Radial Velocity - advice please

[Andrew Smith]

I have been looking at using the "Compute radial velocity difference" cross correlation function in ISIS to give me accurate RV differences. It seems to work well with my data and, as expected, to give random error an order of magnitude lower than trying to line fit individual lines.

However, I would like to fix the measurement to an absolute scale and have been looking a RV standard stars as an option. I picked HIP48113 as it is reasonably bright and well placed at the moment. Looking at it spectra in order 40 although not identical to Rho Cas it has features in common (see attached) and the correlation with my Rho Cas data seems very stable. My question is do you think this is a legitimate way of fixing an absolute scale? Do you think the correlation between different stars acceptable?

I have also been looking at how to monitor long and short term drifts and think that using the correlation of the O2 lines in order 33 is a good way at looking for long term drifts while comparing the Ar calibration lines in order 40 taken at different times during a run a way of looking at short term drifts.
So far with limited data the O2 and Ar drifts give hope that I can get between o.5 to 0.1km/s with my current S/N and processes.

Again comments welcome.

Thanks Andrew

PS in the attached HIP48113 has been shifted by 71km/s as indicated by the RV difference and heliocentric corrections.

[Robin Leadbeater]

Hi Andrew,

Seems a reasonable approach to me. Synthetic spectra are sometimes used as a templates eg with the CfA digital speedometer (presumably tuned with the same physical properties , metalicity, vsini etc as the target) eg
http://adsabs.harvard.edu/abs/1996ASPC...90...31L
but observing an actual star would have the advantage of including the effect of the instrument resolution and cancelling potential systematic calibration offsets.

Robin