# Understanding the gaff2.dat proper dihedral parameters

I have a project where I'm trying to program my own force field, and I decided to use the AMBER gaff2.dat force field parameters.

However, I am really struggling parsing the gaff2.dat file when it is about the dihedral angle parameters.

I am now trying to replicate the rotation around -O-C-C-O, for which I know the profile :

$$V_{std} = 0.25 [1+ \cos 3\omega] + 0.15[1+cos 2\omega]$$

I read that for gaff you have to take every parameters into account, so for $$-O-CH_2-CH_2-O-$$ I have to use the parameters for O-C-C-O, H-C-C-O and H-C-C-H. In the gaff2.dat file I get :

os-c3-c3-os   1    0.000         0.000          -3          p28,suger5ring,suger6ring,coccoc GA AUE=1.1750 RMSE=1.6708 TorType=2
os-c3-c3-os   1    0.000       180.000          -2
os-c3-c3-os   1    0.170       180.000           1

hc-c3-c3-os   1    0.000         0.0            -3.         Junmei et al, 1999 TorType=1
hc-c3-c3-os   1    0.250         0.0             1.         Junmei et al, 1999 TorType=1

hc-c3-c3-hc   1    0.120         0.000           3          m1 SS AUE=0.2420 RMSE=0.2944 TorType=2


Which I guess would mean that :

$$V_{gaff} = 0.170[1+\cos (\omega -\pi)] + 0.25[1+ \cos \omega] + 0.12[1+\cos 3\omega]$$

However, this gives this profile :

Which is not what I expected to get.

I have spent the past week reading a lot about the AMBER force field parameters and I ran out of ideas. So if anyone know how these dihedral angle are parametrized I could really use some explanation.