[gmx-users] Patch for wrong NH2 angles in calch.c

[Charles Karney]

The bond angles used by calch.c when adding "type 3" hydrogens (i.e.,
planar hydrogens in -NH2 groups) is wrong.  The following patch fixes
this and also uses the exact value for the tetrahedral angle.  This
patch applies to the gromacs 3.2.1 source

-- 
Charles Karney                  Email:  ckarney at sarnoff.com
Sarnoff Corporation             URL:    http://charles.karney.info
201 Washington Rd               Phone:  +1 609 734 2312
Princeton, NJ 08543-5300        Fax:    +1 609 734 2323

--- gromacs-3.2.1/src/gmxlib/calch.c.orig	2004-01-23 12:23:13.000000000 -0500
+++ gromacs-3.2.1/src/gmxlib/calch.c	2004-11-18 09:43:28.000000000 -0500
@@ -103,7 +103,8 @@
 
 void calc_h_pos(int nht, rvec xa[], rvec xh[])
 {
-#define alfaH   (DEG2RAD*109.5)
+#define alfaH   (acos(-1/3.0))	/* 109.47 degrees */
+#define alfaHpl (2*M_PI/3)	/* 120 degrees */
 #define distH   0.1
 
 #define alfaCOM (DEG2RAD*117)
@@ -181,8 +182,8 @@
     break;
   case 3: /* two planar hydrogens, e.g. -NH2 */
     for(d=0; (d<DIM); d++) {
-      xH1[d] = xAI[d]-distH*sin(alfaH)*sb[d]-distH*cos(alfaH)*sij[d];
-      xH2[d] = xAI[d]+distH*sin(alfaH)*sb[d]-distH*cos(alfaH)*sij[d];
+      xH1[d] = xAI[d]-distH*sin(alfaHpl)*sb[d]-distH*cos(alfaHpl)*sij[d];
+      xH2[d] = xAI[d]+distH*sin(alfaHpl)*sb[d]-distH*cos(alfaHpl)*sij[d];
     }
     break;
   case 4: /* two or three tetrahedral hydrogens, e.g. -CH3 */