[Difx-users] How to do a clock-search?

[Phillips, Chris (CASS, Marsfield)]

Hi Eskil,

The system we use for real time fringe checking of the LBA should support this approach. In have used this to find fringes up to 300milliseconds in the past. Given the approach it takes (as it is desired to to do fully automated correlation based just on the vex file, with no hand editing of the .v2d file possible), I cannot 100% guarantee it will work for you. It can be a little fiddly to setup, though not particularly difficult. The system does support VDIF and Mark5b fine as well as mis-matched bandwidth (ie automatic zooms, as long as everything lines up).

I would typically run with a spectral resolution of, say 32K, and run difx with a 50% overlap of the defat window. The output is written in a simple text file of peak fringe amplitude vs delay - you can plot this in realtime and quite once a probably fringe is found.

NOTE: I have suspicions DIFX does not support clock offsets > 0.5 sec (fractional, offsets of integer seconds are handled fine).

This essentially does what you suggested in your last sentence but is automated and does not require data to be loaded onto AIPS etc. Probably does sacrifice some SNR, so marginal detections will be missed.

Regards
Chris




> On 27 Nov 2020, at 23:49, Eskil Varenius via Difx-users <difx-users at listmgr.nrao.edu> wrote:
> 
> Dear DiFX users,
> 
> I find myself wanting to explore a huge (1 second) space of possible clock offsets. Testing with spectral resolution 0.01 works fine, but only covers roughly +-50us. Trying the next power of 10, DiFX 2.6.2 runs but difx2mark4 fails, reporting
> 
> "fatal error: # visibilities (32000) exceeds array dimension (8192)"
> 
> Does anyone have a working method for how to clock-search the full 1-second space?
> (Shifting the clock_early values in VEX/v2d by 100us and stepping through using 0.01 resolution should work, but there must be a more elegant way?)
> 
> Kind regards
> 
> Eskil
> 
> _______________________________________________
> Difx-users mailing list
> Difx-users at listmgr.nrao.edu
> https://listmgr.nrao.edu/mailman/listinfo/difx-users