I had a bash on both SPCE TIP 3 n 4 n 5.

The first 3 ive got to work fine in charmm and reproduce literature data....TIP5P does funny stuff.
All sims were 1296 waters in a 33.9 A box (correct physical density). For the implimentation of TIP3,4 I did not use the literature values for the forcefield. In the literature only O has LJ params. The FF i used for TIP3, 4 had LJ for the H as well (mckerrel standard params). TIP5, irrespective of if H has LJ params has v. short HO hbonds (down to about 1.3A), and as a result large voids occur in the water box (clearly a V. unhappy simulation!). All the water mol. geometries are rigid and correct. PSF readout from charmm says all the correct charges are on all the correct atoms. Im outta ideas. Ive been over the LJ params for TIP5 (there is only one!) and it is correct and v. similar to those used for TIP3. I suspect this may be a prob. with the Lonepair command in charmm.

Phil Mason, Brady group, Cornell
as a matter of spam evasion i never post my email anywhere public. I do check the spam account.. for an immediate response check the cornell web directory.