On 24 Feb 2007 05:52:25 -0800, "George Dishman"
wrote:


"Henri Wilson" HW@.... wrote in message
.. .


...
The normal Doppler is there of course, I haven't disputed
that, but it isn't the whole story. The pulse spacing is
also affected by what I describe above and you need to take
that into account AS WELL to get the full answer.

The program takes everything into account.
Why don't you experiment with it?

I have, it gives the wrong value because you only take
the bunching due to c+v vs c-v into account on the
brightness curve, not the velocity curve. The predicted
velocity is derived from the time between pulses so you
need to take into account there too.

Ah, I think I know what you are saying now.

Yes you do vbg!! Put it in your diary, it has taken weeks
for you to see this but the penny has at last dropped :-)

The penny has dropped as to why you are making your mistake.

You are claiming that the closer the bunching, the shorter the
wavelength...and
the higher the observed doppler shift.

Yes Henry. The frequency is the pulse repetition rate
which is what is used to determine the speed, so the
"wavelength" is the distance between consecutive pulses.

see below.

Yes that should be true for the pulsar...but it is not true in my
program.....

Right, that's the program error I have been describing
to you all these weeks :-) At least now you know what
the problem is.

It is not an error.
YOU are making an error of interpretation. see below

So what is the difference?
The difference is that it is the change in the actual width of the pulsar
pulses in each bunch that is analogous to what my program does.
Do you see what I mean?
In reality, the width of the pulsar pulses varies by the same doppler
fraction
as does the distance between pulses. (~90 parts per million)

Correct, the pulse width is about 1.5% of the period and
that factor remains constant as the pulses travel.

So you have to average the WIDTHS of pulses and NOT their spacing to
generate
your equivalent of my red curve.

As you say, the width varies by the same fraction as the
gap so it doesn't matter whether you take the ratio of
the observed width to mean width or of the observed gap
to the mean gap, they should give the same apparent speed.

No No No.
For BaTh, the width of the actual pulses DOES NOT change after emission. The
spacing between them DOES.....because they are moving at different speed wrt
the Bcentre.
So the spacing...or bunching... is not a direct indicator of doppler shift or
relative source velocity..
The PULSE WIDTH is.

So my red curve is a measure of the average pulse width arriving in a fixed
(observer) time interval.
For light, it indicates the average 'wavelength' of the light arriving in the
time interval.

If that was your intention it should have worked but the
curve on the screen doesn't tie up with that and I suspect
if you showed the numerical value of the peak it would be
too low. That's why I have been saying there is a bug in
your software.

Now you know why there is not.

The bunching itself is an indicator of brightness variation.

Yes, but it also affects the _apparent_ Doppler shift so
affects the apparent speed as well, that's the speed
calculated by astronomers which is based only on the PRF
and which you show as the red curve.

No. Bunching due to different relative speeds does not affect the original
wavelengths of the light. (or widths of the pulses)

What a lot of people don't realise is that no doppler shift occurs at the
source, in BaTh.
For instance, all the photons making up H.alpha light from all moving sources
has the same 'absolute distance' between 'wavecrests'.

That absolute distance will change during any change in speed as the photon
crosses space....so the observed wavelength anywhere will still reveal relative
source/observer speed. In other words, gratings still measure true doppler
shift in BaTh.

I'm becoming a bit confused as to what we are actually talking about
now.

At any point arond the orbit, pulses are being
sent with a time gap of 2.95 ms. That gap is
reduced at the receiver for two reasons:

a) the velocity of the pulsar towards the receiver
means that consecutive pulses travel different
distances. That is the normal Doppler effect.

b) if the pulses are transmitted at different speeds
then faster pulses can 'catch up' to slower ones
reducing the gap (or 'fall behind' if the second
pulse is slower increasing the gap) and hence the
time between reception depends on how much of this
effect happens before extinction equalises the
speeds. This effect is not taken into account
in published velocity curves so the published
values will be higher or lower than the simple
Doppler value.

That effect is not indicative of source velocity.

Correct, but it does affect the velocity which an astronomer
would calculate from the pulse timing. Remember you already
have " The blue curve is the true velocity of the source
wrt the observer." so we are here talking about " The red
one is the velocity curve that a distant observer would
calculate as true using doppler shift ...".

Those quotes are from your other post. It is the latter curve
calculated from the pulse rate that I have been telling you
is wrong. Now you understand why.

No George. I now understand why YOU and all the astronomers are wrong.
The true measure of pulsar orbit speed can be obtained by observing the PULSE
WIDTH variations.......which are quite small....and NOT the variation in the
pulse arrival rate.
In fact we have here a good way to check the BaTh....if we can get reliable
data..
I say the pattern of variation of pulse width over each orbit cycle WILL NOT be
quite the same as that of the 'bunching'.
However I think the difference would be too small to measure because the pulses
are generally assumed to have constant width.

Part (a) is dependent on the radial component of velocity
at the time of transmission, part (b) depends on the
acceleration at the same time and of course both vary
round the orbit. Your program includes effect (a) but
not effect (b).

You'll have to rethink this in light of what I have said.

I don't need to rethink, you have now understood and stated
the problem. What you need to do now is make the program
produce the correct red curve and preferably show the peak
velocity value as text like your min/max brightness.

I thank you for pointing out the 'problem' George.

I eagerly await your reply....

George

"When a true genius appears in the world, you may know
him by this sign, that the dunces are all in confederacy against him."
--Jonathan Swift.