How to set the right PME parameters for amber's prmtop and inpcrd files?

From: a a (
Date: Wed Jan 26 2011 - 04:34:33 CST

Dear Sir/Madam,
I have prepared the prmtop and inpcrd with xleap in AMBER10.
In xleap, start with a water-free pdb file, I have prepared xxx_vac.prmtop and xxx_vac.inpcrd.
After adding Na+ and water molecules, with the "addion" and "TIP3BOX 8.0" commands. I have prepared xxx_wat.prmtop and xxx_wat.inpcrd.
I have two questions here, as I want to do the analysis with NAMD
(1) For the parmfile and ambercoor input lines of NAMD, should I use xxx_vac files or xxx_wat files obtained above?
Is NAMD will not read the Na and water molecules in these files, even if I used xxx_wat.prmtop and xxx_wat.inpcrd files?
Can I keep using the boundary/ water molecules's geometries obtained in AMBER for NAMD calculations?
(2) If new boundary parameters need to be set, how can set it up? i started with the xxx_vac files (the molecules is about 70x70x70 A in size, before adding Na and water in AMBER10). I have send the PME section as follows: I read the cellorigin from the centroid of xxx.pdb file in pdb viewer. So, I set boundary with 100A, with a aim to cover all the atoms in the pdb file.
PME on
PMETolerance 1e-06
PMEGridSpacing 1
cellBasisVector1 100.0000000 0.0000000 0.0000000
cellBasisVector2 0.0000000 100.000000 0.0000000
cellBasisVector3 0.0000000 0.0000000 100.0000000
cellOrigin 22.045 0.7493 30.7129
PMEGridSpacing 1.5
wrapAll on

However, the following errors appear.
ERROR: Constraint failure in RATTLE algorithm for atom 9999!
ERROR: Constraint failure; simulation has become unstable.

I tried to increase the cellBasisVector from 100. to 500. The same error message appears. while atom 9999 is one of the surface atom in the molecule.
I double checked the geometries of the xxx_vac.prmtop and xxx_vac.inpcrd files, they are alright, as when I run xxx_wat.prmtop and xxx_wat.inpcrd files in AMBER10. No error message indicating bad geometries appears. So, i suspect I have done something wrong in the NAMD input file or boundary setting, could you mind to give me some hints what may goes wrong?
Best regards,

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:34 CST