Many years ago I had a dream of building a machine unlike anything that was around at the time (and as far as I can determine still doesn't exist in quite the way I envisioned it). Every so often I'd get fired up and attempt to make it, but basically I lacked the skills and the equipment to build a complex machine from scratch.
Then people started building their own CNC mills, and 3D printers came on the market. I was weighing up my options - laser cutter, 3D printer, milling machine - when the ShapeOko came on Kickstarter. I'd already come across Michal Zalewski's guerrilla guide to CNC manufacturing, so I jumped at it and got serial number 877 - it's currently running with a few mods under LinuxCNC with a serial stepper driver, and I've just ordered the parts to add a second motor to what most people call the Y-axis (but which is set up as the X-axis on my machine). Yeah, I've wired my Z-axis in reverse to the standard configuration, too - it's what works for the way I use it.
At first I thought I'd be able to make the parts I wanted out of wood, aluminium, or acrylic. I started with wood, but the results weren't really suitable, I didn't have an adequate method of holding the workpiece in place, and I made my first tool blunt in a couple of weeks. I'd already discovered the limitations of having only a single motor per axis so I didn't even try aluminium. And acrylic was a no-go from the supply point of view. I could only find suppliers that would sell huge sheets (we're talking several square meters), and only up to 5mm thick. They would get something thicker in on special order, but the price was ridiculous. So then I looked at lost-wax casting. I made several prototype parts in wax, but there is only one foundry in the country - it's on the other side of the island (4 hour drive each way) - and my contact with them via email and phone wasn't looking very promising for what I needed to make.
Here in Fiji there are no hobby shops and basically nothing to cater to the maker market except for general hardware stores (with way less variety of stock than developed countries) and engineering suppliers (who are outfitted primarily for commercial and industrial transport and construction). That means any materials like Renshape, cool resins, silicon molding materials, etc. need to be purchased from overseas, attracting costs for shipping, as well as charges for customs & duty, and VAT on arrival. I started experimenting with silicone caulk and Oogoo, along with car body filler, plaster, and wood filler.
I was able to find exactly two resins in the country - a 1:4 epoxy resin (around $100 per litre) and fibre-glass resin (at about $20 per litre). But I needed to invent a way to mold it reliably. and that's when it hit me - instead of making a wax part, then making a crappy mold around it, and then pouring resin into that mold, why not carve the mold shape directly into the wax and pour the resin into it? So, that's what I now do. I buy candles from the supermarket and melt them down into blocks of wax using aluminium baking trays in my oven. Then I draw up the parts I want and hand-write the gcode that will carve the inside of the shape from the bottom up to make a mold. Then mix up the fibre-glass resin and if all goes well I get my part.
Resin curing is an exothermic reaction, and it melts the wax that is used to mold it. That means each mold can only be used once. But cutting wax is much faster than cutting heavier materials - I regularly rough out at feeds of 2000 mm/min and large details can be done accurately enough at 500 mm/min. My parts are a far cry from what Michal Zalewski makes. Sometimes there are flaws in the wax blocks (remember they are just Diwali candles, not specially formulated engineering wax) and large areas of wax tend to create bubbles in the contact surface of the resin, but it's a small price to pay for being able to make things using only materials I can source locally at a reasonable cost.