Thursday, February 8, 2007


Ha - bet you don't get THIS in New Zealand! (Actually, I think that they probably do - in fact, it'll be cleaner, drier, and possibly ski-able. But I bet that they don't make really tall snowmen).

Sorry about the hopeless pic - new camera, left it on ISO1600 by mistake. (GRR - either use MANUAL mode and THINK about what you're doing, or leave it on AUTO. Mucking about in P mode without checking stuff like this leads to hopeless snaps. Ah well.)

Back garden before we started:

Saturday, February 3, 2007

Dice Tower Construction.

First of all I started with an outline plan. I wanted something that was at least bigger and sturdier than the first dice towers I had made. I liked the arrangement of rods placed horizontally across the tower instead of the usual fixed slopes, but the original design only worked well with one size of dice and would "trap" larger dice within.

Those early efforts were 3mm thick acrylic - on the limit of what can be scored and snapped by hand. Going up to 4mm thick meant that everything was going to have to be cut with a jigsaw - no worse than score & snap, but still a pain.

While looking at materials on the supplier's website, I noticed that they did a laser cutting service. Not cheap by any means, but it would eliminate the need to lots of tedious sawing. And the laser cutter could pre-form the holes for me, so no drilling either. Of course, to keep the price as low as possible, they needed a CAD file (to send them hand-drawn plans would have pushed the cost sky-high).

So, the next task was to learn how to use some CAD software. While honing the design - both were worked on in parallel. There were wrong steps along the way with both. For example, I wanted a wider tower, but the rod placement got complicated. Too wide, and you could only fit the outline of one tower on a sheet of plastic - I needed to get two out of each cut sheet.

Finally I had a design that looked fine. It would clearly handle a range of dice sizes and it had a fold-out tray at the bottom. There was no chance to "prototype" it - either the design was right, would work, and I'd be happy, or I would be spending almost £200 on plastic fit for the landfill. The designs were sent off, credit card handed over, and fingers firmly crossed.

Disaster In The Making

The courier arrived with a big package a few days later. In it were 4 sheets of immaculately-cut 4mm acrylic, and enough 4mm diameter acrylic rod to make up 8 towers. Time to assemble the first one.

I had used a glue called Tensol-something-or-other for the earlier towers. Superglue is right out - it leaves a nasty "misting" effect around the join. Tensol is a thick solvent that "welds" two pieces of plastic together, taking about ten minutes to set, and over the course of about 24 hours creates an insanely strong bond. How strong? Well, I tried to snap the ramp out of the tatty test-piece I'd done some time before, and the plastic broke down the middle - and not along the weld joint as I had expected.

The drawbacks are that it has to be applied by brush, and then the pieces held together precisely while the joint set. This was OK for the joins of my 3mm towers - none of the edges were truly square anyway - but it wasn't going to be good enough for towers that I wanted to sell. The other problem is that it was very difficult brushing the glue onto exactly the right places - any excess etches the plastic surface. The benefit of using Tensol is that the joint is nice and clear, though.

The solution lay in using the same solvent in a much thinner consistency. The solvent is sold in the UK under the name "Fusion", and has the same viscosity as water. So if you have two closely-matched components, as my laser-cut pieces were, you can just introduce the solvent to the edge of the join and capillary action causes it to wick into the whole joint. As a result, the tower sides can be clamped together and squared up before the glue is introduced, so the finished towers are less likely to wobble! It's also easier to use the right amount of the solvent (though spills have to be wiped up in a split-second, or they will mark the surface of the plastic. Downside is that the joint isn't quite as clear as with Tensol - but it's still very acceptable.

The first tower of the batch was assembled by hand, with Tensol, while I learned what order things had to be assembled in. I wouldn't perfect the process until tower 7 or 8, but that first tower taught me a few things - ditch the Tensol, and use the "clamp & wick" principle; clean all the parts of fingerprints before you glue them together, because it's well nigh impossible afterwards; and also that I'd made a mistake on the plans.

This was a big set-back. Not a total disaster - the first assembled tower would indeed function with various different dice, and the "ramp" made from a row of rods worked well. But I had got the size of the tray wrong - it was 8mm too narrow, and couldn't be assembled as designed. I tried a bodged solution, but it looked like a bodge. There was no alternative - another £80-odd to get a sheet of replacement parts cut.

Unfortunately the Christmas holidays intervened. Instead of having a few leisurely days to assemble the towers during the break, the request for more plastic languished in someone's in-tray until their return to work after the New Year. Ah well. It was all sorted by the end of the first week of January, the replacement parts had arrived, I'd tracked down a supplier for Fusion (and the necessary syringes to apply it - despite the relative ease that addicts seem to have when it comes to sourcing their syringes, a hobbyist faces a much more protracted search, aided by an overly-aggressive spam filter on my email account).

The most tedious part of the assembly process is cutting the 4mm dowels to the right length. It took over two feature films to get them all done (using a simple "measure, score and snap" process). The New Year television schedule wasn't a total waste of time, then.

The only really tricky part is clamping the pieces together prior to introducing the solvent. I tried a couple of different clamps before hitting on a style, and method, that worked. The tray needs careful handling to get the sides on square (and the tiny G-clamps I had were only just big enough to do the job); the tower itself was easier, but had a nasty habit of collapsing suddenly as I made fine adjustments or tightened the clamps up too much.

Once everything is clamped up, actually injecting the solvent is pretty easy. Almost entertaining, even, watching it wick along the length of the join. This is where having the pieces laser-cut would really score heavily - it made straight, square joins much easier to achieve, and the finished edges were straight without being sharp (on the 3mm thick acrylic, some of the edges were very sharp).

15 minutes later, release the clamps. Glue the base onto the tower, slide the rods into place, and use a bit of solvent on both sides of each rod to fix them permanently in place. The folding tray hinges on the bottom rod of the ramp, and was the only really tricky join - it's important to get enough glue in to fix the rod to the tray, but a tiny bit too much and it will start to wick along the side of the tower, and weld the hinge so it won't work. Thankfully this didn't happen to any of the 8 (one got a tiny bit of solvent in somewhere it shouldn't, and the hinge squeaks a bit as a result, but this will ease with some use).

There we are. Job done. Each tower probably took about 2 hours to prepare and assemble, although having got the technique sorted now I reckon I could probably do a tower in about 45 minutes.

By making 8 of them, I had made 7 "good" ones and one that might be described as a prototype. Although it probably had a value (it's mostly square, reasonably tidy, and works fine) I destroyed what worth it had when I attempted to learn some polishing skills to remove some noticeable blemishes. As a result of the disaster that unfolded, leaving deep scuff marks on the back of the tower, I elected not to try to polish out any of the much less noticeable scratches and marks on the "good" towers. Hey, they're plastic. Although acrylic is pretty tough and hard-wearing, they will inevitably pick up a few marks along the way. Mine just come with a few built-in.

Actually, I think I'm over-stating the problem. You'll not notice any blemishes on the finished towers apart from on the test piece. I am, by nature, something of a perfectionist, though. For the same reason, I'm still not completely happy with the joins. Using Tensol did provide an almost invisible bond between pieces; the Fusion bond is good, but visible. I suspect that the solution may lie in controlling the atmosphere in which the bond is made, using applicators which are perfectly clean and unaffected by the solvent itself - a solution that lies way outside my means.

If one day someone wants to order several hundred of my dice towers, the answer will be to out-source production to somewhere that they perfect the assembly. And flame-polish the finished item. Does the world need several hundred deluxe acrylic dice towers, though?