I promised this post to someone, it’s much overdue, Karlie, you asked for it:
The first time I made the Ripley watch it was just for myself and my friend to go to MCM with. I then redesigned it slightly and made some improvements when I was asked to make one for a customer. Later I redesigned it slightly again when orders increased since the production process is labour intensive and I was getting errors creeping in that cost me time and material wastage. This design (v4) was great for a while, but I figured I could do better.
The Version Five.
This one no-one will ever see. It was an attempt to radically revise the buttons on the side (the main area that takes time) and the overall look of the face, and it didn’t work. In reality there was a v4.1, 4.2, 4.3, 4.4… etc. All failed to deliver on the full package. However from this version came a range of ‘revolutionary’ ideas! The Version Five was a new breed, I redesigned the whole watch from scratch rather than modifying the original V1 file as before. This one was the same size and shape but has some very subtle alterations, a tenth of a millimetre difference here and there can make a big difference in the way the printer approaches the job, this in turn can lead to a cleaner print, tiny adjustments in the way the parts are put together makes it more straight-forward to assemble, but it takes a lot of time and a lot of experimentation, and ultimately the watch looks pretty much the same!
The old way:
Previously the front face and the ‘edge/back’ of the watch were printed as two pieces, which I then had to clean up and glue together. The join had to be sanded back, filled, sanded back again. The whole face then had to be painted (even though they were black anyway) because it helps hide some of the banding from the print and gives a more uniform look. Then I had to carefully fill in the little red and blue bits on the fake front buttons with white, then when that was dry with the relevant colours – this is necessary to stop it looking dull.
Stripping the design back, and starting from scratch.
I started by redesigning the way the LCD panel fits in the watch back itself, this is another area that took some time during assembly to get right. In Version Five I went for a more form-fitting shape that more solidly held it in place from the start, which means I can push it in and know it’s in the right place, very little additional alignment necessary. This took a while to get right since here tiny print irregularities can make a big difference. The trick wasn’t so much measuring the LCD part and making the model to fit, it was doing that to the output – since what comes out of the printer isn’t exactly what the 3D model looks like, it’s a feedback loop of trial and error, making tiny alterations until the actual piece fits, each test takes 35 minutes to print, then alterations need to be made, etc. Took a while.
The more tricky and more major difference is that the front of the face and body are printed as one piece now, this was something I played with in V5 quite a lot. This saves all the time it used to take assembling the bits together. There were some technical issues to overcome in terms of actually making it printable – those of you who know about FDM 3D printing will know that the machine can’t print over a void – I use the analogy of cake icing, you can’t just pipe fondant icing over thin air and expect it to hang there, you need to pipe it over something solid and then you can build up layers – so the big hurdle was printing the front of the face over the void where the watch LCD panel and circuitry would be. There’s a thing called supports that can be used, but often they can be messy to remove, especially on something so small. However, I built into the model custom supports in just the right place to allow the printer to do it’s thing and allowing them to be easily snipped off without causing extra production issues.
The first V5 was an experiment in actually breaking the face down into more parts. Instead of printing the face including the recesses and raised areas for the fake front buttons and panels etc. I printed it completely flat so that I could sand it down to get a really nice surface. I then created a new file of just those rounded rectangular button plates and printed those, and then a file of the thin surround and printed those. This was in the hopes that I’d be able to carefully trim and sand each section, then glue them together. This would mean that I’d be able to make it look very smooth and neat, however this process took a lot longer than I expected and the final result looked bad, really bad. Ultimately I kept the concept of the face being printed as part of the backs, but instead of the part-assembly idea I started experimenting with vapour polishing the printed parts and making adjustments to make this possible.
Getting a smoother finish.
This has always been a big issue for me. I love what the 3D print process can allow, but since I’m trying to replicate an injection moulded part getting a good finish is important for the look of the thing. I’d previously tried vapour polishing the watch but on the older versions the very thin parts necessary for the print to work tended to have issues – vapour polishing is essentially using Acetone to melt away the top layer of the plastic, when it stabilises a lot of the banding has been removed or reduced. On larger pieces this can make them look almost injection moulded, but on something as small as the watch face it’s quite a bit more tricky. The other thing is it’s a fairly hazardous and time consuming process. Heated Acetone vapour is (as you’d expect) corrosive and toxic, it can also explode if the right stoichiometric ratio is reached (when the concentration is correct in the air, it can ignite given an ignition source).
I developed a carefully timed process, along with some modifications to the watch again to shore up thinner areas, to allow me to polish the printed watches in a safe way, and because of all the time I was saving in production with the optimisations mentioned through this post, the extra time this takes per-face isn’t an issue.
Picking the right Paint.
I also did a lot of testing to get just the right paint for the job. I previously airbrushed several very thin layers of black acrylic based paint, then sealed with a clear topcoat, which got the coverage and finish I wanted. Again this is a total of a couple of hours work allowing for drying time and since it required several coats it was a fairly labour intensive process, spray, time the drying, realign, spray, time, realign, spray, time, etc. Plus cleaning the airbrush which is a pain too.
For the V5 I spent quite a bit of time experimenting with different paints to try to get something that would give the right finish but was less time-intensive than the old method. I settled on using enamel paints, which are thicker and give a really nice finish and although the overall drying time is much, much longer, the finish is significantly better with half the amount of time actually spent on the painting – this ties back into creating more time to do the vapour polishing.
The unexpected side effect of this is that I also on a whim tried using red and blue enamel for the button marker infills, and it works great! the thicker more robust paint fills the holes (which I’d engineered to be larger on the V5 than the previous versions to make it easier to infill) more easily and uniformly and retains the strong colours without the need for the white backing infill. The earliest versions of the V5 actually have polymer-clay in these areas, the effect was interesting and certainly more consistent and quicker than the earlier painting method, but the finish wasn’t what I was looking for in the end.
Getting the Glue right.
The small rubbery bit that joins the two faces together is made from black EVA foam, which is tough and can take a lot of twisting and bending, so much so that I’ve not yet managed to break it except by actually twisting it round 360 degrees, and to be honest if that happens to your watch, I hate to think what’s happened to the wrist wearing it! So that wasn’t an issue.
The problem area for me has been attaching it to the faces. The faces in my files are named North and South. The North face has a slot built into the South edge, and South face has a slot built into the North edge, this slot is the right size for the EVA strip to slide in and be glued in place solidly. This process is exactly the same on the V5 as the previous versions, however the glue is the difference. I was previously using contact adhesive because it gives a great bond between two different surfaces and in testing it held up well. The main problem is that it’s a bit imprecise and needs to be timed to work effectively. I discovered a new glue via a friend who makes jewellery, she uses a glue called GS-Hypo cement. It can be used to glue almost anything to almost anything, but here’s the crucial bit, the nozzle is a tiny needle like tube that’s engineered not to clog between uses, the glue is slightly more viscose than super-glue, it doesn’t oxidise plastic like super glue can, and it doesn’t stick to skin. It’s a bit more forgiving than super glue and when it’s dry it’s super-tough. It’s awesome. I bought a tube (read: stole her glue) and this is now the glue that attaches the EVA joining bit to the North and South faces, and also the glue that attaches the face graphic to the white backing and to the watch face itself, and the orange surrounds on the base of the faces. Awesome glue.
Buttons, buttons, I really hate the buttons.
If all that sounds like a lot, it’s nothing in comparison to how long, and how many point versions I went through trying to improve the buttons. This all came about because I had someone contact me who had a faulty button on his watch. We were able to resolve the issue, but it was clear that the older system wasn’t reliably working, and to be honest 50% of the assembly time went into the buttons. It was a fiddly and frustrating part of the process, and the failure-during-testing rate was high.
The issue here is that the 3D printer can’t print small things very well. the nozzle is .4mm across, which sounds pretty small, but the old buttons were so tiny that the printer was struggling to accurately reproduce them. What this means is that the size of the buttons varies, and I’m talking by .1mm or something, which seems small, but when you’re fitting something that only measures 1.3mm into an opening that it needs to slide into but not pop out of, the tolerances are tiny. I went though a dozen or more iterations to redevelop the buttons, I altered everything possible in an effort to get them to print more consistently, and the main purpose behind the redesign in the first place to was to try to get a version that was more consistency reliable.
Ultimately, I went back to the drawing board and took a look at the original prototypes. These ones (if you look back through the blogs) had post-style buttons. These buttons are reliable, easy to make and easy to assemble. The only reason I moved away from them is that visually they aren’t as great as the inline flat style I used later on. However in this case it’s a case of function over form, and I reworked the design to streamline the buttons as new post-style buttons. The early tests proves that they were entirely reliable and the huge amount of time saved here allows for the vapour polishing process and an additional layer of QC on all the watches.
The orange rubber bits… were OK actually.
I did make a tiny change to this area, just thickened the bottom layers up by .2mm, this creates just a little more of a platform and makes it easier to glue together.
And we’re done. The Version Six is born.
From the ashes of the Version Five, the Six is here. A more robust, more consistent design. As of now, with the technology available to me in my workshop, there’s little else I can do to improve this design and keep the price point the same. Version Six is an entirely new watch, rebuilt essentially from the ground up to be more than just-something-I-made-for-cosplay-one-time-and-sell, it’s been designed to be a wearable hand-made watch that looks like the famous Ripley watches.