[daip] Question about SETJY and CVEL

[Eric Greisen]

On 08/01/2014 11:10 AM, Feng Gao wrote:
> Dear Eric,
>
>      I have some questions about the velocity settings/calculation in AIPS, I think you might know the answer.
>
>      Here is the big picture: I had some VLBI  observations of the 22GHz water megamaser and recently I found that there is an offset in frequency/velocity between the VLBI spectrum and the spectrum from GBT single dish observation, which are under the same velocity definition of optical and rest frame of LSR. The typical offset is about 100kHz or 1km/s, as we have three sets of maser lines, the offset are not exactly the same among them. Since our channel size is 25kHz, this kind of offset is much bigger than we expected and is really crucial for our final result, so I'm trying to find out the reason behind.
>
>      As far as I'm aware in our data reduction, the task SETJY & CVEL are the only place that we assign velocity to every channel. We observed the three sets of lines in one 128MHz band, then after correlation, the three sets of lines were "chopped" and saved as three different IFs, with each IF of 32MHz wide. So in SETJY, for each IF, here is an example of what I put in:
>
> sources='NGC5765b'''
> bif=1
> eif=1
> optype =''
> sysvel =8528.975
> restfreq= 22.23e9 5.08e6
> veltyp ='lsr'
> veldef ='optical'
> aparm(1)= 1
>
>      For the starting frequency of each IF, I found the corresponding velocity in the GBT spectrum(under LSR and optical) and use that as the 'sysvel' here.
>
>      Since I don't know how CVEL works, I wonder if CVEL could cause such an offset here. Does CVEL just shift each channel for the same amount or it is frequency depend? BTW, why do we need to specify the 'sysvel'? Would the rest frequency be enough to calculate velocity?
>
>      Another thing that I think might cause such offset is the doppler tracking, since we didn't do doppler tracking in our observations but just used the baseband with fixed frequency. In my understanding, the rotation of the earth is already corrected in the VLBA correlator model, then since our target is an extragalactic source and CVEL already correct for the earth's motion in the solar system, so we don't need to use doppler tracking. But maybe I'm wrong somewhere.
>
>      If necessary, we can talk over the phone, which might be easier to explain things.
>
>      Thanks in advance!
>
> Best,
>
> Feng Gao
>
>

I am not sure I understand what is happening here that is bothering you 
- is CVEL moving the lines around by more than you expect given what you 
entered in SETJY.

An interesting test is to let SETJY compute the velocity of the 
reference channel (optype='vcal').  Then compare what it gets with what 
you are telling it.

Eric Greisen