Re: Pioneer 10 test of light speed delay

From: George Dishman (george_at_briar.demon.co.uk)
Date: 02/27/05 

Date: Sun, 27 Feb 2005 15:15:16 -0000

<r9ns@verizon.net> wrote in message
news:1109465111.878935.202770@g14g2000cwa.googlegroups.com...
>
> George Dishman wrote:
>> <r9ns@verizon.net> wrote in message
>> news:1109351063.725776.320250@o13g2000cwo.googlegroups.com...
>> >
>> >
>> My test is designed to test _your_ theory, not the
>> conventional theory, so again this is irrelevant.
>>
> But if the conventional theory implies something which is impossible

If conventional theory implied something impossible,
it would be wrong, period. That's the end of the
matter. In reality, the radiation pressure needed to
explain the anomalous acceleration is about 3% of the
RTG output so it is nowhere near "impossible".

> such as the larger than claimed anomalous acceleration and that this
> could be avoided by recognizing that the speed of light delay does not
> extrapolate beyond a second or so then your test is invalidated simply
> for this reason.

Wrong. If your theory implies something which is
impossible, the craft being in two places at once,
then your theory is wrong whether the conventional
theory is also wrong or not. The two questions are
entirely independent.

> But lets get to your argument again. You say that the effects I
> mention taken individually are each small but ignore their
> interaction which is large. Lets imagine first that the earth is not
> tilted to the solar plane and is spinning only but not orbiting.

Right. That is equivalent to the Earth's equatorial plane
and is the correct orientation for considering the test.
We can come back to the orbital part.

> I imagine the craft is say at 30 degrees above the solar plane and that
> the earthsites in the two cases are on the equator.

OK.

> Then it would make sense that given one earthsite doppler shift is
> minimal at a certain time when the earthsites are facing away from the
> sun and toward the celestial sphere

No. First think of a plane containing three point,
the North pole, the South pole and the craft. The
Doppler is exactly zero when the motion of the site
carries it through that plane.

Note that is independent of the (example) 30 degree
elevation of the craft and depends only on the fact
that the site is moving perpendicular to the plane.

> that a subsequent earthsite at the
> same latitude eight hours away should be minimal eight hours later and
> you say this is 100 minutes later or earlier.

Treating the second site in the same way, the same
analysis applies and yes, there is a 100 minute
discrepancy. To explain this, your aim is to show
that any errors add up to exactly that time.

> Now if this is made more realistic and the tilt and orbital motion
> and the difference in latitudes and the time of day,etc it is not at
> all clear even from your numerical estimates of these effects taken
> individually on the projections of the orbital and spin motions on the
> earthsite-craft lines at the different times and positions that the
> combined effects here are as you claim.

OK, let's do them one at a time.

Time of day is trivial, the data in the file is all in
the same time system, not local time, so no conversion
is required and there is no error.

In addition I confirmed the final calculation
(converting time difference to angular error) used the
sidereal day so there is no error here either.

The effect of latitude (and altitude incidentally) is
simple. If the distance from the centre of the earth
is R, then the radius of the circle produced by the
rotation of the Earth is r = R * cos(L) where L is the
latitude, and the magnitude of the Doppler is
proportional to r. My method finds the time of the zero
crossing of this contribution hence the magnitude has
no effect whatsoever. Again there is no error here.

Orbital motion is easier to deal with by taking a
different viewpoint. Imagine now we are still using
inertial (non-rotating) coordinates and the x-y plane
is still the equatorial plane of the Earth but the
origin is the barycentre of the solar system. Now
the Earth is moving in a large ellipse and the orbit
is tilted by 23.5 degrees.

That motion has to be projected onto the Earth-craft
line and that projection produces a small error which
depends on the cosine of the angle of projection and
the cosine of the tilt. My previous calculation is for
the worst case where both cosines are 1 (though
obviously that for tilt will actually be less than 1).

So there you go, there is only one source of error and
as I calculated before it is two orders of magnitude
less than the discrepancy in the results.

You talk also of combining the effects which is a
reasonable comment, but the effects of tilt and
projection onto the Earth-craft line both produce
cosine terms so can only _reduce_ the error below
the value I calculated.

> I snipped your numerical estimates because it wasn't clear they were
> correct

OK, but to convince yourself that I didn't cheat, and
that the base numbers from Craig's analysis were
reasonable, you should repeat this test with the values
you downloaded yourself. There is no other way you can
be confident. You will also learn some things that
should explain why your own Excel *** approach needs
some additional work.

> and even if they were, they begged the question of the
> combined effect.

We have now dealt with that. I think the key point is
understanding why the choice of coordinates is important.
Another key aspect is the use of the second derivative
on the magnitude of the orbital error but that's for
another day.

>> >>
>> >> The evidence is there for all to see in the papers.
>> >>
>> > This is a lie.
>>
>> The paper has been peer reviewed and cited by dozens
>> of other researchers, none of whom had any trouble
>> understanding it.
>
> Professional courtesy but not believable science.
>
> The details should have been presented but they were not.

The details have been presented, but perhaps they are
not what you expect. The equations used are all listed
in the paper but the way these are implemented in the
CHASMP and ODP software is proprietary. They do however
explain key differences in terms of the effect of track
durations.

> For example
> on my web page, I show how an initial estimate of craft position and
> velocity is changed to produce a doppler shift which, in combination
> with the assumed transmission earthsite motion a minute before the the
> received earthsite motion, produces a total doppler shift of the
> recorded transmission frequency that equals the observed received
> frequency.
> The next craft position and velocity are calculated from the previous
> position and motion then adding the motion due to the attraction to
> the sun increased slightly by the anomalous acceleration.
> The resulting value gives a predicted received radiation that is not
> the same as the actual radiation so that the initial position and
> velocity must be reevaluated and the earthsite motion at a time a half
> a minute earlier initially and a half a minute earlier for the
> subsequent earthsite motion a minute later, could be used to calculate
> the predicted received radiation for each of these two receptions a
> minute apart.
> Some such description of the method and calculated values should have
> been shown in the Anderson paper but they were not because they ran out
> of funding.

The Anderson paper explains:

 a) that the initial estimates were extrapolated from
    the data prior to loss of the ranging system

 b) that the motion is then calculated over the whole
    eight year period and was subsequently extended

 c) the equations used for the calculations, in
    particular eqn [2], eqn [3] and eqn[4], but also
    eqn [8] through eqn [12].

 d) the effects of conscan manoeuvres and how their
    influence was estimated by correlation with the
    craft spin

 e) that the initial conditions were then optimised
    to minimise the residuals (the total rms error)

 f) describes the tests they did to show that the
    final result was independent of the choice of
    the initial estimates

 g) the possible sources of systematic and random
    errors and tabulates their magnitude.

You have yet to do anything comparable.

> The fact that the errors according to the Horizons ephemeris are much
> larger and indicate a growing trend suggests that even if earthsite
> motions were made accurate to fractions of a minute, the randomness of
> these errors would not eliminate the large trend.

In principle that is what they say too, the random
errors cannot explain the trend, but the trend is
0.001Hz per day so your sample of three results
over two minutes should have a drift of 0.00000139Hz.
The raw data is nowhere near that resolution so
without averaging thousands of points, you don't
have a hope of seeing the drift.

Since there are six parameters in the initial
location and motion vectors, you would need a
minimum of six points plus a seventh even with
perfect data.

The real reason you get higher errors is that you
haven't used the correct equations for the
conventional model as given in Anderson et al.

> Perhaps some sloppy thinking and the need to minimize the anomalous
> acceleration led to the claimed conclusion.

No, using the correct relativistic equations gives
that value, as independently confirmed by Craig.
The "sloppy thinking" is all yours.

<snip>
>> As for the nonsense about "orbital correction", the
>> orbital motion produces less than one minute of
>> error worst case while the discrepancy is more than
>> 100 minutes. Again I have said this over and over
>> and the first time I posted the full table of values
>> so you could check yourself. If you doubt it, use
>> groups.google.com. The reason your theory fails the
>> test is that you have NO explanation for that
>> difference.
>>
>
> Your values do not support this claim.

The claim is merely a restatement of the value. Of
course I expect you to confirm that using the raw
data.

> And the reason your test
> fails is because of the combined effect of all of the so called
> negligible effects taken individually, completely destroy your claim of
> a test.

Since there is only one error, your claim is
meaningless, the result stands.

However, you seem to still be missing the bigger
picture here. When anybody presents a new scientific
theory, it is up to them to do three things:

 a) define the theory and show how others can use
    it (it must be repeatable)

 b) list and explain tests that have been performed
    using existing experimental or observational data

 c) demonstrate that the theory gives accurate
    predictions for all of those tests.

I am helping you out by suggesting a test for you
when really you should be coming up with these
tests on your own. The onus remains on you to do
the calculation and show you get the correct answer.
Finding excuses to avoid doing the work isn't going
to get you anywhere.

George

Quantcast