[daip] cpass

[Eric Greisen]

I think I have found the source of the change in CPASS.  After the
fitting routine returns an error code I added a test on that error
code which exits rather than write a BP record.

revision 3.14
date: 2002/08/07 21:55:26;  author: egreisen;  state: Exp;  lines: +18 -10
ignores error codes!

Studying it further, it appears that there is an error code which
might not mean much (3) as well as then one that says exceeded the max
number of iterations.  Giving it some VLA data (already calibrated
with BPASS and with edge channels discarded) I found that I had to
tell it 400 iterations and a large convergence tolerance (1) to change
the error code from 2 to 3.

I suspect that the program never converged and the "error" was to make
it care about that fact.

I will work on it some more - perhaps change defaults and have it not
care about error 3.

Eric Greisen