Sanity check for homebrew CNC milling machine

I'm about to buy some of the hardware needed to build a bridge type
CNC machine. I'm going to use it for routing and engraving
(wood/plastics) and also want to be able to use it for some light
aluminium work. Before I buy though I wanted to check my ideas with
someone who's got a clue if/how well it will work - that's you
people! :) Perhaps you can give your ideas on what kind of
performance to expect from a machine like this.

For those of you who aren't used to the metric system (numbers are
rounded):
Force: 1 Newton [N] = 0.2248 [lbf]
Torque: 1 Newtonmeter [Nm] = 141.612 [oz-in] (Nm is written just like
that, and not N-m or N*m)
Distance: 1 millimeter [mm] = 0.001 meter [m] = 1/25.4 [in] = 0.039
[in]

This is what I had in mind:

* The frame of the machine will be constructed out of 19 mm MDF
(medium density fibreboard). I will build 150 mm deep and 150 mm wide
U profiles out of the MDF to get maximum rigidity for the bridge, and
use lots of extra pieces for supports to get a machine as heavy and
rigid as I can. The work area of the machine will be 600 mm x 420 mm.
(The reasons for building with MDF are - it's cheap and I can machine
it. I don't have access to a machine shop.)

* Linear guides: for the main axis I already have linear ballbearings
and shafts. These are 1100 mm long and 20 mm dia solid steel bars and
linear ball bearings. Unfortunately they are not the "supported"
kind, but I will support them at the middle as well as at the ends.
For the other axes I haven't made up my mind yet - I thought I would
evaluate the guides I've got first before buying anything more.

* For leadscrews I'm going for 20 mm leadscrews with "Anti-Backlash
Supernuts" (BS&A XCT20x4M). The lead is 4 mm/rev and they are
designed for a load of 1100 N (250 lbs).
Now, to get 1100 N linear force, I need about 1.7 Nm of torque at
the leadscrew. To get this I used:

M = F*p/(2*pi*Eff), where

M = torque [Nm],
F = linear force [N],
p = lead of screw [m/rev],
pi = 3.14159..., and
Eff = efficiency of lead screw, in this case 0.41.

* Motors: I found a source (Brigar Electronics) for surplus PM
synchronous servo motors, 0.33 Nm at 2.6A continous / 0.99 Nm peak at
7.8A. I'm going to couple these to the lead screws with a toothed
belt, with a reduction of 2.5:1, which will give me 0.8 Nm continous
and almost 2.5 Nm peak torque at the lead screw. The maximum RPM of
these motors are 7500 RPM, which is way more than I need (even 2500
RPM gives a linear speed of 4000 mm/min or close to 160 in/min!). I
am however concerned about the available continous torque - will it
be enough or do I risk overheating the motors?

(And yes, I know these motors will be a lot more difficult to drive
than normal DC servo motors - but I will manage somehow because I
think they're cool :)

I'm sorry for this long letter but all these things (together with
end machining of the lead screws and angular contact ball bearings)
is a considerable investment for me as a hobbyist. I hope to get a
homepage on this machine going so that others might get some
inspiration and help also though.

Any thoughts on this - hints, corrections of mistakes, things I
haven't thought about, even "yes, I think it will work!" - are more
than welcome!

Best regards,
Arvid Brodin