[daip] dftpl

[Eric Greisen]

Farhad Zadeh wrote:
> Hi Eric,
> I have been using DFTPL quite a bit for the last few years and now
> am experiencing two major problems. One is a VLBA  data
> that Andreas has reduced and I am trying to get a light curve.
> However, I see these zero amplitude visibilities which
> should not be there when the averaging  time is 5s in DFTPL.
> Andreas tells me that the data is sampled at less than less than a second.
> How can I remove these low level values? see below the parameters
> of DFTPL.
> 
> The other issue is that when I have a cube and use BCHAN and ECHAN in
> DFTPL gives me exactly the light curve of one channel but the 
> uncertainties reduced by the number of channels. DFTPL never looks
> at the values of each channel in the cube. There is certainly a bug in
> DFTPL. I'd be happy to send you the data if you can't confirm.
> Many thanks for your response.
> Cheers,
> Farhad
> 
> INNAME     'SA43_EL15_F1'          Input UV file name (name)
> AIPS 1: INCLASS    'APR1'                  Input UV file name (class)
> AIPS 1: INSEQ         1                    Input UV file name (seq. #)
> AIPS 1: INDISK        1                    Input UV file disk unit #
> AIPS 1: UVRANGE       0        140000      Range of projected spacings
> AIPS 1:                                      (thousands of wavelengths)
> AIPS 1: TIMERANG   *all 0                  Time range to select:
> AIPS 1:                                      Start Day, Hour, Min, Sec
> AIPS 1:                                      End Day, Hour, Min, Sec
> AIPS 1: SHIFT         0           0        1: source offset in RA (asec)
> AIPS 1:                                    2: source offset in DEC
> AIPS 1: FREQID        0                    Frequency ID number: 0 -> 1
> AIPS 1: BCHAN         1                    First channel to include
> AIPS 1: ECHAN         0                    Last channel to include
> AIPS 1: BIF           1                    First IF to include
> AIPS 1: EIF           1                    Last IF to include
> AIPS 1: BPARM         5           5        Control parameters
> AIPS 1:               0           0        1 : Correlator number
> AIPS 1:               0           0.2      2 : averaging interval (sec)
> AIPS 1:               1.2         0        3 : > 0.0 => fixed scale
> AIPS 1:              -2           0            < 0.0 => fixed range
> AIPS 1:                                    4 :
> 
> 

I have found the reason for your problem with channel number - the code 
had a typo that caused the increment for frequency to be zero so it 
always plotted spectral channel 1, summed N times.  I will fix that in 
31DEC08 and 31DEC09 today.

I looked at your plot and cannot tell you why some times have small 
amplitude and some do not.  Try setting a very short TIMERANG so that 
you can identify a time with the near zero amplitude.  Then use PRTUV in 
that time - I bet you will see some baselines with very discrepant 
phases so that the complex sum will cancel.

Eric Greisen