"Henri Wilson" HW@.... wrote in message
news
On Wed, 29 Aug 2007 18:44:27 +0100, "George Dishman"
wrote:
"Henri Wilson" HW@.... wrote in message
. ..
On Sat, 25 Aug 2007 13:32:44 +0100, "George Dishman"

George, as far as I'm concerned, everything you say is riddled with
mistakes.
For a start, I certainly DON'T accept that the radius varies by 12%

Tough, that is the value directly measured by ESO.

it is wrong.

Rubbish, the speed of light in air is the same
as in conventional theory so the results are
the same.

Light traves a long way before it reaches Earth's air.

But that has no effect on the interferometer,
all of it is on Earth ;-)

Interferometry doesn't work in BaTh.

Probably, but interferometry works in the real
world so that is only a problem for the theory.





George, does our sun noticeably vary in size?

No, and if ESO used the interferometric technique
it would show as constant.

I doubt if anything would show up at 50000 LYs...

Probably not, that's why they started with L Car,
it is one of the closest and largest Cepheids and
subtends the largest angle.

Interferomery will give a distorted answer.

Nope, there is no distortion introduced
by ballistic theory.

but if it did it could easily
show a willusory varying radius rather than a constant one....because of
the
varying c+v.

Sorry Henry, you have to do better than hand waving.
The speed at the interferometer is the same across
the instrument so the interference pattern is
unaffected.

George, the technique is highly suspect at best.

Garbage, it is no more suspect than a grating.

Add variable light speed and
it becomes almost useless.

It has no effect, you only want to wave it
away because you cannot stomach the truth.

To a distant observer, our sun will appear to vary in both BRIGHTNESS
and
LUMINOSITY by the same fractional amount every 12 years due to its
orbit
around
its barycentre with Jupiter.

But not in temperature.

there would even be a willusory temperature variation due to ADopler
shifting
of the Planck curve..

Nope, the shift is only 0.01%. The K band is from
2200nm to 2400nm so the median shift is 0.22nm.
How much does that change the intensity in the band
for a Planck curve at ~6000K? It is utterly negligible.

Don't be so hasty George.
The Planck curve deals with PHOTON DENSITY in a particular band.

Intensity Henry.

Photon density variation due to ADoppler DOES NOT include my 'K'
factor...so
your figure of 0.01 is not anywhere near the correct one.

Wrong, the figure is the measured shift. I am saying
Cepheid surface speeds are typically less than 30km/s
so 0.01% is an upper limit. Whether that is caused by
VDoppler or ADoppler doesn't matter, the shift is no
more than that value. That means no more than 0.24nm
worth of the band moves out at one end while about the
same amount moves in at the other.

This is going to become pretty complicated so I will think about it.

Do that, you are obviously missing the point at the
moment.

George, quite clearly, if L Car is a huffpuff, its maximum temperture
should
occur about 30 degrees BEFORE minimum radius....when the 'exploding'
core
bangs
up against the contracting outer layers. Would you not agree?

Nope, you are taking only a single aspect without
considering the overall structure, it is hopelessly
naive. The light valve passes the light as some time
but it then has to work up through ~5% of the star's
radius to the photosphere. The mass of the gas has a
huge inertia but the light pulses are driving an
acoustic resonance and you know what that means for
the phase. There are shock waves propagating through
the region and the gas contains multiple species and
you have to take all of that into account. The bottom
line remains, the conventional modelling matches the
observations.

Naturally it would...because the required answers were already known.

That doesn't mean the model can be made to fit.

When I
match KNOWN curves, you say I just fiddle with a curve matching program
till I
get the right answer.
You should be consistent George.

I am. You know that for simultaneous equations
you can find a solution if you have as many
equations ars you have variables. You have
numerous parameters you can alter to get a fit
and basically if you have say ten variable, you
can do a Fourier fit of up to the fifth harmnonic
with sin and cosine terms (or amplitude and phase)
for each.

For Cepheid models you have basically the mass
of the star and to a degree the elemental
abundance. For any particular star you also have
the age but the model has to fit over the full
evolution of the star so that isn't really free
from a modelling point of view. Also mass, age
and chemistry can all be constrained by observation
so there is no significant scope for fiddling.

You should also realise that both temperature
and size variations MUST affect the luminosity
to some extent therefore you cannot fit that
curve as if it was ONLY due to ballistic effects
and expect to get a valid set of parameters.

True to some extent....but in light of what I said above, the
temperature
phasing should be similar to that of the ADoppler brightness..so there
shouldn't be all that much of an error.

The temperature and luminosity curves are similar
for optical bands since the temperature is the
key driver, but the ADoppler curve can be quite
different, you need to model it by fitting the
velocity curve to find out. That's what I have
been telling you for ages.

...and have been pointing out that the velocity curve should be similar
in
shape an phase to the luminosity curve...but you never listen...

No, check the top of this post, you were arguing
that the luminosity peaked with the acceleration,
not the velocity.

That is why I keep telling you that the only
way you can get a valid analysis is to fit
your predicted curve for the observed velocity
and then work back to get the true velocity.

You still don't understand that I feed into the program the TRUE orbital
parameters including velocity.

I know that. What you need to do is alter your
program so that it predicts what would be the
OBSERVED velocity curve based on spectral line
shift using the values you feed in and ballistic
theory. Then adjust your true values until that
prediction matches the actual OBSERVED spectral
line shift (which you get from the published
velocity curve). Then when you have got a match,
your program will predict the luminosity curve
and you can compare that to the actual curve
_after_ first removing the effects of radius
and temperature. You get the radius by integrating
your true velocity and the temperature from the
published curves by correcting the time of arrival
to account for changing c+v.

You are making lots of assumptions.

I am not making any, I am educating you on
how to go about doing a fit that won't use
information that you don't have.

Since I can produce the exact curves without including a temperature or
radius
changes, my conclusion could easily and quite justifiably be that neither
changes occur...except in the minds of relativists.

Since both changes are directly observed, your
conclusion is wrong. To justify it, you would
need to write down the equations and then solve
them to show that a star whoae temperature didn't
vary would produce a Planck-shaped curve over
multiple bands which varied _as_if_ the temperature
were changing due to some ballistic effect. You
can't do that because photon bunching due to
ADoppler and VDoppler is frequency independent.

That means we know the temperature _does_ change
and with it surface brightness so until you
subtract that part from the luminosity curve,
your results are badly flawed.

From that you can integrate to get the true
radius or differentiate to get the true
acceleration and from those AND the temperature
AND the filter bandwidths you could then predict
the luminosity curves.

No George. You have it all back to front. I can calculate K for a star
by
comparing the ADoppler produced luminosity variation with he OBSERVED
fractional velocity change.

Not until you correct the luminosity for
temperature and radius effects.

These are the willusory temperature changes of course....

The temperature _value_ is valid, only the arrival
time (orbital phase) would be offset by the c+v
effect.

What you have done to date omits so much that
it is meaningless.

You don't get it at all.

Oh I get it Henry, better than you. The vast
majority of the luminosity change is already
explained by radius and temperature changes so
until you remove those, any contribution from
ADoppler is unknown.

No George, you are still living in that imaginary universe in which
willusions
don't exist.

No Henry I live in the real universe, "willusions"
are you imaginary effect, but regardless your
ballistic equations not not result in "willusions"
on temperature measurements and probably not on
the radius measurement to any great extent.

Tough, you have given no alternative analysis.
Until you can apply ballistic theory to the
method and use it to prdict an alternative
radius, you haven't matched the observational
data.

The c+v variations will give the impression of a phase shift and
completely
confuse the interferometer.

Nope, the interferometer is only concerned about
the phase across the instrument of the light that
is arriving at a particular time. In fact the
interferometer will work with single photons (like
the gratings we discussed which are really a
particular type of interferometer) and obviously
each photon only has a single speed. Ballistic
theory doesn't suggest any form of distortion for
the instrument.

All photons arriving at different speeds will adjust to the same c/n on
entering Earth's atmosphere. Their absolute wavelengths will adjust
accordingly.
I gather that inferferometry effectively detects the angle subtended by
the
star.

Yes.

Small differences in emission times and relative velocities from each
side could markedly affect the results.

Emission times cannot matter because the light
is uncorrelated anyway. Speed differences could
matter but the light is moving at the same c/n
value when it reaches the interferometer so there
is no real scope for a distortion that way that
I can see.

You might want to consider the overall setup:

http://tinyurl.com/3dybf3

If you take either of those and plot the
difference between the radii versus the phase,
then square that and convert to the magnitude
scale, you get the residual in terms of
brightness. If you match the velocity curve with
your program as I suggested, you get a template
for the ADoppler from your brightness curve, and
you could then calculate a correlation with the
actual residual to find the magnitude of the
ballistic effect. However, it is obvious from the
plots that the error is so small the ADoppler
will be in the noise.

The ADoppler is responsible for most - if not all - of the luminosity
variation.

Nope, the radius is responsible for most in
K band and surface brightness due to the
temperature change for most in V band.
ADoppler, if it exists, is responsible for
the difference between the radius/brightness
combination and the observed amount.

but the published temperature cuve simply cannot be correct. Like I
said,for a
huffpuff, the maximum temperatue SHOULD and MUST BE just before the
point
of
minimum radius.

Think about a thick piece of metal heated from
one side by a blowlamp that is on for one second
every minute. We view the other side and the peak
temperature is shortly after the blowlamp goes
off. The size variation is an acoustic resonance
driven by the heat pulses via radiation pressure.
Trying to determine the phase relationship is far
more complex than your simplistic model.

The are two separate processes. There is an acoustic pressure wave that
causes
adiabatic compression and temperature rise. As radius increases, there is
also
an expansion that results in an adiabatic temperature DECREASE.

There is also simple heating due to the added
energy which is the more significant contribution.

Frankly I cannot see any obvous connection between the acoustic wave and
your
supposed largish radius change...or surface temperature.

I'll try to find the eigenstate plots which
make it clear.

Real models predict the eigenstates of the
oscillation, they get the harmonic content right,
they predict the 'bump' being in phase at 10 days
and the variation of that phase with luminosity.

They just fiddle the equations till they get the right answer.

Nope, in real science the equations have been
published and reviewed and can't be changed.
They are the same equations used for all other
branches of acoustics.

After all, that's just what Planck did to get his black body curve.

Nope, that's what Wien and those before him
did, they are called "empirical" laws. Planck
derived his equation from the postulate that
energy was emitted in packets each with an
energy proportional to its frequency and from
the statistics for occupancy of different modes
so he had no scope to adjust the equation at
all. The derivation is published so you can
repeat his process and you will get the same
answer. The fact that the equation fits the
observations is what gives confidence that
the postulate was valid.

Sorry Henry, interferometers here on Earth get
light moving at c/n through the atmosphere
even in ballistic theory so the result is
identical with BaTh.

No George. Light travels a long way before it reaches Earth..and the
phasing
between light emitted at slightly different times will vary accordingly
when
mixed with the interfrometer signal.

The phase is relevant to each single photon Henry,
you cannot have interference between light emitted
at slightly different times because it is thermal
(black body) so uncorrelated. The interferometer
is sensitive to arrival time differences for each
photon individually, the resulting curves are the
statistical sum of the photon flux.

I don't see how a single photon could be emitted by both sides of a star.
if it was, it would create NO interference.

No, no Henry each photon is emitted by a single
charged particle. Each photon passes through
both telescopes of the interferometer and
lands with a probability that depends on the
path length difference to create an interference
pattern matching the probability of landing at
some point. It is similar to the usual grating
equation. It is the overlaying of those patterns
from different parts of the star that alters the
contrast ratio of the fringes and tells us the
diameter.

No current astronomical principle is immune to the 'constant c' curse.

Repeating your dogma in the face of facts
just makes it obvious how you let your
religious convictions outweigh scientific
analysis. Ballistic theory says the
interferometric radius and the temperature
values are correct and only the time of
arrival is modified.

Both BaTh and thermodynamics say that the maximum temperature of a
huffpuff
should occur slightly before minimum radius.

The temperature curve is that of a relaxation
oscillator - essentially a sawtooth - with a
delay for the time for the light to reach the
surface. Acoustic theory says the motion will
be a driven and damped resonance with harmonics.
The resulting phase is not trivial to work out.

It certainly is not trivial.

Exactly, but when it is done, the models do match
the observed curves.

George