Spiffy, I had not heard of reprap before the 2007-01-25 LWN and started reading the blog.
by the way the old blog (e.g. 2005-03) entries have a bunch of broken links.
I found the comment about air-drive for producing the 3mm feed stock and thought I would try it
I was thinking however that it might be simpler than the MkII for deposition also, since it can be all plumbing parts.
http://mysite.verizon.net/james.mcmechan/extruder_sm.jpg (140K) is a picture of my test rig.
I used 1"(~25mm) black pipe fittings for the extruder with a 1/16" (~1.6mm) drill.
I was also lazy and just used a wood clamping jig and cardboard as the heat barrier.
I followed the early use of hotmelt glue sticks for plastic, my HDPE just did not get past stiff putty in the oven.
Nor did I do a feed valve like in high end glue guns.
I placed the big pipe in the oven for 15min @ 350F (176C) and then assembled it over the sink and turned up the air pressure slightly.
Hot glue stretches badly after a short distance but I still got about 9ft (3m) of fiber
The sections where I held it low enough were about 2.16mm (0.85") in diameter and quite consistent
When I popped the air hose loose the extrusion stopped and fell to the surface with a mostly square end
A three way air valve with the extruder on the common port should work nicely.
Replacing the PFTE tape with antiseize compound and a better heat barrier like a griddle brick,pumice stone,magnesia or charcoal soldering block or even a piece of gypsum wallboard, should allow for temperatures up to molten lead without problem.
I was also reading about the use of DC motors instead of steppers, it looks like the problem is trying to encode after the gear train,
where I would think before the gear train would be best.
example 20:1 worm gear small 12000rpm motor 12000/20 = 600rpm = 10rps, 10rps * 0.8 pitch => 8mm/s
which would have a resolution of 0.8/20 => 0.04mm per revolution of the motor and since the controller knows which way
it is being driven it could just add or subtract based on the driving direction requiring just a revolution counter in the PIC.
Great start James. I'm really glad to see someone with the interest going after that work again.
I've got so far into doing Tommelise that I'd put aside the whole filament production question once I found that I could buy filament from plastic welding rod people. Mind, that still doesn't address the problem of how you recycle plastic.
Sounds like we need to make a plan to get you some CAPA (polycapralactone) plastic to work with.
As for HDPE you have to leave it in the oven for a LONG time for the bubbles to work their way to the top of it. It wouldn't even get started in 15 minutes. Most plastics have very high viscosities, about like molten glass except maybe worse.