Making some progress. Adhering pretty much to Tom's design except for a couple of changes that might be noteworthy.
Tom used channel aluminum for his rails. I had some channel steel that I used. It is the stuff you put vertically on a wall and has slots in it that shelf brackets fit into. It was nice and straight but a bit rusty. I gave it a good working over on the wire brush of my bench grinder and got it where you could run your finger over it and it was smooth to the touch. But when I assembled the X-stage and installed the little bearing assembly, I was concerned over the amount of friction there was. It was a lot harder to push and pull than I expected. In the video, Tom's looked like it slid quite easily.
While I was mulling this development over, my eye strayed to a pile of junk on the side of my workbench in which sat a marble. Hmmmm... I put the marble on the slot of the rail and put another piece of channel steel on top of it. It slid practically frictionless. So I hunted down 3 more marbles and tried it again. The top rail still slid easily and there was no discernible latitudinal slop. "Hey... I might be on to something here!" said I. So, as you can see in the pictures, I put it together with glass marbles to act as bearings. It remains to be seen how well it will work in actual practice, but if it works, marbles can be had 100 for a buck at the local dollar store. The bearings Tom used cost a bit more and are harder to find. I put a screw in the end of each rail to keep the marbles from escaping their duty. (I hate it when I loose my marbles!)
(July 18 EDIT): Okay, it's a month later and though the marbles ~almost~ worked, they had a few small problems. While the stage was traversing, the line of marbles would hit the stop screw but the ones in back would keep on pushing. This would finally push a marble out of the rail and onto the floor. Also, I discovered that the marbles all looked the same size, but were not. Slight differences caused uneven routing depths. I've come up with something even simpler than the marbles but I won't present it until I know it works.
I used 1/4" x 20 threaded rod and found that "3/16" ID Fuel Hose (Lowes #22283) was perfect to join the rod to the 1/4" shaft of the motor.
In the video, Tom showed how to wrap metal around a nut to make a bracket for the threaded rod to attach to the underside of the stage. I did it a little differently and I think this was easier. I just took a standard size cable clamp, smeared a very thin layer of gorilla glue inside it and inserted two 1/4" x 20 nuts. I threaded the rod into them so their threads were in sync and let the glue cure. I'm sure this looks better than what I would have wound up with trying to copy Tom's metal gizmo.
(July 18 EDIT): Another good idea gone bad. Although securely bolted down, the clamp was still slightly flexible and gave me a small backlash. Also, as Gabriel pointed out, steel nuts are not good. The lithium grease has already turned black from metal particles in it. I really like the picture though.
So I need to finish up the Y stage. I hope it won't matter if the motor is mounted on the side rather than in the center of the stage. I can see it making a difference if I were routing wood or metal, but I guessing that just buzzing the copper off a pcb won't give any side torque.
I find myself guessing a lot here. I looked high and low for a cnc noob faq. All the faq's I found were forum rule faq's... not things like "is 1/4 x 20 threaded rod good to use for a pcb mill?" Guess that depends on a lot of variables though.
Haven't even started on the Z stage. I'm thinking about moving the post over to the corner and having it point out diagonally to the center of the work area.
Oh well, 'nuff fer now.