"Henri Wilson" HW@.... wrote in message
...
On 29 Mar 2007 00:09:21 -0700, "George Dishman"
wrote:
On 29 Mar, 01:32, HW@....(Henri Wilson) wrote:

I think I had it right before.
The distance for 45 deg phase difference is about 0.0007 LY.
It is independent of velocity.

OK, that is the sort of value I would expect. Now
the general gist of my argument is this: you get
a 45 degree phase shift at 0.0007 LY so you would
expect to get of the order of 5 degrees at a 1/10th
of that distance where the ADoppler only adds a
small fraction to the VDoppler.

You made the point that an elliptical orbit could
look circular provided the periastron was on the
line of sight because the distortion of the sine
wave from the variable speed is cancelled by the
distortion caused by the c+v effect.

A slight change in your yaw factor could then
change the relative phase of those factors to
give a net phase change of a few degrees. That
could cancel the phase shift due to ADoppler
and again make the orbit look circular.

The distortion of the brightness curve for circular orbits looks quite
symmetrical. I tried varying the yaw angle very slightly but it skewed the
curve away from a sine wave.
I think the major axis has to be aligned witrh teh LOS. However, we don;t
know
how acccurate the published curves are....so you are probably right.

I've already replied but I didn't have time to explain
where the numbers came from so here's a bit more detail.

The residuals on the timing measurements are measured at
74ns compared with a pulse period of 2.295ms. They say
somewhere that if they can reconfigure the receivers to
make better use of multiple channels, they should be
able to get that down to 10ns.

If yaw distorts the shape rather than changing the phase,
a crude estimate of the phase accuracy is 74ns in 2.295ms
or about 3 parts per million, that is 0.011 degrees on
the phase.

The bottom line then is that knowing we see what
looks like a circular orbit (or at least very low
eccentricity) there is a relationship between the
extinction distance, the true eccentricity and the
yaw.

Well I can telll you one thing. The extinction distance is directly
proportional to period.
The 0.0007 value is for a period of 0.0042 years.
It becomes 0.007 for 0.042 years, 0.07 for 0.042 years..etc.
...always independent of peripheral velocity.

How can you explain THAT?

See my addition to Leonard's response, you require an
_incredible_ coincidence between the inclination from
which we are viewing the system and the properties of
space along the line of sight.

The end result should be an upper limit on the speed
equalisation distance based on the uncertainty in the
orbital phase and the eccentricity.

I will upload my program to the website George so you can fiddle with it.
It is by no measns complete but is OK for circular orbits.

Click the red button after selecting eccentricity then click either the
yellow
one (for my original method) or 'george' for your quick method.
George has the VDoppler correction included...
Increase 'output size' to see the curve at short distances.
If you hold the mouse button down, a vertical line appears on the screen
to
compare phases.

http://www.users.bigpond.com/hewn/newvariables.exe

I'll have a go over the weekend but it depends whether you
can set numbers sufficiently low. If extinction is around
one light minute as I suspect, the GUI is going to be
inconvenient.

George