Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > sci.physics.relativity > #372785 > unrolled thread
| Started by | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| First post | 2015-12-26 14:49 +0300 |
| Last post | 2016-01-09 14:02 +0000 |
| Articles | 20 on this page of 37 — 8 participants |
Back to article view | Back to sci.physics.relativity
This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by
below is the oldest one visible, not the original post.
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2015-12-26 14:49 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2015-12-26 04:38 -0800
Re: A Particle Model of Light Waye Kitchens <wscom@complang.net> - 2015-12-26 15:06 +0000
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2015-12-26 21:57 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2015-12-26 13:17 -0800
Re: A Particle Model of Light Waye Kitchens <wscom@complang.net> - 2015-12-27 12:15 +0000
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2015-12-27 15:40 +0300
Re: A Particle Model of Light Waye Kitchens <wscom@complang.net> - 2015-12-27 12:41 +0000
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2015-12-27 20:11 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2015-12-27 14:32 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-02 22:05 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-02 16:24 -0800
Re: A Particle Model of Light Tom Roberts <tjroberts137@sbcglobal.net> - 2016-01-03 12:57 -0600
Re: A Particle Model of Light John Gogo <jfgogo22@yahoo.com> - 2016-01-03 13:26 -0800
Re: A Particle Model of Light John Gogo <jfgogo22@yahoo.com> - 2016-01-03 13:13 -0800
Re: A Particle Model of Light John Gogo <jfgogo22@yahoo.com> - 2016-01-03 13:21 -0800
Re: A Particle Model of Light Tom Roberts <tjroberts137@sbcglobal.net> - 2016-01-03 12:26 -0600
Re: A Particle Model of Light Kathleen Winder <katwin@mcgatlin.org> - 2016-01-03 18:48 +0000
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-04 05:29 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-05 20:19 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-05 11:15 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-06 01:03 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-05 17:28 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-07 15:41 +0300
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-07 05:02 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-07 21:00 +0300
Re: A Particle Model of Light John Heath <heathjohn2@gmail.com> - 2016-01-08 01:04 -0800
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-08 18:45 +0300
Re: A Particle Model of Light John Heath <heathjohn2@gmail.com> - 2016-01-09 08:33 -0800
Re: A Particle Model of Light Gary Harnagel <hitlong@yahoo.com> - 2016-01-08 20:50 -0800
Re: A Particle Model of Light Tom Roberts <tjroberts137@sbcglobal.net> - 2016-01-05 19:50 -0600
Re: A Particle Model of Light Kathleen Winder <katwin@mcgatlin.org> - 2016-01-06 15:33 +0000
Re: A Particle Model of Light Tom Roberts <tjroberts137@sbcglobal.net> - 2016-01-06 12:40 -0600
Re: A Particle Model of Light Kathleen Winder <katwin@mcgatlin.org> - 2016-01-07 22:27 +0000
Re: A Particle Model of Light Anton Shepelev <anton.txt@gmail.com> - 2016-01-08 19:15 +0300
Re: A Particle Model of Light Carmine Portacio <carpo@trilogybearing.org> - 2016-01-09 14:00 +0000
Re: A Particle Model of Light Carmine Portacio <carpo@trilogybearing.org> - 2016-01-09 14:02 +0000
Page 1 of 2 [1] 2 Next page →
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2015-12-26 14:49 +0300 |
| Subject | Re: A Particle Model of Light |
| Message-ID | <20151226144943.217e76b59ecc1b7155226772@gmail.com> |
Thomas 'PointedEars' Lahn to Anton Shepelev:
> > Thomas 'PointedEars' Lahn
> > > Anton Shepelev:
> > >
> > > > http://bourabai.kz/shtyrkov/shtyrkov.pdf
> > >
> > > Have you read this?
> >
> > I have, though I misdoubt me [that...]
> ^^^^^^^^^^^^^
> YGCIB.
Underlined is a perfect English phrase meaning "I
suspect," while such acronyms are best left to SMSes
and other instant messaging media plagued by illit-
erate teenagers.
> I have scanned the "paper". That was enough for
> me to see that it is a crackpot theory, not worth
> investing more time into it. The premise is
> wrong, and /ex falso quodlibet/.
/Diabolus in singulis est/. Ether wind is not pre-
sumed, but tested as a hypothesis:
a. If there is no ether wind, a geostationary
satellite will show no aberration.
b. If there is a galactic ether wind, aberration
will be observed with a geostationary satel-
lite.
Did you mean some other wrong "premise," which you
didn't care to name?
--
() ascii ribbon campaign - against html e-mail
/\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [next] | [standalone]
| From | Gary Harnagel <hitlong@yahoo.com> |
|---|---|
| Date | 2015-12-26 04:38 -0800 |
| Message-ID | <18e5c3b2-0a6f-4648-89f1-4b7b8f5c4300@googlegroups.com> |
| In reply to | #372785 |
On Saturday, December 26, 2015 at 4:49:21 AM UTC-7, Anton Shepelev wrote: > > Thomas 'PointedEars' Lahn to Anton Shepelev: > > > > > > > http://bourabai.kz/shtyrkov/shtyrkov.pdf > > > > I have scanned the "paper". That was enough for > > me to see that it is a crackpot theory, not worth > > investing more time into it. The premise is > > wrong, and /ex falso quodlibet/. > > /Diabolus in singulis est/. Ether wind is not pre- > sumed, but tested as a hypothesis: > > a. If there is no ether wind, a geostationary > satellite will show no aberration. > > b. If there is a galactic ether wind, aberration > will be observed with a geostationary satel- > lite. > > Did you mean some other wrong "premise," which you > didn't care to name? I don't know what Lahn was referring to in his judgment of the paper, but there was no error analysis, which makes the numbers and conclusions meaningless. That the "measured" motion only varied by 1% is very strange because that means any supposed ether whould be virtually stationary wrt the sun, yet the sun is orbiting the galaxy at 200 km/sec. The whole thing smells really bad.
[toc] | [prev] | [next] | [standalone]
| From | Waye Kitchens <wscom@complang.net> |
|---|---|
| Date | 2015-12-26 15:06 +0000 |
| Message-ID | <n5mad6$16u$1@speranza.aioe.org> |
| In reply to | #372789 |
Gary Harnagel wrote: >> Did you mean some other wrong "premise," which you didn't care to >> name? > > I don't know what Lahn was referring to in his judgment of the paper, but You must be smiling. Lahn and judgement? Try exclusive or. > there was no error analysis, which makes the numbers and conclusions > meaningless. That the "measured" motion only varied by 1% is very strange > because that means any supposed ether whould be virtually stationary wrt > the sun, yet the sun is orbiting the galaxy at 200 km/sec. > The whole thing smells really bad. Make sure your nose is properly qualified making such a preposterous statement. Error Analysing is not aimed to what you just implied.
[toc] | [prev] | [next] | [standalone]
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2015-12-26 21:57 +0300 |
| Message-ID | <20151226215753.2b0cb7b827872127ecc887be@gmail.com> |
| In reply to | #372789 |
Gary Harnagel about http://bourabai.kz/shtyrkov/shtyrkov.pdf : > but there was no error analysis, which makes the > numbers and conclusions meaningless. Less credible -- sure, but why meaningless? Whereas he got a sinusoidal signal instead of noise, the random error is small enough, and if the singal is to be accounted for by a systematic error then I ought to ask what kind of error would produce a di- urnally varying lag between the actual and observed locations? I agree that the question is better ad- dressed to Shtyrkov, though. The corresponding Russian article mentions briefly that meausurement accuracy was about 10% but the av- eraging over many observations yielded a precise- enough result: the observed average orbital velocity (29.4 km/s) coincides with the value accepted in as- tronomy (29.8 km/s) to about 1%. Is it a miracle? Tom Roberts thinks that this value arose from Shtyrkov's correction for a non-existent aberration, but I did not understand what he meant. > That the "measured" motion only varied by 1% is > very strange because that means any supposed ether > whould be virtually stationary wrt the sun, yet > the sun is orbiting the galaxy at 200 km/sec. The > whole thing smells really bad. No, it is simply that the galactic component of the ether wind (declinaton of 89.5 degrees) is nearly coincidental with the polar axis and has therefore but a tiny effect on diurnal aberration from the satellite. To explain -- the diurnal variations of the ether wind in the plane orthogonal to the source-observer line are caused mainly by the rotation of the Earth as it changes orientation with respect to its or- bital speed, while the galactic component, having a near-zero projection upon the equatorial plane, causes only a negligible variation. -- () ascii ribbon campaign - against html e-mail /\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [prev] | [next] | [standalone]
| From | Gary Harnagel <hitlong@yahoo.com> |
|---|---|
| Date | 2015-12-26 13:17 -0800 |
| Message-ID | <31e1800c-fae0-48d1-baa2-297230be6f29@googlegroups.com> |
| In reply to | #372799 |
On Saturday, December 26, 2015 at 11:57:38 AM UTC-7, Anton Shepelev wrote: > > Gary Harnagel about > http://bourabai.kz/shtyrkov/shtyrkov.pdf : > > > > but there was no error analysis, which makes the > > numbers and conclusions meaningless. > > Less credible -- sure, but why meaningless? Whereas > he got a sinusoidal signal instead of noise, the > random error is small enough, and if the singal is > to be accounted for by a systematic error then I > ought to ask what kind of error would produce a di- > urnally varying lag between the actual and observed > locations? The effect of temperature upon the instruments comes to mind. > I agree that the question is better ad dressed to Shtyrkov, though. > > The corresponding Russian article mentions briefly > that meausurement accuracy was about 10% but the av- > eraging over many observations yielded a precise- > enough result: the observed average orbital velocity > (29.4 km/s) coincides with the value accepted in as- > tronomy (29.8 km/s) to about 1%. Is it a miracle? Reducing the error by making many observations is an acceptable procedure. Is that what the little circles on the graph mean? Do they represent sigma? They are almost as big as the variation and you need at least three-sigma (preferably more) to have any validity. And they may well have over-massaged the data. See http://arxiv.org/ftp/physics/papers/0608/0608238.pdf as an example of how this can happen. Another point is that reducing error by multiple measurements only works if the data approaches a normal distribution. > Tom Roberts thinks that this value arose from > Shtyrkov's correction for a non-existent aberration, > but I did not understand what he meant. Well, stellar aberration is caused by a variation in relative velocity between earth and star. A GPS satellite has NO relative velocity variation so maybe that's what he was referring to? If Shtyrkov, et al, assumed there was stellar aberration and "compensated" for it. Similar mistakes in attempts to measure the supposed ether have been made by others: http://www.newtonphysics.on.ca/illusion/index.html http://www.ccsenet.org/journal/index.php/apr/article/view/10410 http://www.k1man.com/f61.pdf These, of course, go against a mountain of evidence that there is no such ether and that the speed of light is not variable. Here's a paper that refutes the last one above: http://arxiv.org/pdf/1002.3968v1.pdf > > That the "measured" motion only varied by 1% is > > very strange because that means any supposed ether > > whould be virtually stationary wrt the sun, yet > > the sun is orbiting the galaxy at 200 km/sec. The > > whole thing smells really bad. > > No, it is simply that the galactic component of the > ether wind (declinaton of 89.5 degrees) is nearly > coincidental with the polar axis and has therefore > but a tiny effect on diurnal aberration from the > satellite. > > To explain -- the diurnal variations of the ether > wind in the plane orthogonal to the source-observer > line are caused mainly by the rotation of the Earth > as it changes orientation with respect to its or- > bital speed, while the galactic component, having a > near-zero projection upon the equatorial plane, > causes only a negligible variation. > > -- > () ascii ribbon campaign - against html e-mail > /\ http://preview.tinyurl.com/qcy6mjc [archived] Well, you can see the galactic plain (i.e., the Milky Way). I don't think the normal to that goes anywhere near Polaris. Consequently, the component of the sun's velocity should be quite large, shouldn't it?
[toc] | [prev] | [next] | [standalone]
| From | Waye Kitchens <wscom@complang.net> |
|---|---|
| Date | 2015-12-27 12:15 +0000 |
| Message-ID | <n5okoo$khb$1@speranza.aioe.org> |
| In reply to | #372805 |
Gary Harnagel wrote: > On Saturday, December 26, 2015 at 11:57:38 AM UTC-7, Anton Shepelev > wrote: >> >> Gary Harnagel about http://bourabai.kz/shtyrkov/shtyrkov.pdf : >> > >> > but there was no error analysis, which makes the numbers and >> > conclusions meaningless. >> >> Less credible -- sure, but why meaningless? Whereas he got a >> sinusoidal signal instead of noise, the random error is small enough, >> and if the singal is to be accounted for by a systematic error >> then I ought to ask what kind of error would produce a di- >> urnally varying lag between the actual and observed locations? > > The effect of temperature upon the instruments comes to mind. ha ha ha ha, Gary doesn't know what an Numerical Error Analysing is all about. Listen my Gary, the temperature concern is taken care of by the manufacturer. It stays written inside their datasheet. Ha ha ha ha, thank you Gary.
[toc] | [prev] | [next] | [standalone]
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2015-12-27 15:40 +0300 |
| Message-ID | <20151227154055.34583e99802717367141b6dc@gmail.com> |
| In reply to | #372807 |
Waye Kitchens: > ha ha ha ha Your behavior only discredits proponents of the ether. If that's what you are trying to achieve then you have succeeded. Why not keep silent if you have nothing to say on the subject? -- () ascii ribbon campaign - against html e-mail /\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [prev] | [next] | [standalone]
| From | Waye Kitchens <wscom@complang.net> |
|---|---|
| Date | 2015-12-27 12:41 +0000 |
| Message-ID | <n5omag$nm6$1@speranza.aioe.org> |
| In reply to | #372809 |
Anton Shepelev wrote: > Waye Kitchens: > >> ha ha ha ha > > Your behavior only discredits proponents of the ether. If that's > what you are trying to achieve then you have succeeded. Why not keep > silent if you have nothing to say on the subject? Because you are stupid, not knowing what an Error Analysis is all about? You two are moreons, end of story.
[toc] | [prev] | [next] | [standalone]
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2015-12-27 20:11 +0300 |
| Message-ID | <20151227201105.7df72532b4eca77d29f8682a@gmail.com> |
| In reply to | #372805 |
Gary Harnagel to Anton Shepelev about
http://bourabai.kz/shtyrkov/shtyrkov.pdf :
> > Gary Harnagel:
> >
> > > but there was no error analysis, which makes
> > > the numbers and conclusions meaningless.
> >
> > Less credible -- sure, but why meaningless?
> > Whereas he got a sinusoidal signal instead of
> > noise, the random error is small enough, and if
> > the singal is to be accounted for by a systemat-
> > ic error then I ought to ask what kind of error
> > would produce a diurnally varying lag between
> > the actual and observed locations?
>
> The effect of temperature upon the instruments
> comes to mind.
It is possible, but without knowldedge of the instu-
ments employed by Shtyrkov, we can only discuss it
as a black box, i.e. analysing the results observed,
which are:
1. The observed azimuth oscillates a little ahead
of the reference curve, as would be expected
if there were an ether (eq. 7).
2. The average orbital velocity of the ether wind
calculated from the observed aberration of the
azimuth may be explaned by exactly such aber-
ration due to ether wind as would be expected
to arise from the orbital motion of the Earth.
3. Both seasonal and diurnal behaviours of the
observed aberration correspond to the predic-
tion (eq. 8) derived form the assumption of
the ether. Even the small variation of the
Earth's orbital speed due to its slightly el-
liptic orbit was resolved.
To ascribe it all to a systematic error seems far-
fetched.
> > The corresponding Russian article mentions
> > briefly that meausurement accuracy was about 10%
> > but the averaging over many observations yielded
> > a precise-enough result: the observed average
> > orbital velocity (29.4 km/s) coincides with the
> > value accepted in astronomy (29.8 km/s) to about
> > 1%. Is it a miracle?
>
> Reducing the error by making many observations is
> an acceptable procedure. Is that what the little
> circles on the graph mean? Do they represent sig-
> ma?
I fear not, and see no reason why they should be
circles and not errorbars. Time was measured with
very high precision. Might those circles be a sort
of eye-candy in the Origin program, where Shtyrkov
said he had processed his data?
> They are almost as big as the variation and you
> need at least three-sigma (preferably more) to
> have any validity.
Shtyrkov says nothing about the magnitude of his
signal with respect to the sigma.
> And they may well have over-massaged the data.
> See
>
> http://arxiv.org/ftp/physics/papers/0608/0608238.pdf
>
> as an example of how this can happen.
I have read Roberts's paper and disagree with some
of his points, about which I have written in this
very thread, in the subthread called "Dayton
Miller's measurements". I have not had time to an-
swer all of his counter-arguments, though.
What do you mean by over-massaging? Processing with
the assumption of a periodic effect as would make
this effect "appear" in the result, as Roberts
claims Miller did? I see no such error in
Shtyrkov's work, although fitting the predicted
curve and estimating the error would have been bet-
ter.
> Another point is that reducing error by multiple
> measurements only works if the data approaches a
> normal distribution.
Well, standard error of the mean is lower than that
of a single value by the square root of the number
of measurements as long the values are statictically
independent. They need not be normally distributed
for that to hold. Confidence intervals, however,
cannot be calculated without knowing the type of the
distribution. Normal distibution seems likeliest,
describing the sum of many small additive values.
> > Tom Roberts thinks that this value arose from
> > Shtyrkov's correction for a non-existent aberra-
> > tion, but I did not understand what he meant.
>
> Well, stellar aberration is caused by a variation
> in relative velocity between earth and star.
Exactly, and even if space were filled with ether
that would not change, for the (assumed-)constant
ether wind across the radius vector would effect a
*constant* bending of light-rays without variation.
Bradley's stellar aberration is therefore the same
for empty space and for ether.
> A GPS satellite has NO relative velocity variation
> so maybe that's what he was referring to?
I don't think GPS satellites are geostationary, so
their relative speed can't be zero, but is negli-
bibly small to produce a detectable *stellar* aber-
ration.
The situation will different if we assume there is
an ether. As the radius-vector of a geostationary
satellite rotates with the Earth, its orientation
with respect to the ether wind varies, and so does
the aberration angle. Mark you it is not stellar
aberration, where the radius vector points in a con-
stant galactic direction. If you will calculate the
effect of the hypothetical ether wind you will see
it is similar to the one in Bradley's ballistic mod-
el:
sin(phi) = Vep/c ,
where Vep is the projection of the ether wind onto
the plane orthogonal to the radius-vector.
If e denote a unity vector:
_ _ _
e(r) = r / |r|
_ _
and proj(a,n) the projection_ of vector a onto the
plane orthogonal to vector n:
_ _ _ _ _ _
proj(a, r) = a - (a, r) e (r),
then the apparent radius-vector of the source will
be: _ _ _ _ _ _
a = c * e(r) - proj(Vsrc,r) - proj(Veth,r),
where Vsrc and Veth are the speeds of the source and
ether with respect to the observer. In Bradley's
stellar aberraton the third term is zero. In
Shtyrkov's aberration from a geostationary satellie
the second term is zero.
> Similar mistakes in attempts to measure the sup-
> posed ether have been made by others:
>
> http://www.newtonphysics.on.ca/illusion/index.html
> http://www.ccsenet.org/journal/index.php/apr/article/view/10410
> http://www.k1man.com/f61.pdf
>
> These, of course, go against a mountain of evi-
> dence that there is no such ether and that the
> speed of light is not variable. Here's a paper
> that refutes the last one above:
>
> http://arxiv.org/pdf/1002.3968v1.pdf
I will comment on these articles when/if I shall
have read and understood them.
> > > That the "measured" motion only varied by 1%
> > > is very strange because that means any sup-
> > > posed ether whould be virtually stationary wrt
> > > the sun, yet the sun is orbiting the galaxy at
> > > 200 km/sec. The whole thing smells really bad.
> >
> > No, it is simply that the galactic component of
> > the ether wind (declinaton of 89.5 degrees) is
> > nearly coincidental with the polar axis and has
> > therefore but a tiny effect on diurnal aberra-
> > tion from the satellite.
> >
> > To explain -- the diurnal variations of the
> > ether wind in the plane orthogonal to the
> > source-observer line are caused mainly by the
> > rotation of the Earth as it changes orientation
> > with respect to its orbital speed, while the
> > galactic component, having a near-zero projec-
> > tion upon the equatorial plane, causes only a
> > negligible variation.
>
> Well, you can see the galactic plain (i.e., the
> Milky Way). I don't think the normal to that goes
> anywhere near Polaris. Consequently, the compo-
> nent of the sun's velocity should be quite large,
> shouldn't it?
I think you misunderstood me--I meant the Celestial
poles in the Equatorial coordinate system.
--
() ascii ribbon campaign - against html e-mail
/\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [prev] | [next] | [standalone]
| From | Gary Harnagel <hitlong@yahoo.com> |
|---|---|
| Date | 2015-12-27 14:32 -0800 |
| Message-ID | <80df2809-ac30-4673-8552-3269d5e29225@googlegroups.com> |
| In reply to | #372816 |
On Sunday, December 27, 2015 at 10:10:49 AM UTC-7, Anton Shepelev wrote: > > Gary Harnagel to Anton Shepelev about > http://bourabai.kz/shtyrkov/shtyrkov.pdf : > > > > > Gary Harnagel: > > > > > > > but there was no error analysis, which makes > > > > the numbers and conclusions meaningless. > > > > > > Less credible -- sure, but why meaningless? > > > Whereas he got a sinusoidal signal instead of > > > noise, the random error is small enough, and if > > > the singal is to be accounted for by a systemat- > > > ic error then I ought to ask what kind of error > > > would produce a diurnally varying lag between > > > the actual and observed locations? > > > > The effect of temperature upon the instruments > > comes to mind. > > It is possible, but without knowldedge of the instu- > ments employed by Shtyrkov, we can only discuss it > as a black box, i.e. analysing the results observed, > which are: But we can't even do that without error bars. Raw data would be more satisfactory. Do you know how many are required to reduce the error from 10% to 1? > 1. The observed azimuth oscillates a little ahead > of the reference curve, as would be expected > if there were an ether (eq. 7). > > 2. The average orbital velocity of the ether wind > calculated from the observed aberration of the > azimuth may be explaned by exactly such aber- > ration due to ether wind as would be expected > to arise from the orbital motion of the Earth. > > 3. Both seasonal and diurnal behaviours of the > observed aberration correspond to the predic- > tion (eq. 8) derived form the assumption of > the ether. Even the small variation of the > Earth's orbital speed due to its slightly el- > liptic orbit was resolved. > > To ascribe it all to a systematic error seems far- > fetched. Systematic error was what caused Dayton Miller to proclaim an ether wind. > > > The corresponding Russian article mentions > > > briefly that meausurement accuracy was about 10% > > > but the averaging over many observations yielded > > > a precise-enough result: the observed average > > > orbital velocity (29.4 km/s) coincides with the > > > value accepted in astronomy (29.8 km/s) to about > > > 1%. Is it a miracle? > > > > Reducing the error by making many observations is > > an acceptable procedure. Is that what the little > > circles on the graph mean? Do they represent sig- > > ma? > > I fear not, and see no reason why they should be > circles and not errorbars. Time was measured with > very high precision. Might those circles be a sort > of eye-candy in the Origin program, where Shtyrkov > said he had processed his data? Then I'm afraid that anything garnered from their paper is purely speculation. > > They are almost as big as the variation and you > > need at least three-sigma (preferably more) to > > have any validity. > > Shtyrkov says nothing about the magnitude of his > signal with respect to the sigma. Indeed, thus the paper is unscientific claptrap. > > And they may well have over-massaged the data. > > See > > > > http://arxiv.org/ftp/physics/papers/0608/0608238.pdf > > > > as an example of how this can happen. > > I have read Roberts's paper and disagree with some > of his points, about which I have written in this > very thread, in the subthread called "Dayton > Miller's measurements". I have not had time to an- > swer all of his counter-arguments, though. Apparently, you don't understand statistical analysis. > What do you mean by over-massaging? Processing with > the assumption of a periodic effect as would make > this effect "appear" in the result, as Roberts > claims Miller did? I see no such error in > Shtyrkov's work, although fitting the predicted > curve and estimating the error would have been bet- > ter. Just how did they reduce the error to 1% from 10%? How many measurements did they make per data point? > > Another point is that reducing error by multiple > > measurements only works if the data approaches a > > normal distribution. > > Well, standard error of the mean is lower than that > of a single value by the square root of the number > of measurements as long the values are statictically > independent. Were they? Who knows? > They need not be normally distributed for that to hold. Again, how many measurements did they make for each point on the graph? Was there a trend while they were measuring or were results purely random? > Confidence intervals, however, cannot be calculated without > knowing the type of the distribution. Normal distribution > seems likeliest, Yes. If they aren't, then there are some big questions. > describing the sum of many small additive values. I have no idea what you're talking about here. > > > Tom Roberts thinks that this value arose from > > > Shtyrkov's correction for a non-existent aberra- > > > tion, but I did not understand what he meant. > > > > Well, stellar aberration is caused by a variation > > in relative velocity between earth and star. > > Exactly, and even if space were filled with ether > that would not change, for the (assumed-)constant > ether wind across the radius vector would effect a > *constant* bending of light-rays without variation. > Bradley's stellar aberration is therefore the same > for empty space and for ether. I don't believe so: https://en.wikipedia.org/wiki/Aberration_of_light#Luminiferous_aether > > A GPS satellite has NO relative velocity variation > > so maybe that's what he was referring to? > > I don't think GPS satellites are geostationary, Oh, right. I was confusing them with the geostationary ones. My bad. > so their relative speed can't be zero, but is negligibly > small to produce a detectable *stellar* aberration. > > The situation will different if we assume there is > an ether. As the radius-vector of a geostationary > satellite rotates with the Earth, its orientation > with respect to the ether wind varies, and so does > the aberration angle. Mark you it is not stellar > aberration, where the radius vector points in a con- > stant galactic direction. If you will calculate the > effect of the hypothetical ether wind you will see > it is similar to the one in Bradley's ballistic mod- > el: > sin(phi) = Vep/c , > > where Vep is the projection of the ether wind onto > the plane orthogonal to the radius-vector. > > If e denote a unity vector: > > _ _ _ > e(r) = r / |r| > _ _ > and proj(a,n) the projection_ of vector a onto the > plane orthogonal to vector n: > _ _ _ _ _ _ > proj(a, r) = a - (a, r) e (r), > > then the apparent radius-vector of the source will > be: _ _ _ _ _ _ > a = c * e(r) - proj(Vsrc,r) - proj(Veth,r), > > where Vsrc and Veth are the speeds of the source and > ether with respect to the observer. In Bradley's > stellar aberraton the third term is zero. In > Shtyrkov's aberration from a geostationary satellie > the second term is zero. > > > Similar mistakes in attempts to measure the sup- > > posed ether have been made by others: > > > > http://www.newtonphysics.on.ca/illusion/index.html > > http://www.ccsenet.org/journal/index.php/apr/article/view/10410 > > http://www.k1man.com/f61.pdf > > > > These, of course, go against a mountain of evi- > > dence that there is no such ether and that the > > speed of light is not variable. Here's a paper > > that refutes the last one above: > > > > http://arxiv.org/pdf/1002.3968v1.pdf > > I will comment on these articles when/if I shall > have read and understood them. > > > > > That the "measured" motion only varied by 1% > > > > is very strange because that means any sup- > > > > posed ether whould be virtually stationary wrt > > > > the sun, yet the sun is orbiting the galaxy at > > > > 200 km/sec. The whole thing smells really bad. > > > > > > No, it is simply that the galactic component of > > > the ether wind (declinaton of 89.5 degrees) is > > > nearly coincidental with the polar axis and has > > > therefore but a tiny effect on diurnal aberra- > > > tion from the satellite. > > > > > > To explain -- the diurnal variations of the > > > ether wind in the plane orthogonal to the > > > source-observer line are caused mainly by the > > > rotation of the Earth as it changes orientation > > > with respect to its orbital speed, while the > > > galactic component, having a near-zero projec- > > > tion upon the equatorial plane, causes only a > > > negligible variation. > > > > Well, you can see the galactic plain (i.e., the > > Milky Way). I don't think the normal to that goes > > anywhere near Polaris. Consequently, the compo- > > nent of the sun's velocity should be quite large, > > shouldn't it? > > I think you misunderstood me--I meant the Celestial > poles in the Equatorial coordinate system. The "celestial pole" points to Polaris. Anyway, I don't see anything in Shtyrkov's paper to justify throwing out the mountain of evidence against it: "Extraordinary claims require extraordinary evidence." - Marcello Truzzi And Shtyrkov's paper does not provide it.
[toc] | [prev] | [next] | [standalone]
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2016-01-02 22:05 +0300 |
| Message-ID | <20160102220504.bca84019cedb65cdc000790f@gmail.com> |
| In reply to | #372823 |
Gary Harnagel to Anton Shepelev:
> > Gary Harnagel:
> >
> > > The effect of temperature upon the instruments
> > > comes to mind.
> >
> > It is possible, but without knowldedge of the
> > instuments employed by Shtyrkov, we can only
> > discuss it as a black box, i.e. analysing the
> > results observed, which are:
>
> But we can't even do that without error bars.
We can estimate the random error by fitting a sine
and calculating the dispersion of the mesurements
with respect to it.
> Raw data would be more satisfactory.
I am sure you would appreciate it. If I get it I
will post it, with the author's permission.
> Do you know how many are required to reduce the
> error from 10% to 1?
A hundred. Remember that he tracked the satellite
for several years and used several days' data to de-
termine the average azimuth lead angle at every se-
lected date in every year.
> >
> > 1. The observed azimuth oscillates a little
> > ahead of the reference curve, as would be
> > expected if there were an ether (eq. 7).
> >
> > 2. The average orbital velocity of the ether
> > wind calculated from the observed aberra-
> > tion of the azimuth may be explaned by ex-
> > actly such aberration due to ether wind as
> > would be expected to arise from the or-
> > bital motion of the Earth.
> >
> > 3. Both seasonal and diurnal behaviours of
> > the observed aberration correspond to the
> > prediction (eq. 8) derived form the as-
> > sumption of the ether. Even the small
> > variation of the Earth's orbital speed due
> > to its slightly elliptic orbit was re-
> > solved.
> >
> > To ascribe it all to a systematic error seems
> > far-fetched.
>
> Systematic error was what caused Dayton Miller to
> proclaim an ether wind.
Are you so convinced of it because of Tom Roberts's
paper? Well, he didn't publish the raw data either
and commited a serious overlook in section III where
he writes Miller combined the half-periods, which he
did not. The section's concusion is therefore base-
less:
And in the final plot the frequency bin in which
the real signal would appear is bin 1, the lowest
nonzero frequency bin. From the nature of
Fourier analysis, it should be clear that after
this analysis any initial data with a falling
spectrum (e.g. 1/f noise) will have a spectrum
similar to Fig. 9, and a time domain appearance
similar to the plot at the bottom of Fig. 1.
This is a simple consequence of the fact that the
1/2-turn Fourier component is the lowest fre-
quency retained by the algorithm, and it will
dominate because of the falling spectrum.
Not that it renders the whole paper wrong, though.
I still have suspicions to test, when time permits.
> > Shtyrkov says nothing about the magnitude of his
> > signal with respect to the sigma.
>
> Indeed, thus the paper is unscientific claptrap.
Then do you consider claptrap any reports from pre-
error-analysis days?
Maybe he didn't deem it necessary as long as the
measurements so clearly correspond with the expected
result regarding the orbital motion of the Earth.
> > I have read Roberts's paper and disagree with
> > some of his points, about which I have written
> > in this very thread, in the subthread called
> > "Dayton Miller's measurements". I have not had
> > time to answer all of his counter-arguments,
> > though.
>
> Apparently, you don't understand statistical anal-
> ysis.
I may be wrong and may be deluded, but to say that
anyone disagreeing with a certain report is wrong
and does not understand statistics, without pointing
out at least one specific error by way of example,
seems prejudiced.
> > What do you mean by over-massaging? Processing
> > with the assumption of a periodic effect as
> > would make this effect "appear" in the result,
> > as Roberts claims Miller did? I see no such er-
> > ror in Shtyrkov's work, although fitting the
> > predicted curve and estimating the error would
> > have been better.
>
> Just how did they reduce the error to 1% from 10%?
> How many measurements did they make per data
> point?
Each data point consists of several sessions of
1-3-day tracking, each taken at a certain epoch in
every year. Here is a typical session:
http://oi64.tinypic.com/k532i9.jpg
The orbital speed calculated from the data turned to
be within 1% of the known astronomical value.
> > Well, standard error of the mean is lower than
> > that of a single value by the square root of the
> > number of measurements as long the values are
> > statictically independent.
>
> Were they? Who knows?
>
> > They need not be normally distributed for that
> > to hold.
>
> Again, how many measurements did they make for
> each point on the graph? Was there a trend while
> they were measuring or were results purely random?
Does the above graph help?
> > Normal distibution seems likeliest, describing
> > the sum of many small additive values.
>
> I have no idea what you're talking about here.
About the central limit therem.
> > > Well, stellar aberration is caused by a varia-
> > > tion in relative velocity between earth and
> > > star.
> >
> > Exactly, and even if space were filled with
> > ether that would not change, for the (as-
> > sumed-)constant ether wind across the radius
> > vector would effect a *constant* bending of
> > light-rays without variation. Bradley's stellar
> > aberration is therefore the same for empty space
> > and for ether.
>
> I don't believe so:
>
> https://en.wikipedia.org/wiki/Aberration_of_light#Luminiferous_aether
Yes, I was wrong about the effect of ether drag upon
aberraion. If photons light completely and instan-
teously entrained by the ether there would be no
aberration observable from the Earth.
One proposition is that light is qualitatively dif-
ferent from electro-magnetic waves and that tran-
verse ether flow exerts a very small force upon
light, so that large distances are required. Not so
with electro-magnetic waves.
There are at least two theories in which light is
not simly an electromagnetic wave. How deep should
light penetrate into sea water according to Maxwell
equations, and how deep does it in reality?
> > > > > That the "measured" motion only varied by
> > > > > 1% is very strange because that means any
> > > > > supposed ether whould be virtually sta-
> > > > > tionary wrt the sun, yet the sun is orbit-
> > > > > ing the galaxy at 200 km/sec. The whole
> > > > > thing smells really bad.
> > > >
> > > > No, it is simply that the galactic component
> > > > of the ether wind (declinaton of 89.5 de-
> > > > grees) is nearly coincidental with the polar
> > > > axis and has therefore but a tiny effect on
> > > > diurnal aberration from the satellite.
> > > >
> > > > To explain -- the diurnal variations of the
> > > > ether wind in the plane orthogonal to the
> > > > source-observer line are caused mainly by
> > > > the rotation of the Earth as it changes ori-
> > > > entation with respect to its orbital speed,
> > > > while the galactic component, having a near-
> > > > zero projection upon the equatorial plane,
> > > > causes only a negligible variation.
> > >
> > > Well, you can see the galactic plain (i.e.,
> > > the Milky Way). I don't think the normal to
> > > that goes anywhere near Polaris. Consequent-
> > > ly, the component of the sun's velocity should
> > > be quite large, shouldn't it?
> >
> > I think you misunderstood me--I meant the Celes-
> > tial poles in the Equatorial coordinate system.
>
> The "celestial pole" points to Polaris.
Ah, ok. The Sun is orbiting the galaxy, the galaxy
too is moving somewhither. What is the resultant
motion and is it to be considered absolute?
> Anyway, I don't see anything in Shtyrkov's paper
> to justify throwing out the mountain of evidence
> against it:
>
> "Extraordinary claims require extraordinary evidence."
> --Marcello Truzzi
>
> And Shtyrkov's paper does not provide it.
Can't dismiss a paper on such a basis.
--
() ascii ribbon campaign - against html e-mail
/\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [prev] | [next] | [standalone]
| From | Gary Harnagel <hitlong@yahoo.com> |
|---|---|
| Date | 2016-01-02 16:24 -0800 |
| Message-ID | <d5bd5c4b-5810-47a7-a1ef-f19aaea9385a@googlegroups.com> |
| In reply to | #373136 |
On Saturday, January 2, 2016 at 12:04:20 PM UTC-7, Anton Shepelev wrote: > > Gary Harnagel to Anton Shepelev: > > > > > Gary Harnagel: > > > > > > > The effect of temperature upon the instruments > > > > comes to mind. > > > > > > It is possible, but without knowldedge of the > > > instuments employed by Shtyrkov, we can only > > > discuss it as a black box, i.e. analysing the > > > results observed, which are: > > > > But we can't even do that without error bars. > > We can estimate the random error by fitting a sine > and calculating the dispersion of the mesurements > with respect to it. Dayton Miller did that and got fooled. "The first principle is that you must not fool yourself and you are the easiest person to fool." -- Richard P. Feynman > > Raw data would be more satisfactory. > > I am sure you would appreciate it. If I get it I > will post it, with the author's permission. Maybe I'm just a skeptic, but I doubt you'll ever get it :-) > > Do you know how many are required to reduce the > > error from 10% to 1? > > A hundred. Remember that he tracked the satellite > for several years and used several days' data to de- > termine the average azimuth lead angle at every se- > lected date in every year. Yes, but only if the data is truly random. It is not so in the case of systematic error or other kind of bias. > > > 1. The observed azimuth oscillates a little > > > ahead of the reference curve, as would be > > > expected if there were an ether (eq. 7). > > > > > > 2. The average orbital velocity of the ether > > > wind calculated from the observed aberra- > > > tion of the azimuth may be explaned by ex- > > > actly such aberration due to ether wind as > > > would be expected to arise from the or- > > > bital motion of the Earth. > > > > > > 3. Both seasonal and diurnal behaviours of > > > the observed aberration correspond to the > > > prediction (eq. 8) derived form the as- > > > sumption of the ether. Even the small > > > variation of the Earth's orbital speed due > > > to its slightly elliptic orbit was re- > > > solved. > > > > > > To ascribe it all to a systematic error seems > > > far-fetched. > > > > Systematic error was what caused Dayton Miller to > > proclaim an ether wind. > > Are you so convinced of it because of Tom Roberts's > paper? Well, he didn't publish the raw data either Hmmm, I don't see any justification for that since that data is available: http://arxiv.org/ftp/physics/papers/0608/0608238.pdf "Acknowledgements "Thanks to the Case Western Reserve University Archives for making copies of Miller's data sheets available." > and commited a serious overlook in section III where > he writes Miller combined the half-periods, which he > did not. The section's concusion is therefore base- > less: > > And in the final plot the frequency bin in which > the real signal would appear is bin 1, the lowest > nonzero frequency bin. From the nature of > Fourier analysis, it should be clear that after > this analysis any initial data with a falling > spectrum (e.g. 1/f noise) will have a spectrum > similar to Fig. 9, and a time domain appearance > similar to the plot at the bottom of Fig. 1. > This is a simple consequence of the fact that the > 1/2-turn Fourier component is the lowest fre- > quency retained by the algorithm, and it will > dominate because of the falling spectrum. > > Not that it renders the whole paper wrong, though. > I still have suspicions to test, when time permits. Ummm, I trust Roberts to have done the right thing. He is very knowledgeable about such things. He HAS to be. You would have to spell it out to me in detail why what he did was wrong. > > > Shtyrkov says nothing about the magnitude of his > > > signal with respect to the sigma. > > > > Indeed, thus the paper is unscientific claptrap. > > Then do you consider claptrap any reports from pre- > error-analysis days? That depends upon the noise sources. Miller had much more noise than M&M. > Maybe he didn't deem it necessary as long as the > measurements so clearly correspond with the expected > result regarding the orbital motion of the Earth. No one is condemning Miller. He used what was available to him. > > > I have read Roberts's paper and disagree with > > > some of his points, about which I have written > > > in this very thread, in the subthread called > > > "Dayton Miller's measurements". I have not had > > > time to answer all of his counter-arguments, > > > though. > > > > Apparently, you don't understand statistical anal- > > ysis. > > I may be wrong and may be deluded, but to say that > anyone disagreeing with a certain report is wrong > and does not understand statistics, without pointing > out at least one specific error by way of example, > seems prejudiced. Huh? All I've seen you post is a disagreement about 1/2-turn statistics with no valid reason for doing so. > > > What do you mean by over-massaging? Processing > > > with the assumption of a periodic effect as > > > would make this effect "appear" in the result, > > > as Roberts claims Miller did? I see no such er- > > > ror in Shtyrkov's work, although fitting the > > > predicted curve and estimating the error would > > > have been better. > > > > Just how did they reduce the error to 1% from 10%? > > How many measurements did they make per data > > point? > > Each data point consists of several sessions of > 1-3-day tracking, each taken at a certain epoch in > every year. Here is a typical session: > > http://oi64.tinypic.com/k532i9.jpg > > The orbital speed calculated from the data turned to > be within 1% of the known astronomical value. Yes, and they averaged all the measurements in that session to ONE data point in Figure 4. BTW, do you see the equation they used to get v from the azimuth angle? > > > Well, standard error of the mean is lower than > > > that of a single value by the square root of the > > > number of measurements as long the values are > > > statictically independent. > > > > Were they? Who knows? > > > > > They need not be normally distributed for that > > > to hold. > > > > Again, how many measurements did they make for > > each point on the graph? Was there a trend while > > they were measuring or were results purely random? > > Does the above graph help? Yes, but I need the equation for v to determine what the error bars are in Figure 4. rom that figure, it looks like they took two sets of sequences in many cases, and those two sets differed by about 1 km/sec, which is greater than the variation they're claiming to have detected. > > > Normal distibution seems likeliest, describing > > > the sum of many small additive values. > > > > I have no idea what you're talking about here. > > About the central limit therem. > > > > > Well, stellar aberration is caused by a varia- > > > > tion in relative velocity between earth and > > > > star. > > > > > > Exactly, and even if space were filled with > > > ether that would not change, for the (as- > > > sumed-)constant ether wind across the radius > > > vector would effect a *constant* bending of > > > light-rays without variation. Bradley's stellar > > > aberration is therefore the same for empty space > > > and for ether. > > > > I don't believe so: > > > > https://en.wikipedia.org/wiki/Aberration_of_light#Luminiferous_aether > > Yes, I was wrong about the effect of ether drag upon > aberraion. If photons light completely and instan- > teously entrained by the ether there would be no > aberration observable from the Earth. > > One proposition is that light is qualitatively dif- > ferent from electro-magnetic waves and that tran- > verse ether flow exerts a very small force upon > light, so that large distances are required. Not so > with electro-magnetic waves. > > There are at least two theories in which light is > not simly an electromagnetic wave. How deep should > light penetrate into sea water according to Maxwell > equations, and how deep does it in reality? It's wavelength dependent and is caused by absorption. > > > I think you misunderstood me--I meant the Celes- > > > tial poles in the Equatorial coordinate system. > > > > The "celestial pole" points to Polaris. > > Ah, ok. The Sun is orbiting the galaxy, the galaxy > too is moving somewhither. What is the resultant > motion and is it to be considered absolute? Motion is relative, it is observer dependent. > > Anyway, I don't see anything in Shtyrkov's paper > > to justify throwing out the mountain of evidence > > against it: > > > > "Extraordinary claims require extraordinary evidence." > > --Marcello Truzzi > > > > And Shtyrkov's paper does not provide it. > > Can't dismiss a paper on such a basis. > > -- > () ascii ribbon campaign - against html e-mail > /\ http://preview.tinyurl.com/qcy6mjc [archived] Yes, I can. The "extraordinary evidence" does not exist.
[toc] | [prev] | [next] | [standalone]
| From | Tom Roberts <tjroberts137@sbcglobal.net> |
|---|---|
| Date | 2016-01-03 12:57 -0600 |
| Message-ID | <Pvmdnakpx_2F8hTLnZ2dnUU7_82dnZ2d@giganews.com> |
| In reply to | #373136 |
On 1/2/16 1/2/16 1:05 PM, Anton Shepelev wrote: > We can estimate the random error by fitting a sine > and calculating the dispersion of the mesurements > with respect to it. Yes, if you do this with the RAW DATA. Doing it to the averages given in his paper is meaningless, BECAUSE THE VARIATIONS THAT CONTRIBUTE TO THE STATISTICAL ERRORBAR HAVE ALREADY BEEN REMOVED. > A hundred. Remember that he tracked the satellite > for several years and used several days' data to de- > termine the average azimuth lead angle at every se- > lected date in every year. Yes. So averaging measurements a year apart is invalid. To get reliable results one MUST fit a theoretical function of time to the raw data. > Are you so convinced of it because of Tom Roberts's > paper? Well, he didn't publish the raw data either I referenced where I got it, and how others can get it (the CWRU archives). And like Miller I displayed the raw data from one run. > and commited a serious overlook in section III where > he writes Miller combined the half-periods, which he > did not. You SERIOUSLY overstate your case. I'll discuss your claim in a response to your original post on this. Bottom line: a half-rotation is what Miller plotted and what people look at and claim "sinusoid"; omitting that last averaging does not significantly affect my conclusion -- the variations in the data remain MUCH smaller than the statistical errorbars. > [Shtyrkov and error analysis] > Maybe he didn't deem it necessary as long as the > measurements so clearly correspond with the expected > result regarding the orbital motion of the Earth. THAT IS NONSENSE! Today there is no excuse for omitting an error analysis. And one CANNOT "deduce" small errorbars because the data match some expected pattern -- that is imposing your pattern onto the data, which is not only invalid, but also highly improper. > Yes, I was wrong about the effect of ether drag upon > aberraion. If photons light completely and instan- > teously entrained by the ether there would be no > aberration observable from the Earth. OK. Recognizing that, apply it to Shtyrkov's entire approach. He is being grossly self-inconsistent. (Light being "completely and instantaneously entrained by the ether" is what one means by "ether theory".) But in ANY model, a geosynchronous satellite would have no annually-varying or diurnally-varying aberration, because it is essentially at rest in the lab doing the observing. > One proposition is that light is qualitatively dif- > ferent from electro-magnetic waves Hopeless. There are zillions of experiments throughout the electromagnetic spectrum, including light; one model of electrodynamics handles them all (QED). Several experimenters have used comb filters to make highly-accurate measurements of the frequencies of light from various lasers; they started from microwaves whose frequency can be directly counted (and which are indisputably EM waves). > The Sun is orbiting the galaxy, the galaxy > too is moving somewhither. What is the resultant > motion and is it to be considered absolute? That's merely one aspect of the difficulties surrounding "absolute motion". Tom Roberts
[toc] | [prev] | [next] | [standalone]
| From | John Gogo <jfgogo22@yahoo.com> |
|---|---|
| Date | 2016-01-03 13:26 -0800 |
| Message-ID | <75a3729f-80ee-4d35-b4e6-16ec7160150b@googlegroups.com> |
| In reply to | #373208 |
On Sunday, January 3, 2016 at 12:57:30 PM UTC-6, tjrob137 wrote: > On 1/2/16 1/2/16 1:05 PM, Anton Shepelev wrote: > > We can estimate the random error by fitting a sine > > and calculating the dispersion of the mesurements > > with respect to it. > > Yes, if you do this with the RAW DATA. Doing it to the averages given in his > paper is meaningless, BECAUSE THE VARIATIONS THAT CONTRIBUTE TO THE STATISTICAL > ERRORBAR HAVE ALREADY BEEN REMOVED. > > > > A hundred. Remember that he tracked the satellite > > for several years and used several days' data to de- > > termine the average azimuth lead angle at every se- > > lected date in every year. > > Yes. So averaging measurements a year apart is invalid. To get reliable results > one MUST fit a theoretical function of time to the raw data. > > > > Are you so convinced of it because of Tom Roberts's > > paper? Well, he didn't publish the raw data either > > I referenced where I got it, and how others can get it (the CWRU archives). And > like Miller I displayed the raw data from one run. > > > > and commited a serious overlook in section III where > > he writes Miller combined the half-periods, which he > > did not. > > You SERIOUSLY overstate your case. I'll discuss your claim in a response to your > original post on this. Bottom line: a half-rotation is what Miller plotted and > what people look at and claim "sinusoid"; omitting that last averaging does not > significantly affect my conclusion -- the variations in the data remain MUCH > smaller than the statistical errorbars. > > > > [Shtyrkov and error analysis] > > Maybe he didn't deem it necessary as long as the > > measurements so clearly correspond with the expected > > result regarding the orbital motion of the Earth. > > THAT IS NONSENSE! Today there is no excuse for omitting an error analysis. And > one CANNOT "deduce" small errorbars because the data match some expected pattern > -- that is imposing your pattern onto the data, which is not only invalid, but > also highly improper. > > > > Yes, I was wrong about the effect of ether drag upon > > aberraion. If photons light completely and instan- > > teously entrained by the ether there would be no > > aberration observable from the Earth. > > OK. Recognizing that, apply it to Shtyrkov's entire approach. He is being > grossly self-inconsistent. > > (Light being "completely and instantaneously entrained by > the ether" is what one means by "ether theory".) > > But in ANY model, a geosynchronous satellite would have no annually-varying or > diurnally-varying aberration, because it is essentially at rest in the lab doing > the observing. > > > > One proposition is that light is qualitatively dif- > > ferent from electro-magnetic waves > > Hopeless. There are zillions of experiments throughout the electromagnetic > spectrum, including light; one model of electrodynamics handles them all (QED). > Several experimenters have used comb filters to make highly-accurate > measurements of the frequencies of light from various lasers; they started from > microwaves whose frequency can be directly counted (and which are indisputably > EM waves). > > > > The Sun is orbiting the galaxy, the galaxy > > too is moving somewhither. What is the resultant > > motion and is it to be considered absolute? > > That's merely one aspect of the difficulties surrounding "absolute motion". > > > Tom Roberts Perpetual motion- motion that is observable- but has no beginning or end.
[toc] | [prev] | [next] | [standalone]
| From | John Gogo <jfgogo22@yahoo.com> |
|---|---|
| Date | 2016-01-03 13:13 -0800 |
| Message-ID | <07864b2d-c686-46bb-865d-d0c4c9cb440d@googlegroups.com> |
| In reply to | #373136 |
On Saturday, January 2, 2016 at 1:04:20 PM UTC-6, Anton Shepelev wrote: > Gary Harnagel to Anton Shepelev: > > > > Gary Harnagel: > > > > > > > The effect of temperature upon the instruments > > > > comes to mind. > > > > > > It is possible, but without knowldedge of the > > > instuments employed by Shtyrkov, we can only > > > discuss it as a black box, i.e. analysing the > > > results observed, which are: > > > > But we can't even do that without error bars. > > We can estimate the random error by fitting a sine > and calculating the dispersion of the mesurements > with respect to it. > > > Raw data would be more satisfactory. > > I am sure you would appreciate it. If I get it I > will post it, with the author's permission. > > > Do you know how many are required to reduce the > > error from 10% to 1? > > A hundred. Remember that he tracked the satellite > for several years and used several days' data to de- > termine the average azimuth lead angle at every se- > lected date in every year. > > > > > > > 1. The observed azimuth oscillates a little > > > ahead of the reference curve, as would be > > > expected if there were an ether (eq. 7). > > > > > > 2. The average orbital velocity of the ether > > > wind calculated from the observed aberra- > > > tion of the azimuth may be explaned by ex- > > > actly such aberration due to ether wind as > > > would be expected to arise from the or- > > > bital motion of the Earth. > > > > > > 3. Both seasonal and diurnal behaviours of > > > the observed aberration correspond to the > > > prediction (eq. 8) derived form the as- > > > sumption of the ether. Even the small > > > variation of the Earth's orbital speed due > > > to its slightly elliptic orbit was re- > > > solved. > > > > > > To ascribe it all to a systematic error seems > > > far-fetched. > > > > Systematic error was what caused Dayton Miller to > > proclaim an ether wind. > > Are you so convinced of it because of Tom Roberts's > paper? Well, he didn't publish the raw data either > and commited a serious overlook in section III where > he writes Miller combined the half-periods, which he > did not. The section's concusion is therefore base- > less: > > And in the final plot the frequency bin in which > the real signal would appear is bin 1, the lowest > nonzero frequency bin. From the nature of > Fourier analysis, it should be clear that after > this analysis any initial data with a falling > spectrum (e.g. 1/f noise) will have a spectrum > similar to Fig. 9, and a time domain appearance > similar to the plot at the bottom of Fig. 1. > This is a simple consequence of the fact that the > 1/2-turn Fourier component is the lowest fre- > quency retained by the algorithm, and it will > dominate because of the falling spectrum. > > Not that it renders the whole paper wrong, though. > I still have suspicions to test, when time permits. > > > > Shtyrkov says nothing about the magnitude of his > > > signal with respect to the sigma. > > > > Indeed, thus the paper is unscientific claptrap. > > Then do you consider claptrap any reports from pre- > error-analysis days? > > Maybe he didn't deem it necessary as long as the > measurements so clearly correspond with the expected > result regarding the orbital motion of the Earth. > > > > I have read Roberts's paper and disagree with > > > some of his points, about which I have written > > > in this very thread, in the subthread called > > > "Dayton Miller's measurements". I have not had > > > time to answer all of his counter-arguments, > > > though. > > > > Apparently, you don't understand statistical anal- > > ysis. > > I may be wrong and may be deluded, but to say that > anyone disagreeing with a certain report is wrong > and does not understand statistics, without pointing > out at least one specific error by way of example, > seems prejudiced. > > > > What do you mean by over-massaging? Processing > > > with the assumption of a periodic effect as > > > would make this effect "appear" in the result, > > > as Roberts claims Miller did? I see no such er- > > > ror in Shtyrkov's work, although fitting the > > > predicted curve and estimating the error would > > > have been better. > > > > Just how did they reduce the error to 1% from 10%? > > How many measurements did they make per data > > point? > > Each data point consists of several sessions of > 1-3-day tracking, each taken at a certain epoch in > every year. Here is a typical session: > > http://oi64.tinypic.com/k532i9.jpg > > The orbital speed calculated from the data turned to > be within 1% of the known astronomical value. > > > > Well, standard error of the mean is lower than > > > that of a single value by the square root of the > > > number of measurements as long the values are > > > statictically independent. > > > > Were they? Who knows? > > > > > They need not be normally distributed for that > > > to hold. > > > > Again, how many measurements did they make for > > each point on the graph? Was there a trend while > > they were measuring or were results purely random? > > Does the above graph help? > > > > Normal distibution seems likeliest, describing > > > the sum of many small additive values. > > > > I have no idea what you're talking about here. > > About the central limit therem. > > > > > Well, stellar aberration is caused by a varia- > > > > tion in relative velocity between earth and > > > > star. > > > > > > Exactly, and even if space were filled with > > > ether that would not change, for the (as- > > > sumed-)constant ether wind across the radius > > > vector would effect a *constant* bending of > > > light-rays without variation. Bradley's stellar > > > aberration is therefore the same for empty space > > > and for ether. > > > > I don't believe so: > > > > https://en.wikipedia.org/wiki/Aberration_of_light#Luminiferous_aether > > Yes, I was wrong about the effect of ether drag upon > aberraion. If photons light completely and instan- > teously entrained by the ether there would be no > aberration observable from the Earth. > > One proposition is that light is qualitatively dif- > ferent from electro-magnetic waves and that tran- > verse ether flow exerts a very small force upon > light, so that large distances are required. Not so > with electro-magnetic waves. > > There are at least two theories in which light is > not simly an electromagnetic wave. How deep should > light penetrate into sea water according to Maxwell > equations, and how deep does it in reality? > > > > > > > That the "measured" motion only varied by > > > > > > 1% is very strange because that means any > > > > > > supposed ether whould be virtually sta- > > > > > > tionary wrt the sun, yet the sun is orbit- > > > > > > ing the galaxy at 200 km/sec. The whole > > > > > > thing smells really bad. > > > > > > > > > > No, it is simply that the galactic component > > > > > of the ether wind (declinaton of 89.5 de- > > > > > grees) is nearly coincidental with the polar > > > > > axis and has therefore but a tiny effect on > > > > > diurnal aberration from the satellite. > > > > > > > > > > To explain -- the diurnal variations of the > > > > > ether wind in the plane orthogonal to the > > > > > source-observer line are caused mainly by > > > > > the rotation of the Earth as it changes ori- > > > > > entation with respect to its orbital speed, > > > > > while the galactic component, having a near- > > > > > zero projection upon the equatorial plane, > > > > > causes only a negligible variation. > > > > > > > > Well, you can see the galactic plain (i.e., > > > > the Milky Way). I don't think the normal to > > > > that goes anywhere near Polaris. Consequent- > > > > ly, the component of the sun's velocity should > > > > be quite large, shouldn't it? > > > > > > I think you misunderstood me--I meant the Celes- > > > tial poles in the Equatorial coordinate system. > > > > The "celestial pole" points to Polaris. > > Ah, ok. The Sun is orbiting the galaxy, the galaxy > too is moving somewhither. What is the resultant > motion and is it to be considered absolute? > > > Anyway, I don't see anything in Shtyrkov's paper > > to justify throwing out the mountain of evidence > > against it: > > > > "Extraordinary claims require extraordinary evidence." > > --Marcello Truzzi > > > > And Shtyrkov's paper does not provide it. > > Can't dismiss a paper on such a basis. > > -- > () ascii ribbon campaign - against html e-mail > /\ http://preview.tinyurl.com/qcy6mjc [archived] There has to be a point where the wheel and man's ability to measure far or small runs out.
[toc] | [prev] | [next] | [standalone]
| From | John Gogo <jfgogo22@yahoo.com> |
|---|---|
| Date | 2016-01-03 13:21 -0800 |
| Message-ID | <e0d80211-4dc8-43d8-b12e-7850fa205d9d@googlegroups.com> |
| In reply to | #373231 |
On Sunday, January 3, 2016 at 3:13:04 PM UTC-6, John Gogo wrote: > On Saturday, January 2, 2016 at 1:04:20 PM UTC-6, Anton Shepelev wrote: > > Gary Harnagel to Anton Shepelev: > > > > > > Gary Harnagel: > > > > > > > > > The effect of temperature upon the instruments > > > > > comes to mind. > > > > > > > > It is possible, but without knowldedge of the > > > > instuments employed by Shtyrkov, we can only > > > > discuss it as a black box, i.e. analysing the > > > > results observed, which are: > > > > > > But we can't even do that without error bars. > > > > We can estimate the random error by fitting a sine > > and calculating the dispersion of the mesurements > > with respect to it. > > > > > Raw data would be more satisfactory. > > > > I am sure you would appreciate it. If I get it I > > will post it, with the author's permission. > > > > > Do you know how many are required to reduce the > > > error from 10% to 1? > > > > A hundred. Remember that he tracked the satellite > > for several years and used several days' data to de- > > termine the average azimuth lead angle at every se- > > lected date in every year. > > > > > > > > > > 1. The observed azimuth oscillates a little > > > > ahead of the reference curve, as would be > > > > expected if there were an ether (eq. 7). > > > > > > > > 2. The average orbital velocity of the ether > > > > wind calculated from the observed aberra- > > > > tion of the azimuth may be explaned by ex- > > > > actly such aberration due to ether wind as > > > > would be expected to arise from the or- > > > > bital motion of the Earth. > > > > > > > > 3. Both seasonal and diurnal behaviours of > > > > the observed aberration correspond to the > > > > prediction (eq. 8) derived form the as- > > > > sumption of the ether. Even the small > > > > variation of the Earth's orbital speed due > > > > to its slightly elliptic orbit was re- > > > > solved. > > > > > > > > To ascribe it all to a systematic error seems > > > > far-fetched. > > > > > > Systematic error was what caused Dayton Miller to > > > proclaim an ether wind. > > > > Are you so convinced of it because of Tom Roberts's > > paper? Well, he didn't publish the raw data either > > and commited a serious overlook in section III where > > he writes Miller combined the half-periods, which he > > did not. The section's concusion is therefore base- > > less: > > > > And in the final plot the frequency bin in which > > the real signal would appear is bin 1, the lowest > > nonzero frequency bin. From the nature of > > Fourier analysis, it should be clear that after > > this analysis any initial data with a falling > > spectrum (e.g. 1/f noise) will have a spectrum > > similar to Fig. 9, and a time domain appearance > > similar to the plot at the bottom of Fig. 1. > > This is a simple consequence of the fact that the > > 1/2-turn Fourier component is the lowest fre- > > quency retained by the algorithm, and it will > > dominate because of the falling spectrum. > > > > Not that it renders the whole paper wrong, though. > > I still have suspicions to test, when time permits. > > > > > > Shtyrkov says nothing about the magnitude of his > > > > signal with respect to the sigma. > > > > > > Indeed, thus the paper is unscientific claptrap. > > > > Then do you consider claptrap any reports from pre- > > error-analysis days? > > > > Maybe he didn't deem it necessary as long as the > > measurements so clearly correspond with the expected > > result regarding the orbital motion of the Earth. > > > > > > I have read Roberts's paper and disagree with > > > > some of his points, about which I have written > > > > in this very thread, in the subthread called > > > > "Dayton Miller's measurements". I have not had > > > > time to answer all of his counter-arguments, > > > > though. > > > > > > Apparently, you don't understand statistical anal- > > > ysis. > > > > I may be wrong and may be deluded, but to say that > > anyone disagreeing with a certain report is wrong > > and does not understand statistics, without pointing > > out at least one specific error by way of example, > > seems prejudiced. > > > > > > What do you mean by over-massaging? Processing > > > > with the assumption of a periodic effect as > > > > would make this effect "appear" in the result, > > > > as Roberts claims Miller did? I see no such er- > > > > ror in Shtyrkov's work, although fitting the > > > > predicted curve and estimating the error would > > > > have been better. > > > > > > Just how did they reduce the error to 1% from 10%? > > > How many measurements did they make per data > > > point? > > > > Each data point consists of several sessions of > > 1-3-day tracking, each taken at a certain epoch in > > every year. Here is a typical session: > > > > http://oi64.tinypic.com/k532i9.jpg > > > > The orbital speed calculated from the data turned to > > be within 1% of the known astronomical value. > > > > > > Well, standard error of the mean is lower than > > > > that of a single value by the square root of the > > > > number of measurements as long the values are > > > > statictically independent. > > > > > > Were they? Who knows? > > > > > > > They need not be normally distributed for that > > > > to hold. > > > > > > Again, how many measurements did they make for > > > each point on the graph? Was there a trend while > > > they were measuring or were results purely random? > > > > Does the above graph help? > > > > > > Normal distibution seems likeliest, describing > > > > the sum of many small additive values. > > > > > > I have no idea what you're talking about here. > > > > About the central limit therem. > > > > > > > Well, stellar aberration is caused by a varia- > > > > > tion in relative velocity between earth and > > > > > star. > > > > > > > > Exactly, and even if space were filled with > > > > ether that would not change, for the (as- > > > > sumed-)constant ether wind across the radius > > > > vector would effect a *constant* bending of > > > > light-rays without variation. Bradley's stellar > > > > aberration is therefore the same for empty space > > > > and for ether. > > > > > > I don't believe so: > > > > > > https://en.wikipedia.org/wiki/Aberration_of_light#Luminiferous_aether > > > > Yes, I was wrong about the effect of ether drag upon > > aberraion. If photons light completely and instan- > > teously entrained by the ether there would be no > > aberration observable from the Earth. > > > > One proposition is that light is qualitatively dif- > > ferent from electro-magnetic waves and that tran- > > verse ether flow exerts a very small force upon > > light, so that large distances are required. Not so > > with electro-magnetic waves. > > > > There are at least two theories in which light is > > not simly an electromagnetic wave. How deep should > > light penetrate into sea water according to Maxwell > > equations, and how deep does it in reality? > > > > > > > > > That the "measured" motion only varied by > > > > > > > 1% is very strange because that means any > > > > > > > supposed ether whould be virtually sta- > > > > > > > tionary wrt the sun, yet the sun is orbit- > > > > > > > ing the galaxy at 200 km/sec. The whole > > > > > > > thing smells really bad. > > > > > > > > > > > > No, it is simply that the galactic component > > > > > > of the ether wind (declinaton of 89.5 de- > > > > > > grees) is nearly coincidental with the polar > > > > > > axis and has therefore but a tiny effect on > > > > > > diurnal aberration from the satellite. > > > > > > > > > > > > To explain -- the diurnal variations of the > > > > > > ether wind in the plane orthogonal to the > > > > > > source-observer line are caused mainly by > > > > > > the rotation of the Earth as it changes ori- > > > > > > entation with respect to its orbital speed, > > > > > > while the galactic component, having a near- > > > > > > zero projection upon the equatorial plane, > > > > > > causes only a negligible variation. > > > > > > > > > > Well, you can see the galactic plain (i.e., > > > > > the Milky Way). I don't think the normal to > > > > > that goes anywhere near Polaris. Consequent- > > > > > ly, the component of the sun's velocity should > > > > > be quite large, shouldn't it? > > > > > > > > I think you misunderstood me--I meant the Celes- > > > > tial poles in the Equatorial coordinate system. > > > > > > The "celestial pole" points to Polaris. > > > > Ah, ok. The Sun is orbiting the galaxy, the galaxy > > too is moving somewhither. What is the resultant > > motion and is it to be considered absolute? > > > > > Anyway, I don't see anything in Shtyrkov's paper > > > to justify throwing out the mountain of evidence > > > against it: > > > > > > "Extraordinary claims require extraordinary evidence." > > > --Marcello Truzzi > > > > > > And Shtyrkov's paper does not provide it. > > > > Can't dismiss a paper on such a basis. > > > > -- > > () ascii ribbon campaign - against html e-mail > > /\ http://preview.tinyurl.com/qcy6mjc [archived] > > There has to be a point where the wheel and man's ability to measure far or small runs out. Mankind's UNIFYING ability lies in the capability to put FAR and SMALL together; whether, they share a FUNDAMENTAL theory or should be SEPARATED.
[toc] | [prev] | [next] | [standalone]
| From | Tom Roberts <tjroberts137@sbcglobal.net> |
|---|---|
| Date | 2016-01-03 12:26 -0600 |
| Message-ID | <ALCdnbRAB_gj-hTLnZ2dnUU7_8ydnZ2d@giganews.com> |
| In reply to | #372805 |
On 12/26/15 12/26/15 3:17 PM, Gary Harnagel wrote: > On Saturday, December 26, 2015 at 11:57:38 AM UTC-7, Anton Shepelev wrote: >> Gary Harnagel about >> http://bourabai.kz/shtyrkov/shtyrkov.pdf : >>> but there was no error analysis, which makes the >>> numbers and conclusions meaningless. >> >> Less credible -- sure, but why meaningless? So much less credible that it is tantamount to being meaningless. Without errorbars one cannot determine whether patterns in the data are significant. > Whereas >> he got a sinusoidal signal instead of noise, the >> random error is small enough, This is BACKWARDS -- you CANNOT deduce the errorbars from comparing to a desired pattern like that. As I have said before: Amateurs look for patterns, professionals look at errorbars. You and Shtyrkov brand yourselves as poorly-educated amateurs. > Reducing the error by making many observations is an acceptable procedure. ONLY SOMETIMES. For measuring the length of a static object, averaging can usually be used to reduce the errorbar. But averaging data has long been known to fool experimenters, because the process of averaging can have additional effects. When the data comprise a time-sequence of values, one should fit the theoretical function of time to the data. For Miller's experiment his averaging imposed a comb filter onto the data, with disastrous results (it fooled him into thinking he had a real signal). It is not difficult to fit to the entire time sequence -- see what I did in Part IV of my paper on Miller's results. Granted such a fit would have been beyond Miller, in 1933. But today nobody has such an excuse. > Is that what the little circles on the graph mean? (Here is a case where experience is quite valuable.) I believe he simply chose open circles to plot points in whatever plotting program he used (programs offer many selections, and the user chooses among them to make the points visible). So their size is irrelevant. I doubt very much that they are errorbars (sigma), because a) today nobody uses that for errorbars, and b) he surely would have mentioned that. > Another point is that reducing error by multiple measurements only works > if the data approaches a normal distribution. If you are going to average data, with a sufficient number of raw data points the average will always approach a normal distribution, regardless of the distribution of the raw data, as long as the data points are statistically independent [#]. In practice, five to ten data points is usually sufficient for this to be valid. Moreover the statistical errorbar on the mean can be computed from the sigma of the raw data divided by the square-root of the number of data points being averaged. [#] This is the central limit theorem of statistics. >> Tom Roberts thinks that this value arose from >> Shtyrkov's correction for a non-existent aberration, >> but I did not understand what he meant. He "corrected" all of his data for aberration. But aberration DOES NOT APPLY to a geosynchronous satellite. See my post from yesterday on this. Tom Roberts
[toc] | [prev] | [next] | [standalone]
| From | Kathleen Winder <katwin@mcgatlin.org> |
|---|---|
| Date | 2016-01-03 18:48 +0000 |
| Message-ID | <n6bqd5$c7s$1@gioia.aioe.org> |
| In reply to | #373201 |
Tom Roberts wrote: >>> Gary Harnagel about http://bourabai.kz/shtyrkov/shtyrkov.pdf : >>>> but there was no error analysis, which makes the numbers and >>>> conclusions meaningless. >>> >>> Less credible -- sure, but why meaningless? > > So much less credible that it is tantamount to being meaningless. > Without errorbars one cannot determine whether patterns in the data are > significant. Okay, how small those errorbars need to be making the data significant enough to you? You must know the answer already, you seems to. No data exists without allocated errorbar(s). You make it sound like data cannot be significant since it comes with errorbars in it.
[toc] | [prev] | [next] | [standalone]
| From | Gary Harnagel <hitlong@yahoo.com> |
|---|---|
| Date | 2016-01-04 05:29 -0800 |
| Message-ID | <818c0929-b083-4a23-bd05-1dcd261807b2@googlegroups.com> |
| In reply to | #373201 |
On Sunday, January 3, 2016 at 11:26:09 AM UTC-7, tjrob137 wrote: > > On 12/26/15 12/26/15 3:17 PM, Gary Harnagel wrote: > > > > Reducing the error by making many observations is an acceptable procedure. > > ONLY SOMETIMES. For measuring the length of a static object, averaging can > usually be used to reduce the errorbar. But averaging data has long been > known to fool experimenters, because the process of averaging can have > additional effects. > > When the data comprise a time-sequence of values, one should fit the > theoretical function of time to the data. Isn't that what Miller did? Of course, it would have been really hard for him to have done a complete analysis. I would think that, today, one could choose a series of sinusoidal periods and phases and calculate a least-squares fit to each. Is that a valid technique? > For Miller's experiment his averaging imposed a comb filter onto the data, > with disastrous results (it fooled him into thinking he had a real signal). > It is not difficult to fit to the entire time sequence -- > see what I did in Part IV of my paper on Miller's results. I'll have to look at that more carefully. > Granted such a fit would have been beyond Miller, in 1933. > But today nobody has such an excuse. > > > > Is that what the little circles on the graph mean? > > (Here is a case where experience is quite valuable.) > > I believe he simply chose open circles to plot points in whatever plotting > program he used (programs offer many selections, and the user chooses among > them to make the points visible). So their size is irrelevant. I doubt very > much that they are errorbars (sigma), because a) today nobody uses that for > errorbars, and b) he surely would have mentioned that. I agree. I was just trying to be charitable :-) > > Another point is that reducing error by multiple measurements only works > > if the data approaches a normal distribution. > > If you are going to average data, with a sufficient number of raw data > points the average will always approach a normal distribution, regardless > of the distribution of the raw data, as long as the data points are > statistically independent [#]. Yes, that is what I was trying to say. I was thinking about the case of a systematic error, one that was not changing randomly. > In practice, five to ten data points is usually sufficient for this to > be valid. Moreover the statistical errorbar on the mean can be computed > from the sigma of the raw data divided by the square-root of the number > of data points being averaged. ?? sigma/sqrt(N)? > [#] This is the central limit theorem of statistics. > > > > > Tom Roberts thinks that this value arose from > > > Shtyrkov's correction for a non-existent aberration, > > > but I did not understand what he meant. > > He "corrected" all of his data for aberration. But aberration DOES NOT > APPLY to a geosynchronous satellite. See my post from yesterday on this. > > > Tom Roberts Which explains why he got the "results" he did. Gary
[toc] | [prev] | [next] | [standalone]
| From | Anton Shepelev <anton.txt@gmail.com> |
|---|---|
| Date | 2016-01-05 20:19 +0300 |
| Message-ID | <20160105201932.8b3d5f1257f2df4c2c0b2c98@gmail.com> |
| In reply to | #373267 |
Gary Harnagel to Tom Roberts:
> > When the data comprise a time-sequence of val-
> > ues, one should fit the theoretical function of
> > time to the data.
>
> Isn't that what Miller did?
No. For each data run, usually consisting of twenty
interferometer turns, he averaged the drift-correct-
ed fringe-shifts for each of the sixteen (Tom, not
eight!) azimuths, and plotted the result to a large
scale. These full-period (Tom, not half-period!)
curves he fed to a mechanical harmonic analyser mea-
suring the first, the second, and all the higher
harmonics:
http://oi66.tinypic.com/209q2dl.jpg
For each epoch shown are four graphs: observed full-
period signal, first harmonic, second harmonic, and
residual.
It is the magnitude and phase of the thus-determined
second-order effect that he used in the subsequent
analysis. Hicks showed that the first-order effect
was also to be expected -- because finite fringe
width implies asymmetry -- but Miller was not aware
of it at the time of the experiment and couldn't use
it later because it depends on the frige width,
which was not recorded.
Tom was not aware of it either:
Because of the 180 degree symmetry of the inter-
ferometer, any real signal can depend only on
orientation modulo 180 degrees.
and so in section III he averaged the half-periods,
while Miller never did it.
> Of course, it would have been really hard for him
> to have done a complete analysis.
I find his analysis quite "complete," but, as Tom
says, modern error analysis was not at the time
available.
> Tom Roberts:
> > Anton Shepelev:
> >
> > > Tom Roberts thinks that this value arose from
> > > Shtyrkov's correction for a non-existent aber-
> > > ration, but I did not understand what he
> > > meant.
> >
> > He "corrected" all of his data for aberration.
> > But aberration DOES NOT APPLY to a geosyn-
> > chronous satellite. See my post from yesterday
> > on this.
>
> Which explains why he got the "results" he did.
Do you really agree with Tom? I mean, have you re-
ally found the said correction in Shtyrkov's paper
and can point the equation(s) into which it has
crept?
--
() ascii ribbon campaign - against html e-mail
/\ http://preview.tinyurl.com/qcy6mjc [archived]
[toc] | [prev] | [next] | [standalone]
Page 1 of 2 [1] 2 Next page →
Back to top | Article view | sci.physics.relativity
csiph-web