[evlatests] spectral line observers beware

[Walter Brisken]

There is a very misleading statement in the on-line spectral line 
observers guide:

http://www.vla.nrao.edu/astro/guides/sline/current/node14.html

"Note that the calibrators must be observed at the same frequency as the 
target, although there is a few MHz leeway."

This statement is array configuration and zenith angle dependent in many 
cases an error of less than a MHz can be detrimental.  [A array near 
horizon can give 1.5 radian errors with a 0.1 MHz frequency difference, if 
my numbers are correct] Crystal and her student encountered this problem a 
month or so ago and originally thought it was a new EVLA issue.  Even if 
we don't make special mention of the calibration requirement that the 
expected phase error in radians after calibration is

delta_phi = 1/2pi * (w_cal*nu_cal - w_target*nu_target) / c

in EVLA literature, we should at least fix the web page.

-Walter