I’ve been working for about 6 months on finding a suitable silicone rubber material to use when making the keys for my beatseqr project. I started on this journey right after I’d set up my blacktoe cnc router. I knew I wanted a clear rubbery material to put in between the LED Tact switches on my circuit board for the user to push. This is a pretty well understood paradigm, so I figured there had to be a low-volume, relatively low-cost way to at least do some research into the feasibility of being able to do this. Of course you never know everything you need to know when you just start stabbing wildly at solving a problem that doesn’t have an obvious solution.  So, below is a small sampling of the attempts I’ve made.

TAP Platinum Silicone

When I started, I thought I knew what I needed. A clear silicone rubber. Right? Should be easy to find. I didn’t really know what exactly to look for, but then I saw that TAP had one! I drove on down to one of my local TAP stores and was a little sticker shocked at the price. Oh well, one credit card swipe later, I was ready to test it out. The product is really simple to use. It’s a 1:1 mix ratio, so you pour parts A and B in equal amounts, mix it all up, and start pouring it into your mold. It’s super runny before it cures so it pours easily.

Once I had tested it out, I realized that there was a critical third qualifying attribute to the silicone that would ultimately win the day: hardness. Smooth-on has an excellent FAQ posted about this topic. It explains everything you need to know and has a printable visual aid. Ok, so with that new knowledge, I can see that this TAP silicone has a Shore hardness of A8, which explains why it feels like floppy gelatin. I’m going for a push button with some give, without being squishy. So this silicone was not going to work. (It did, however, come back later as an excellent mold making material, its intended use anyway.)

Smooth-on Sorta clear 40

So while we were showing Beatseqr at the Bay Area Maker Faire 2010, I was able to locate the Smooth-on booth and discovered a couple of clear silicone rubbers that were MUCH firmer than the silicone I’d tried. Sorta clear 40, and Dragonskin 30. (Note that those product pages contain more than one shore hardness in the same product line!) I figured I’d go with the firmest one they sold, so I located a local dealer and bought a quart sized trial kit of Sorta clear 40.  As you can see from the photo above, the optical properties are excellent with this product.  However, after running a couple of test castings with this product, I discovered a fourth critical qualifying property that I’d need to consider: viscosity.

Sorta clear 40 is an excellent material. I really like it.  It does, however, trap air bubbles… permanently, if you don’t do something about it.  That something you can do is called “degassing”.  Degassing essentially means that you use a vacuum chamber to forcibly remove trapped gasses, usually air, but sometimes other gasses depending on what you’re working with.  Well, I didn’t have a degassing vacuum chamber, and the ones that I found when searching for them seemed to cost multiple hundreds of dollars on the low end, and thousands on the high end. Definitely not within my budget for this test.

I found an article explaining how to build one, so I did that.  I’ll save the explanation for another post, but I managed to cut the price down from $800 down to about US$250.  I’ve always been interested in different materials, and I knew I could use this degassing chamber for other stuff later, so I justified the cost and assembled the parts.  

The viscosity of sorta clear 40 is pretty high, meaning, it does not pour like water. More like honey. Or taffy. The way you combine the 10:1 mix ratio parts is to open the canister of the part A liquid and scoop out 10 parts into a cup on a scale, and then add drops of part B to the mix until you reach the 10:1 ratio. I say “scoop” because the part A liquid is barely a liquid. If you hold the container upside down, it does not pour out. It’s perfectly happy to stay there without pouring at all. When you mix parts A and B together, it does become a somewhat pourable liquid, but before you get there, you have to stir stir stir for 3 minutes to get it fully mixed. Then the material has a bazillion air bubbles in it. So off to the degassing chamber it goes for several minutes, all while the “pot life” is ticking away. Pot life, meaning that you have a certain number of minutes after you combine the parts before they start to cure and solidify, so you need to work pretty quickly to get everything done once you start the process. So, out of the degassing chamber, and into the casting mold, quick!

above, you can see the worst case scenario of rushing to get the silicone into the casting mold, and not being thorough about making sure all of the cavities are filled and big air bubbles are out. Not only does this produce an unusable finished product, it wastes time and money.

Above is a test I ran to mix, degas, pour, and then degas in the mold. It sorta worked. Small bubbles were reduced significantly, but big bubbles formed and by the time the silicone started setting up, they were trapped forever.

Above is another test I ran in a similar fashion. I got pretty good clarity by working faster and leaving more time for the silicone to sit in an uncured state in the mold, but while working feverishly to make sure all big bubbles were eliminated, I overworked the material past the pot life, and so the back of the flange is mottled and uneven. This might be usable, but seems like waaaay too much work to get it right every time.

Originally, the first beatseqrs had acrylic bar stock cut into rectangular shapes and hand sanded to be used as the button interface. The look and feel was pretty good, but the amount of work it took was too much for any kind of mass production, so I thought I’d try a casting acrylic to get the same/similar results without having to do so much hand work finishing. For all of the silicone castings, I am using machinable wax as the negative casting mold. Casting soft rubbery materials into a hard mold is fine, but… you can’t do that for rigid materials. So I made a “master positive” casting out of the TAP silicone, and then a silicone negative from that master positive. Actually, i did two. the first time, I didn’t adequately spray down the master positive with release, and silicone loves being cast on top of silicone, so I adequately destroyed the first silicone negative mold that I needed to do it again. The second silicone negative worked fine, and I was able to do the above casting in the casting acrylic I found at Michael’s.

Ok, BUT … when I went to do the acrylic casting, I discovered the downside to this material immediately. It basically smelled like a chemical factory took a crap, exploded, and died. And it stayed like that for upwards of 3 weeks. I asked a guy from Smooth-on about that and he said it sounded like it was a polyester casting material. Ok, so as pretty as it looks, I will definitely avoid using that again. The jury’s still out on rigid keys… Smooth-on has a material I may run some tests with, but maybe later if some customers dislike the silicone keys. So far, that hasn’t happened too much.

I had these silicone negative molds, and started looking around my garage. I noticed how nice and translucent hot melt glue sticks are. After I realized that silicone would more than withstand the heat, I did a test and got some interesting results. Maybe this is a good fast casting material in some cases. It’s relatively cheap and easy to work with. But… really… not what I was going for.

Enter the dragon. Smooth-on Dragonskin 30, to be precise. While at World Maker Faire 2010 in NYC, I got a great recommendation to try dragonskin 30. It seemed to have the right properties for this need:

1. translucent. Not transparent like sortaclear 40, but early tests indicate that LED light transmission is pretty good, which is what I want.
2. shore hardness A30, which is pretty firm. Not quite as much as sortaclear 40, but firm enough.
3. 1:1 mix ratio, so easy peasy mixing.
4. low viscosity, easy pouring. Less trapped air bubbles, should be easier to degas.
5. good working time and self leveling. Here, take another look:

The flange is super smooth. That lets it sit perfectly flat on the LED Tact buttons so they they’re all seated evenly. There are some air bubbles trapped in the casting, but that’s because I didn’t bother degassing the early tests. I’m confident that if I need to get more clarity in the final castings, degassing will go well because this silicone is so much less viscous than sortaclear 40.

Full castings in the mold cavities. Awesome. Minimal hand-working to get it to that state. Since the material pours much more easily, it doesn’t need much coercion to fill all the nooks. I declare a winner! Dragonskin 30.

There are a few loose ends to tie up on this story, so maybe when I complete all of the research I want to do on this process, I’ll post up the results. I’m hopeful my market for beatseqr is just starting, and I’ll need to keep refining my products to meet demand. If that happens, I will have more to post.

UPDATE … it occurred to me that you might be interested in see what the casting mold looked like (or, one similar to the current one I’m using) and how the casting mold looks when it’s filled, and how the casting looks as it’s coming out of the mold, so here you go!

UPDATE 2: just a reminder to never leave your CNC program running unattended. I stepped away for one minute right at the end of this run, and KABLAM… ruined casting mold. I was lucky it was just wax!

The end result is a mold to use for casting translucent silicone into for use in my beatseqr project, as a replacement for the hand-cut-and-sanded acrylic buttons that I’m currently using. You never know how an experiment will really go until it’s over, and now that this experiment is over, I’ve learned a lot and will make a bunch of adjustments:

1. change the silicone product from tap plastics platinum cure translucent (way too wobbly for this application) to smooth-on sorta clear 40. I saw this product at the Maker Faire this past weekend and really liked the firm quality of the cured product. I think it will be a really good candidate. Smooth-on dragon skin 30 was my second-best candidate.
2. cut a new button mold using the techniques I’ve learned by doing this one.
3. adjust the height of the button to be not as tall as this test.
4. adjust the flashing surrounding the buttons to be not nearly as tall. I had it set to be 0.1″ tall, and that’s probably 0.08″ too much. all I need is for it to basically be thick enough to keep all of the buttons together, and that’s it.
5. think about how to diffuse the LED light going into the button… this new silicone is water-clear… so … I’ll need a way to adjust that quality so it looks good. maybe just a sheet of paper under the silicone flashing… we’ll see how that goes.

So… progress continues. I’ll post the result of the next prototype when it’s done.

And then I’m into all kinds of new new new new stuff thanks to my training at Techshop. Here’s some acrylic test pieces. I’m over the moon on the cut quality on these. I have some fairly big ideas that I’m going to try out on this front.

Here are some larger hardboard abstracts that I’ll be sealing and painting. I’m so excited to be able to get this level of precision into my build quality. Once I paint them, I think that will be a nice combination of machine precision and human imprecision :

And then we’re working on a mobile for our coming addition to our family. Here’s the first pass at the cast of characters. I output these at 100% of the design spec, but we both agree that I can redo these and bump up the size. A lot. Although, I’m totally shocked at the fidelity to the design that the combination of the laser cutter working on the 1/4″ hardboard can reproduce. I probably just got lucky with where I put the holes and how much space I added around them.

check out the detail:

Note the circles at the white arrows. No problem at all. Every one of the holes I put into the pieces produced a clean 1/4″ thick hardboard peg. I love this machine. It’s freakin’ awesome.

Technorati Tags: art, epilog, fine art, techshop

I would highly recommend taking some classes there. The instructors are nice, and the environment is really conducive to exchanging ideas and helping each other learn. I’ll be starting up my month long membership next week! I’m very excited to get in there and get some stuff done!

Technorati Tags: cnc, cutter, epilog, laser, milling, plasma, TIG, techshop

So, some history… Last year I got the crazy idea to build a projector from scratch, using an LCD screen and a 400 watt bulb. This isn’t really anything new or innovative, I used the tried and proven designs from Lumen Lab to achieve the results I had originally planned for. After the projector was built, I needed a screen to project onto. I’d already spent waaaay too much money on the project as a whole, so I needed a screen solution that would be:

• cheap cheap cheap to build
• easy to locate materials, or reuse stuff I already had
• free standing and structurally stable
• knock-down construction so I could assemble and disassemble it as my needs required.
• big when assembled
• compact when disassembled
• reusable for other purposes (more on this later)

So, with these goals in mind, here’s the solution I came up with:

materials and approximate costs:

• four “2×2″ studs, these are about 8 feet long, and cost about $2 each.
• four shelf brackets for the feet, 2 on each side. This cost can very widely, but maybe call it $4 each. This is even debatable.. you could just just whatever 90 degree brackets you have lying around. Or just run 2 or 3 screws into the bottom on the side legs into the feet and skip the brackets altogether. I ended up using two shelf brackets and two 90 degree steel brackets.
• length of PVC pipe. It’s cheap, easy to work with, and light.. sort of. – $8ish
• a little bit of hardboard for the u shaped holding things for the PVC – $free – $5 for a quarter sheet
• Various woodscrews, 8-32 machine screws, washers, and wingnuts, maybe $5 total in hardware
• a canvas tarp. I got mine from home depot along with the rest of the materials, and it has a nasty sewn seam across the middle of it. Not ideal, but again, I’d spent a bunch on everything else, so I decided I could live with a seam in my canvas screen. $10-ish
• finally, I used acrylic gesso on the tarp for two reasons. One is to fill the small gaps in the woven fabric, and two was to make the natural tan canvas closer to a white color. You could go completely bonkers on finding the right combination of primer, base coat, and reflective additives to cover your screen.. I just went with gesso. I had a bucket of it sitting right in my garage already. $optional

These construction details are more intended to be a guide than strict instructions. In fact the only major things I remember about the design decisions I made were that I’d have it be 6 feet tall, which would leave me a 2 foot long foot from an 8 foot tall 2″x2″ board. Also, I decided that I’d make the crossbar widths as big as I could to get the biggest 16:9 ration screen out of the 2x2x8 boards. So, if I have a 4 foot tall side from the top to the cross bar, that puts my width at like 7 feet 2 inches or something like that. Oh well. You get the idea. 16:9 good.

I had an extra length of PVC with a little room to spare, so I came up with the u shaped brackets to hold onto the PVC pipe off the back side of the top. It works really well, but you could probably achieve the same results with a couple of cloth straps looped and screwed into the top of the side pieces… in fact, I may end up changing it to that kind of system so that I don’t have to worry about the special shapes I cut out of hardboard breaking in transit to or from a gig. I just grabbed the edge of the tarp and screwed it into the PVC pipe with some short self tapping screws. Doesn’t have to be insanely secure, just enough to make sure it doesn’t rip apart and fall off the frame.

Lastly, I decided that I’d so with half-lap construction and use 8-32 bolts, washers, and wingnuts so that I could assemble and disassemble the frame easily and quickly. It’s all very cheap hardware, again, stuff I already had lying around. There’s very little on this design that’s “hardcoded” together, so it does break down to a small, if long, space.

Here are some more photos:

This is a detail photo of the foot construction:

Half lap joinery, at the crossbar:

8-32 bolts, big washers, and wingnuts on one of the top corners:

Here’s the canvas rolled up on the PVC and sitting in the u-shaped cradle brackets I made for this purpose:

Here it is, torn down and strapped together with some cloth straps I sewed together and used with some quick-disconnect clips. It’s leaning up against my garage door, so you may identify some of the door’s anatomical things in the photo…

and finally, here’s a rendered image of what it looks approximately like from the front when you have the screen extended to near the floor in “4:3-ish mode”… 16:9 mode can be achieved by only extending the canvas down to the crossbar.

And of course since I’m showing you 3d renders from Sketchup, you can grab the models from the 3d warehouse:
Exploded construction view and assembled view.

So far I’ve used the screen a lot at home in my garage, and taken it to a few gigs, and it’s done very well. Hopefully the design will inspire you to make a version to suit your needs. Let me know if you found this useful.

Technorati Tags: diy, LCD, make, maker faire 2007, makerfaire, makerfaire2007, projector, screen, sketchup, stevecooley

Since I can’t get my act together to build this any time soon, and since I did this design a year and a half ago, I figured I might as well put it out there since I’ve seen two new teleprompter ideas lately.

Technorati Tags: teleprompter

ok, here’s the link to the 3d warehouse sketch up file … man, that is cool what they’ve done with google and sketch up. Proper respect.

Oh, go check out my main blog Some junk we like.