Dear UE,
Since your pulse sequence has D1 of constant duration, the line broadening you specified has no effect – because it dampens each data point by exactly the same factor. Thus, a saw-looking spectrum is quite justified if you do a zero-fill in D1.
You can play with the option -dw to access other modes of computing these things.
I noticed, however, that the explanation of this option in the Reference Manual is obsolete, inaccurate, and not very clear. So I've edited it as follows (which will be reflected in the new version of the Manual):
–dwnm...
Include the pulse sequences n, m, ... in the calculation of the time variable(s) t1 (for 1D pulse sequences) or t1 and t2 (for 2D pulse sequences). These variables label the data in the dimensions
of the pulse sequence and are also used in the computation of the line broadening specified at the line_broaden(Hz) line. By default, all non-D0 sequences are included in the calculation of t1 and t2. For example, in the pulse sequence
timing(usec) (100)256 100 (100)-256
t1 will be (0.1 0.2 0.3 ... )
if the option –dw12 is specified, and it will be (25.5 25.5 25.5 ... )
by default.
Best regards,
Mikhail