I am working on NAMD recently. I used AMBER force field files. NAMD did have much better parallel computing for big system.
The solution was got from http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/3969.html.
1) you need 1st create a new .pdb file with B col specified as ‘0’ for unaffected atoms, and ‘1’ for fixed or constrained atoms. you can do it with VMD.
after you load your protein and open the TK console, type the following for example,
set all [atomselect top all]
set fixatom [atomselect top “protein”]
$all set beta 0
$fixatom set beta 1
$all writepdb fixprotein.pdb
2) then you simply implement the fix or constrain patch into your .conf file
a short example,