Thanks to a variety of respondents, especially Clemens Anklin, for putting
me on the right track. The fundamental issue is that f2 and f1 are
swapped in the vconv-converted data. (I found that trying to think about
"used to be f2, now f1" made my head hurt, but it was a significant
improvement just to think about the indirect planes as carbon, nitrogen,
or proton.)
The default imported parameters set fnmode to echo-antiecho for f2 and
States/TPPI in f1. That has 0% likelihood of working, but I found I could
quickly find a working recipe by typing s fnmod and initially just
reversing those, plus setting 90 degrees of zero order phase correction in
f1. From there, there were just three ways things could go wrong. If
signals appeared both in their correct position and their mirror image
position vertically, the fnmode setting of echo-antiecho vs. some type of
States was wrong. If the spectrum was split in half and the top was
switched with the bottom, the setting of States vs. States/TPPI was wrong.
If the entire plane was upside down, the setting of the reverse flag was
wrong, and the 90 degrees of zero order phase correction in f1 needed to
be -90 as well. Once all those were set, ft3d gives the correct result.
The indirect chemical shifts are intially somewhat arbitrary, but typing s
sfo1 brings up the absolute carrier frequencies for the 3 channels, which
can be edited to match the nucleus actually in each indirect plane. Then
I used the spectrum calibration widget and plugged in the
sample-temperature-corrected chemical shifts corresponding to the center
of each plane calculated from tof, dof, and dof2 through the nmrPipe
"varian" conversion GUI. That covered the essentials.
David VanderVelde
Manager, High Resolution NMR Facility, Caltech
davidv_at_caltech.edu
Received on Fri Feb 04 2011 - 08:09:27 MST