The basic principle governing reinforcement is this:

To first order, an E/G beam supported at the ends without reinforcement has to be at least twice as heavy and likely twice as deep as a reinforced one because all of the material below the neutral axis contributes virtually no strength since E/G has a negligible tensile strength. See my ascii art a few posts back for the meaning of neutral axis. If the doubled mass(weight) is an asset in your application rather than a liability then reinforcement may not be necessary.

Harlow's commerical machine at http://www.cncbridges.com has rebar however:
http://www.cnczone.com/forums/showpo...9&postcount=51

This is not to say all machines need rebar but Harlow is selling a lightweight mini-mill and it's certainly the way to make it easily in the size it is.

See also
http://www.cnczone.com/forums/showpo...59&postcount=9 for a bit more info.

Generally, the heavier a gantry, the harder it will be for the machine to change direction. The lighter the gantry, the more it will vibrate. There is likely a happy medium somewhere. Likewise, there are some designs that will deflect more than others depending on how the materials are used.

As for the thermal issues, the E/G has a higher coefficient of expansion than reinforcing steel. In an extremely thin member, or one that is not symmetrically reinforced on the top and bottom, the reinforcement will tend to bow the member concave down. This effect will constrain how thin composite members can be made.

Anybody have a good sense of the temperatures differences we might see across a baseplate or a gantry beam?

Here's a food for thought: If you use a metallic tensile reinforcement and can preferentially heat the reinforcement material by some method like induction heating by a reasonably large amount just seconds before the epoxy sets, you could prestress the member and get much greater strength with much less reinforcement and less worry about the exact Granite content. You might have to heat the material from it's current temperature by something like 100 degrees C after everything is done and right before the epoxy sets. If you use a reinforcement that exceeds the thermal expansion of E/G then this might be a bit easier (steel does not.) I don't remember the numbers but aluminum might work.
This may not by possible in the HSM environment but I may work more on the idea after I get out some info on beam bending. You could also induction heat metallic reinforcement to help cure recalcitrant epoxy.