[daip] Fwd: Re: UVFIT

[Eric Greisen]

-------- Forwarded Message --------
Subject: Re: [daip] UVFIT
Date: Thu, 24 Mar 2016 15:17:33 -0600
From: Eric Greisen <egreisen at nrao.edu>
Organization: National Radio Astronomy Observatory
To: Andy Biggs <abiggs at eso.org>

On 02/19/2016 12:24 PM, Andy Biggs wrote:
> Hi there. Is UVFIT still believed to work? I am finding it difficult to
> get anything sensible out of it. I am using it to fit to VLA data which
> include e.g. about a minute's worth of 3C84 i.e. very nearly a point
> source of 21 Jy. Unless I set dopos=-1, the fluxes are just gibberish
> and usually returns the input guess.
>
> If you'd like to investigate, I have put a dataset at
> ftp://ftp.eso.org/pub/general/abiggs/UVFIT.FITS
>

There is something I do not understand about this data set.  UVFIT and
the new task I wrote with the other method fail on it so long as DOPOS=1
whether with your data or model data.

But I just tried other old VLA data sets with models and they work even
if I cut the time range down to almost nothing and scale the baselines
so that they are as long as yours.

I do not understand....

Eric Greisen