OK, So I just whipped up a quick diagram in Solidworks for you folks who are interested in the belt tensioner. It works very well and if you need more information on how to make it (assuming you have the means to) I can shoot you .stl, .dxf. files or anything in between to help you out. The whole setup requires:
5 Printed parts
2 M3x30mm Sockethead Capscrews (SHCS)
2 M3 nuts
2 M3 Washers
12 M5x10 SHCS (For two of the M5 SHCS, I recommend getting low profile heads from Openrail, it's not required but it makes for a cleaner build)
12 M5 T-slot nuts
1 10-24 Eye bolt
1 10-24 Nylok nut
14 M5 Washers (these will also work with the Eyebolt)
2 2020 HFS Misumi Extrusions
(Let me know if I missed something)
In any case, as the drawing shows, when you tighten the nut on the eye bolt, it pulls on the belt. Very little force is need since there is a slight pulley system in effect here. The belt feeds through the motor and pulley system as it normally would and it is clamped at either end using 2mm (I think) Zipties(cable ties)
[Make sure the teeth of the belt are locked together! ]
If there's anything I missed with the drawing or whatever. Let me know and I'll update the information as soon as I can. Now onto the frame...
JuiceBox, I like your frame very, very much. Do you think it's stronger than the original machine? Any difficulties getting it aligned? Can we have a detail image of your belt anchors?
First, getting the frame built took a bit longer than I had expected but I had ordered Misumi's kit which came in at the wrong dimensions which set me back a couple days. The frame costs about 100-120 bucks with all bolts, nuts and brackets included. I don't necessarily recommend their kit but it might save you a couple bucks, and, if you can follow their confusing diagram you should be fine. Just remember they measure from the outside of the frame, not the volume of the frame which is what I was thinking originally.
Rigidity: If I did my math right, this frame should offer a substantial gain in rigidity (strength) over the basic steel plates. The main reason is that the forces are spread out across a number of members instead of just the two. Additionally, there are thicker and more rigid corner joints with this method verses just two M5 SHCSs.
Alignment: When I built the frame, I built it on a Black & Decker portable table. What I didn't realize, was that it was crooked and I made everything 'flush' to the table. I then moved it back to my work bench to finish the build and I found that the frame was teetering on two corners (kinda like a bar stool does when two legs are longer than the other two). I started to check and measure everything and after about 10 minutes of doing that I got fed up and hit each corner with a mallet. I got 3 thunks and on the 4th I got a loud thud. I shook the frame again and it didn't move at all. So, in affect, one vertical extrusions wasn't aligned properly and whacking it with a mallet, fixed it. (I did some other minor adjusting as well but nothing worth noting). As for tramming and aligning the y-axis perpendicular to the x-axis I have not done but I know it needs it. Fist, since I am using the stock Shapeoko spindle clamps, the DeWalt 660 doesn't sit 100% vertical in the mount since the body has a taper to it. So I'm going to need to shim the upper bracket just a hair to fix this. Since I use this machine solely for x-y hole plotting, this isn't much of a concern to me but I'll fix it never the less. As for the y vs x axis alignment. I still need to drill 2 holes on both y-axis motor plates so I can fully secure the 4040 extrusion. Once this is done, I believe that the y-axis will be perpendicular to the x-axis.
Wiki:
I hope you'll document that and your belt anchors on the wiki.
If you explain to me how, I have no problem doing so.
Edit: Just noticed a typo in the drawing, the 18 inches should be 24 inches. 
Thanks Again! 
