Originally I was using a direct belt drive system shown here.  While it worked ok, the belts themselves present a few problems.

As the belts stretch, the ratio to the pinion gear changes slightly.  This will affect accuracy.

The belts cause the gantry and X-carriage to wiggle under high G forces.  The get good cuts, you need to be able to run your plasma table at at least .3Gs.

For these reasons I decided to change over to a Rack & Pinion Drive system.  While a R&P system will cost more than a belt system, the rigidity that you gain will be worth the extra cost.

Here are a couple rack and pinion sets.  The smaller one is 1/4" and I originally played with this set by mounting the pinion directly on the motor and attaching the racks to the side of the CNC.  This worked but due to the small racks, it would flex too much under any kind of tension.

I then moved to the 1/2" rack and pinion set shown on the right. Because of the size of the gear, I lost a quite a bit of resolution.

I decided to design a motor reduction assembly that would increase the accuracy, and torque and lower the top speed.  

Shown here is a prototype I used to help work out some of the dimensions for the various gears and belts that I have available.

I chose a belt drive system to keep the backlash to a minimum. Due to the belts small size, I didn't have the issues I had with my original belt drive system.

The gears I have on hand are designed for a 1/4" shaft.  The pinion gear unfortunately has a 3/8" shaft.

Using my mini-lathe, I was able to fabricate a couple of the shafts shown here.

Shown here is the first working reduction drive prototype.  Its made from 1/2" thick Corian.

The reduction drive mounts to the same 2" steel angle located under the Y-carriage that I used for the direct drive.  The drive is hinged to this angle with a 1/4" bolt that is tightened just enough, but still allows the spring to do its job.

The drives work pretty good.  I was able to get backlash down to about .001, with the speeds at over 1500IPM. 

I have some bearings on the way that will allow me to support the drive shaft.  This should help me improve the backlash even more as I think most of it is coming from the raw hole in the Corian.

I have started working on the second prototype. It will have bearings supporting the driveshaft.  The mount will be a little longer to allow me to move the pivot point further under the Y-carriage. I'm loosing about 2" of travel due to the current design.