how do you mean tolerance, distance from the pvc to the pipe? its a little less that 1/16 between the pvc and the pipe i think.

i think i should also check to see if after a while do the o-rings on the top fall down. would they require somethign to keep them in place or will the pressure hold them up?

I had actually planned on having some sort of attachment between the swivel and the back frame to prevent the swivel from trying to turn, but also to keep it centered on the drill string.  If there is only a 1/16th gap there now, it is likely that after a bit of running that the gap will expand from wear.  I would do a bit of testing like this and then take some sand paper or something and then just open it up so the gap is a bit larger and do the same set of tests for a range of gap sizes.  Ideally at each gap size you would want to test with 1) water only, 2) with water and coarse sand like for a sandbox, and 3) with a fairly thick mixture of water/clay.

The third test is the one that interests me the most.  If we can use a very large gap between the swivel and the drill string but just restrict ourself to always using thick drilling fluid, that would be ok.  Many sites will hit a clay deposit at least some distance along the drilling anyway and since the amount of clay needed to make thick mud is relatively small it can be brought from a clay pit in the area.  You could reuse it as well, however this may be a bad idea due to the potential for well contamination.  Always using thick mud could also simplify the design of the mud pump as we can optimize it for that consistency.