Dirk:

I have remade the codes that write netcdf files for tgcmidl,
with checks for negative ions. This includes mkcdf_sut, 
mkcdf_mut, and mkcdf_mutglb as tgcmidl reads these 3 types
of netcdf files. You should be able to rerun your mkcdf_sut
case w/ Barb's history.

I agree tgcmidl is behaving only so-so. The 5-min histories
are creating some problems esp in mut and mutglb codes. As
you probably know, idl is not as good as ncar graphics when
it comes to rotating map projections to fractions of a degree.
(if you want to make animations, a different idl application
is better -- see $TGCMROOT/animxdr -- let me know if you want
it, it reads xdr files made by tgcmproc).

I did make 2 minor changes to tgcmidl which you may as well
do as well. Copy $TGCMROOT/tgcmidl/sut/newfile.pro (this names
model type "TIEGCM-P7" when reading netcdf file made from 
history that goes up to zp +7). Also copy 
$TGCMROOT/tgcmidl/glbanim/rdcdf_ga.pro (this fixes small bug).

What is the specific error msg re the msis.so file? Is it
finding msis.so at all, or undefined symbols or what? We should
be able to remake msis.so on your machine if necessary. Do
hwm93 and iri work ok? I assum We will be going over to idl 5.0
in the near future, so maybe I will get same problems then.

--Ben