I machined mine from M6 stainless steel threaded rod (60mm long works well so there is continuous tube from nozzle to hobbled gear).  You could also use regular steel which is easier to get from a local hardware store and easier to machine, but has a higher conductivity (You want as low conductivity as possible so that the heat stays near the nozzle and doesn't creep up the barrel.  This keeps the melt zone narrow and also prevents heat damage to plastic parts it is mounted to).  I also bought 4 copper washers and nuts to put at the top of the barrel to help with heat dissipation.

Here's what I did:
1) Bore a 2mm hole straight through the threaded rod
2) Mark where the rod will emerge from the heater block and grind a narrow ring there.  The best way to do this is to use a dremel with cutting wheel in one hand and spin the threaded rod in a drill with the other hand or use a drill press (photo attached).  While you are spinning the bit, it is helpful to smooth and shape the ends of the rod as well.  I ground a ring down to 3.5 mm diameter.  The reduced cross section will slow the heat travel up the barrel.  I also used dremel sanding discs to smooth the ends to a shine.
3) Screw the nozzle on tight, then the heater block.
4) Starting right above the ground spot, alternate nut/copper washer/nut/copper washer…etc. until about 15mm from the top.
5) My hardware store had M6 washers with the exact diameter of the original PEEK barrel mount.  Using two washers, two nuts, and a split washer, I was able to clamp the above extruder assembly onto the acrylic jigsaw very tightly (see photo).  This design makes it very easy to remove the assembly as well, you just loosen the nut with needle nose pliers and it slides right out.
6) Finally, I mounted a fan to actively cool the top of the barrel with the copper washers.  This may be unnecessary, but with it running the copper washers are cool enough to touch, so I know my melt zone is quite narrow.

A couple weeks ago one of my 8 hour prints ran out of filament about 5 hours in and I discovered it two hours later.  So my Nozzle was just moving over air for 2 hours with the nozzle at 205 degrees.  I was sure I would have a clog, but when I added filament, it just pushed right out.

One quick note about machining a center bore.  If you have access to a lathe you will be in good shape.  Otherwise, it is possible to do on a high quality drill press with a bed clamp using the following method: 1) insert drill bit into chuck upside down (cutting side up), 2) lower bit into a bed clamp and clamp tight.  Also clamp the bed clamp down to the bed, 3) loosen chuck and raise up off of the bit (you now have a bit clamped to the bed which is perfectly in line with the drill axis), 4) insert threaded rod into chuck, 5) with drill press on, lower threaded rod onto clamped bit SLOWLY clearing shavings often.  You can get a pretty straight hole this way.

Why is the I.D. of the threaded tube so much larger (3.0mm)than the filament diameter (1.75)?  Seems to me that keeping this oversized reservoir of molten plastic would be a bad thing.  If the I.D. were closer to 1.8mm wouldn't there be a more precise push/extraction action going on?

Yes, other hotends have different designs for this specific reason.

Why is the I.D. of the threaded tube so much larger (3.0mm)than the filament diameter (1.75)?  Seems to me that keeping this oversized reservoir of molten plastic would be a bad thing.  If the I.D. were closer to 1.8mm wouldn't there be a more precise push/extraction action going on?