> I am thinking about making a Mendel, but i want
> one that can print things larger than 8"x8"x5.5"
Just keep in mind that build times can exponentially increase as the volume of your printed area expands. Also, leveling the build surface such that the extreme ends still bond well to the bed is more difficult with a larger build plate due to how the offset will grow as you travel outwards along an angle. Warpage will also become a bigger issue and making a heated bed will require more power. Having a larger build plate has its advantages, but there are serious drawbacks as well. I suggest giving serious thought to whether or not you can split things into smaller printable segments or use other part orientations to print big things before automatically building a bigger printer. For what it would cost, you might end up disappointed with the performance.
> How should i go about making a larger version of
> the Mendel?
A design more like the shapercube might be better suited to larger scaling. On this printer, the extuder moves in the X and Y direction and the part moves in the Z direction. The extruder may end up being lighter than your build plate and part. Choosing to move the lightest load along the more complex axis should simplify the design, allow the use of cheaper components, and make it easier to build. Squaring up the different axes may be easier on a box-shaped construction too.
>Also i was wondering if there is a way
> to modify the original design so i can make
> accurate color changes?
I think this is still under development. I saw some work on piezoelectric print heads that could allow printing colors much like normal printers do, but I have not seen a working model actually put ink on a part. Some people run multiple hot ends on one extruder, which allows them to print with multiple materials. We can't mix two colors to achieve a third (with any extent of control anyhow) like on a inkjet printer, so even with two materials we are still stuck with relatively monotone coloring.