(Andr? Michaud) wrote in message ...
(Spud) wrote in message ...
Is there enough information to state the acceleration is directly
towards the sun ?

Yes.

What if there is an acceleration towards the earth ?

Isn't this worth investigating.

There is no need. Both crafts are on escape trajectories from
the Solar system in opposite directions, the Sun being the
central body of the system.

The attraction of the Earth is totally negligible at the
distances involved.

André Michaud

The reason why it matters, and is a viable target for investigation,
whether the acceleration is heliocentric or Earth-centric is not
because one thinks it might be a real effect. But because the most
probably systematic experimental errors would give an apparent
Earth-centric effect rather than heliocentric.

Something no-one seems to mention in the classic paper at
http://xxx.lanl.gov/abs/http://www....gr-qc/0104064.
is that they see a systematic DIURNAL acceleration approximately 100
times as great as the overall long-term average they are purporting to
measure. Which they cannot find any explanation for, and ignore after
1 paragraph of discussion.

Let me repeat. There is no way they could see ANY diurnal effect, let
alone 100 times greater than the long-term one, unless at this level
of accuracy there was a serious systematic error in the direction of
the assumed Earth-station - spacecraft vector.

Actually I think the explanation is probably quite mundane for the
whole effect. Either the position of the Earth antenna is not exactly
where the GPS calculations put it (working in the industry, I know how
few people ever bother about the proper conversion between WGS-84
ellipsoidal co-ordinates and ECEF). Or the spacecraft has a position
co-ordinate tangential to the Earth-spacecraft line not exactly where
they think. Remember, only the RANGE is known to within metres.

Either way, that puts the observation vector off at an angle, which
introduces a small diurnal variation and probably a small systematic
error.

Also, there is the fact that the RANGE itself is not well-defined in
their measurement procedure (although the rate of change ought to be).
This is because of the PLL which is used to "reflect" the outgoing
transmission at the spacecraft. The assumption is that the spacecraft
transmitted frequency is instantaneously equal to N times the received
frequency. Of course, we don't have the engineering details, but the
spacecraft PLL is almost certainly second-order, and the phase noise
spec given in the paper mandates the PLL bandwidth to be sub-1 Hz. The
implication is that for the PLL to "converge", the output frequency
equals N times the input frequency as it was at least 1 second ago.
However, the PLL bandwidth - signal lag will remain constant over
time.

Similarly, the modulated ranging signal (as opposed to Doppler) has a
finite latency through the spacecraft, which does not seem to have
been modelled. These are small errors, but when the claimed accuracy
is ~10^-10 m/s^2........