oct. 1976  r.goldstein
effect of general relativity on tmdly and doppler shift
this routine replaces an older one of the same name written by
mike ash in sept. 1967. the old routine was only good for
planetary radar.

note: this routine calculates a correction to a value of tmdly
that has been calculated and sent through common. for
doppler, dop is computed using a relativistic formulation
and previous "classical" calculations are overwritten.

redesigned 1978 june - j.f.chandler
now all radar routines should call radrel for both time delay
and doppler (skip doppler call if no doppler).  the logic to
determine whether apply relativistic corrections is performed
here.  some of the setup needed for phase delay doppler is done
in radrel(1).

---------------------------------doppler-------------------------
this code is for instaneous doppler only (ict(21)=1).
when doing instantaneous doppler, time delay is assumed to be
also calculated, and the upper half of this routine (for tmdly)
is assumed to have been executed.

the next calculations  give delf/f (doppler).dop1,2,3 are the first,
second and third order parts of the 'classical' calc.except for a 3rd
order coef.which is included with a gen.rel. term in dop 4.   terms
from inclusion of gen.rel.(i.e. grav.effects) are in dop5 and dop6.
this expression is acc to order e-14. for e-12 accuracy dop4=dop6=0.