out sooner than others. once any kit becomes ready to go, it'll go out.

[mkawa]

thanks to dork vader, rainbow should have the samples next  week. we're going to have to run some experiments to do costing and determine colors. once everything is set, we order paint, get the WJ plates sent and that will at least be in motion again. the gaskets _are_ in motion with this last sample. mechanically, something about these seems wrong though..

[mkawa]

then again, there are a lot of things wrong in my life right now, so it may be worth just pushing ahead on that one..

[Skuloth]

I would imagine most of us will be happy as long as the plates look good, and the pcbs work xD

[byker]

I would imagine most of us will be happy as long as the plates look good, and the pcbs work xD

Hell yeah! I guess the gasket is cool, but I like the looks of the jd40 without the gasket anyways. 

[r3nfr0]

Plates and pcb are what matter to me. Gasket not that big of a deal IMO

[tbc]

i was actually really looking forward to the gasket

[geniekid]

I'm very curious about this ceramic coating and it's thickness since I imagine the plates were not cut with paint in mind.

Also it sounds like you could use a drink, mkawa )

[mkawa]

100-150microns. coating thickness is negligible. it's an epoxy with ceramic powder in it. there are multiple formulations, but in the version we'll most likely use, the ceramic contributes hardness but not structure. there are also formulations where the resin cures after the ceramic is able to form hardening structures. this kind of coating is stronger and tolerates higher temperatures, but comes in fewer colors, because the amount of ceramic per unit volume is much larger, and the epoxy has to be pretty opaque. you usually lay this stuff on a lot thicker as well, since it's often used as an insulator. at 100mics, it's a slick and protective surface. at a few mils it's a ceramic layer that happens to have a resin substrate for adhesion and application promotion.

if we're forced to do 40 duro ptfe coated gaskets, i'm going to give people the option not to get one with the basic warning that with wear the polyurethane inner can adhere to objects and potentially act as an abrasive. we'll see if they can run 50 duro ptfe in the same mold. if they can, it'll have a bit more shrinkage and a bit less squish.

[dashing_girlboy]

If there is anyone who is looking to get a kit, I'm letting go of mine. With extra PCB, teensy. PM me if interested.

[Spopepro]

100-150microns. coating thickness is negligible....

To put this in perspective, this coating method is popular for firearms.  There are many very small interference fit pins in most firearms, and no provision is needed.  You drill the same hole whether you are using the coating or not.  It is hard to imagine how thin it really is.  It also does not wear on slides.  Impressive stuff and overboard high-performance for a keyboard.

[mkawa]

to be pedantic (ahahaha when am i not), very precisely machined firearms are machined to tolerances of about 50 mics, particularly for things like small pins that determine trigger tension, and so on. the trick is that the process is two parts, and it's not electrostatic.

first, the branded coating requires a significantly more precise blasting chamber than is typically used in parts cleaning or powdercoating. the blasting chamber needs to be able to produce controlled reductions in the 100-300 mics range Ra (a statistical measure of surface roughness).

next, the coating process is _wet_. acrylic powdercoating is accomplished by creating a voltage differential between a powder and the object to be coated such that the two attract. you then melt the acrylic while under charge so that it coats the surface. the acrylic particles are fairly large and the process is really only self-leveling if you assume a perfect sphere and even then, only to an integral number of particles. (in reaction, many precise powdercoating shops have started using a 'fluidizing' process. the basic idea is that bulk particulates can start behaving like bulk fluids if you toss enough kinetic energy at them. but, instead of using electrical charge to stick the fluidized powder to the object, you usually get the object really hot and then spray or dip it.)

the ceramic coating is a wet thermoset polymer that starts out as a liquid resin with ceramic structures and/or particles in suspension. the polymer liquid can have much smaller polymer structures than a thermoplastic powdered solid. further, the ceramic portion is extremely hard even when layered at a couple molecules thick. finally, actual liquids are actually self-leveling. hence, this particular ceramic composite can be used to consistently coat many more types of surfaces with a highly controllable layer thickness.

THE END... OR IS IT?!

[Melvang]

to be pedantic (ahahaha when am i not), very precisely machined firearms are machined to tolerances of about 50 mics, particularly for things like small pins that determine trigger tension, and so on. the trick is that the process is two parts, and it's not electrostatic.

first, the branded coating requires a significantly more precise blasting chamber than is typically used in parts cleaning or powdercoating. the blasting chamber needs to be able to produce controlled reductions in the 100-300 mics range Ra (a statistical measure of surface roughness).

next, the coating process is _wet_. acrylic powdercoating is accomplished by creating a voltage differential between a powder and the object to be coated such that the two attract. you then melt the acrylic while under charge so that it coats the surface. the acrylic particles are fairly large and the process is really only self-leveling if you assume a perfect sphere and even then, only to an integral number of particles. (in reaction, many precise powdercoating shops have started using a 'fluidizing' process. the basic idea is that bulk particulates can start behaving like bulk fluids if you toss enough kinetic energy at them. but, instead of using electrical charge to stick the fluidized powder to the object, you usually get the object really hot and then spray or dip it.)

the ceramic coating is a wet thermoset polymer that starts out as a liquid resin with ceramic structures and/or particles in suspension. the polymer liquid can have much smaller polymer structures than a thermoplastic powdered solid. further, the ceramic portion is extremely hard even when layered at a couple molecules thick. finally, actual liquids are actually self-leveling. hence, this particular ceramic composite can be used to consistently coat many more types of surfaces with a highly controllable layer thickness.

THE END... OR IS IT?!

Very cool info there mkawa.  However, my big question is this.  Is it a true no bake cure process or is it just "low temp bake"?

[mkawa]

checking out some datasheets, it looks like the basic formulation is a room temperature initial cure with a 300f final bake. it's single part, so there's no catalyst. it looks like there is some kind of room temperature reaction, i'm guessing that this is for adhesion or a leveling skin or something and then a low temperature heated reaction that cures the remainder?

also interesting is that a single pass is listed as typically producing a 5 to 10 micron coating

that is bat**** insane right there.

also super interesting is the singular sentence "electrostatic application may be an option". i don't know if this is with certain formulations or with certain parts or at high potential difference or what, but that could help lower Ra so that you're getting like a 10 +- 1 micron coating. that's crazy

[Melvang]

checking out some datasheets, it looks like the basic formulation is a room temperature initial cure with a 300f final bake. it's single part, so there's no catalyst. it looks like there is some kind of room temperature reaction, i'm guessing that this is for adhesion or a leveling skin or something and then a low temperature heated reaction that cures the remainder?

also interesting is that a single pass is listed as typically producing a 5 to 10 micron coating

that is bat**** insane right there.

also super interesting is the singular sentence "electrostatic application may be an option". i don't know if this is with certain formulations or with certain parts or at high potential difference or what, but that could help lower Ra so that you're getting like a 10 +- 1 micron coating. that's crazy

Very intersting but where I was heading is would this be possible to do to caps.

[mkawa]

there are room temperature set formulations. check out the C-series firearm coatings.

so yes, definitely possible.

[Melvang]

there are room temperature set formulations. check out the C-series firearm coatings.

so yes, definitely possible.

I wonder how well that would work as stem lubrication and/or how it would feel as a coating on the top of caps.

[Firebolt1914]

Just wondering, if I chose raw titanium, it won't come coated at all correct?

[mkawa]

correct!

[tahuds]

Just a heads up to those in the thread, at Pimp my keyboard the Deep Space group buy has a JD40 kit for $30 that would look really snazzy with a purple, gray, or raw kit. http://www.pimpmykeyboard.com/deals/deep-space-jd40/

[sypl]

How close are these kits to completion? Do you think they'll be out for Christmas?

[Firebolt1914]

correct!

Slightly off topic, mkawa, do you know if I would be able to laser engrave the keyboard, even with the cerakote? I want to know if it is possible as I've just gotten access to a laser printer.

[mkawa]

laser engraving of titanium is going to be problematic, as it will cause rainbowing from oxidization at the focal point. having to burn through the cerakote means turning the power up, which will only make this problem worse.

unfortunately, getting into holiday territory means that things move even slower :\. shipments take longer, shops go on vacation, etc. i've decided that i'm going to start mailing plates alone and pcbs if i have them (coordinating with ctrlalt on this one) to domestic customers as soon as soon as i have them. so, packages with only raw plates will go out first. then, colored plates, as rainbow turns things around quickly. finally, i'll resolve this mess with the gasket and as soon as the pcbs come in, i'll send those out with teensies to domestic customers.

domestic customers won't be charged for the additional shipping. unfortunately, the cost of shipping even tiny little parts internationally makes it prohibitive to send most things out piecemeal. if you have a special circumstance, eg you _only_ need plates and you're international, please let me know asap so i can mark your order and get it out as soon as parts come in.

which brings us to the question: will this kit ever be produced again? the answer is that it will never happen at this scale, ever. you would think i would have learned my lesson from those panavises but i'm very good at repeating my mistakes XD

[cjhard]

Mistake? That feeling you have deep in the pit of your guts, was that not fun? Wholesome, sadistic fun

[r3nfr0]

Running a group buy is a huge stress!  Thanks for sticking with it

[Ceedog]

So, how exactly is color going to be decided, seeing as how anodizing produces unique colors, and we have no idea what colors of Cerakote are available?

[mkawa]

anodizing titanium actually tends to produce rainbows. i chose the anodizer based on their claimed ability to process 6 uniform colors without rainbowing. communications after the quote however did not inspire much confidence.

in fact, cerakote, being a kind of paint, has an insane number of available colors:

http://www.cerakoteguncoatings.com/finishes/

that said, what rainbow and i will be doing is choosing 6 finishes that are the best looking approximation of the 6 colors the anodizer used. as much fun as it would be to let people pick any color they wanted, as you can see from the link, cerakote is really really expensive. so we're going to keep it down to six. by far the largest contingent is gray, which will help keep things manageable, but i'll be working with rainbow to see what they have stock of that we can use and what we will have to buy.

[HoffmanMyster]

anodizing titanium actually tends to produce rainbows. i chose the anodizer based on their claimed ability to process 6 uniform colors without rainbowing. communications after the quote however did not inspire much confidence.

in fact, cerakote, being a kind of paint, has an insane number of available colors:

http://www.cerakoteguncoatings.com/finishes/

that said, what rainbow and i will be doing is choosing 6 finishes that are the best looking approximation of the 6 colors the anodizer used. as much fun as it would be to let people pick any color they wanted, as you can see from the link, cerakote is really really expensive. so we're going to keep it down to six. by far the largest contingent is gray, which will help keep things manageable, but i'll be working with rainbow to see what they have stock of that we can use and what we will have to buy.

I'm confused.  Will we still be getting the colours we selected before?  If not, what colours will we be getting?

[Skuloth]

anodizing titanium actually tends to produce rainbows. i chose the anodizer based on their claimed ability to process 6 uniform colors without rainbowing. communications after the quote however did not inspire much confidence.

in fact, cerakote, being a kind of paint, has an insane number of available colors:

http://www.cerakoteguncoatings.com/finishes/

that said, what rainbow and i will be doing is choosing 6 finishes that are the best looking approximation of the 6 colors the anodizer used. as much fun as it would be to let people pick any color they wanted, as you can see from the link, cerakote is really really expensive. so we're going to keep it down to six. by far the largest contingent is gray, which will help keep things manageable, but i'll be working with rainbow to see what they have stock of that we can use and what we will have to buy.

I'm confused.  Will we still be getting the colours we selected before?  If not, what colours will we be getting?

Sounds like we will be getting the closest color match to the one's we selected.

[mkawa]

yes, you'll be getting the closets match to the ICC color you requested. i can't match the actual anodizing, because despite repeated requests, the anodizer never sent me color samples. hence, cerakote.

[r3nfr0]

As long as I can get those GREEN plates to match the green tea keycap set I ordered!

[Skuloth]

laser engraving of titanium is going to be problematic, as it will cause rainbowing from oxidization at the focal point. having to burn through the cerakote means turning the power up, which will only make this problem worse.

unfortunately, getting into holiday territory means that things move even slower :\. shipments take longer, shops go on vacation, etc. i've decided that i'm going to start mailing plates alone and pcbs if i have them (coordinating with ctrlalt on this one) to domestic customers as soon as soon as i have them. so, packages with only raw plates will go out first. then, colored plates, as rainbow turns things around quickly. finally, i'll resolve this mess with the gasket and as soon as the pcbs come in, i'll send those out with teensies to domestic customers.

domestic customers won't be charged for the additional shipping. unfortunately, the cost of shipping even tiny little parts internationally makes it prohibitive to send most things out piecemeal. if you have a special circumstance, eg you _only_ need plates and you're international, please let me know asap so i can mark your order and get it out as soon as parts come in.

which brings us to the question: will this kit ever be produced again? the answer is that it will never happen at this scale, ever. you would think i would have learned my lesson from those panavises but i'm very good at repeating my mistakes XD

I am excited to hear this. I'm so excited to finally get started building my JD40.

[mkawa]

As long as I can get those GREEN plates to match the green tea keycap set I ordered!

you have no idea how excited i am to pick a seafoam semi-translucent cerakote color. OH GOD SO MANY BEAUTIFUL COLORS!!!!

[byker]

As long as I can get those GREEN plates to match the green tea keycap set I ordered!

you have no idea how excited i am to pick a seafoam semi-translucent cerakote color. OH GOD SO MANY BEAUTIFUL COLORS!!!!

Can you let us know when we should repick our colours. I need to take a look once the new ones are decided. 

[Firebolt1914]

As long as I can get those GREEN plates to match the green tea keycap set I ordered!

you have no idea how excited i am to pick a seafoam semi-translucent cerakote color. OH GOD SO MANY BEAUTIFUL COLORS!!!!

Can you let us know when we should repick our colours. I need to take a look once the new ones are decided. 

Same here, can we also choose colors from the website?

[mkawa]

unfortunately, no. in order to keep costs sane, i'm sticking with the six color scheme that danco used, just substituting cost effective cerakote colors.

that said, if you want to take your plates raw and send them separately to rainbow paint & blast to color, you will pay a heck of a lot extra, but you can get any cerakote color you want on your plates. suggestion? mini-gb the coloring. if you can find a few people willing to share the cost of the paint itself (a lot, unfortunately), the blasting cost and the cost of painting a separate small batch, then by all means go for it. i don't want to stifle people's creativity here. i just want to keep myself and rainbowrpb(.com) sane on the large batch i have to process.

ps, i send small batches for powdering to rainbow all the time and i find their pricing to be reasonable (and i run some weird stuff through them), the quality top notch and the turnaround to be as fast or slow as you need. tammi at rainbow is awesome, and is always happy to help geekhackers. we've both been incredibly excited about the cerakote process (sooo many exhaust parts). you can see some of her work on their facebook page: https://www.facebook.com/pages/Rainbow-Paint-Blasting-Inc/337716255868

[byker]

So should I go by the old colour scheme or will you post a new one with the 6 colours?

[mkawa]

go by the old one. we'll start with the straight ICC colors and do a nearest neighbor, then look at what's in stock already, and potential modifiers. assume you're getting your original chosen color. i'm still waiting for rainbow to get the sample plates, and much will be informed by what they think will work best with that material.