26 replies to this topic

[Mike W]

Dr. Don Bruns, optical physicist chose the Televue- NP101is to repeat the 1919 experiment of measuring star deflections during a solar eclipse to confirm Einsteins theory of relativity. This experiment will be conducted during the august 2017 solar eclipse. Dr. Bruns chose the NP101is, FLI Microline 8051 ccd camera, and the Software Bisque Myt Paramount with field tripod. The Televue- NP101is was chosen for it's pinpoint sharp star images corner to corner for accurate measurements.
More about this experiment can be found on Televues Blog on their website.
 
TV 102
Gibraltar mount
Celestron CPC 800 HD edge
 
TV 32 plossl
"     24 Pan
"     18.2 Delite
"     15   "      "
"     11   "      "

Edited by Mike W, 08 July 2017 - 02:15 PM.

[DLuders]

Yes indeed, that Tele Vue Blog article can be found at http://televue.com/n...r/#.WWDyruSWyM8

Edited by DLuders, 08 July 2017 - 09:58 AM.

[Jon Isaacs]

Dr. Don Bruns, optical physicist chose the Televue NP101-is to repeat the 1919 experiment of measuring star deflections during a solar eclipse to confirm Einsteins theory of relativity. This experiment will be conducted during the august 2017 solar eclipse. Dr. Bruns chose the NP101-is, FLI Microline 8051 ccd camera, and the Software Bisque Myt Paramount with field tripod. The Televue NP101-is was chosen for it's pinpoint sharp star images corner to corner for accurate measurements.

More about this experiment can be found on Televues Blog on their website.

 

Dr. Bruns lives in the San Diego area and I had the good fortune of meeting him.  He was selling a 6 inch Starblast he had used in a demonstration of a principle.  I spent an hour and a half just shooting the breeze. 

 

I remember him telling me that he had previously used an NP-101 in his research.. Don was/involved in adaptive optics but I had the feeling that the small company he had been part of specialized in those projects that are well funded but require a top secret clearance.. 

 

He also is in the process of developing multiaxial adaptive optics for the amateur.

 

He's also a member of Cloudy Nights.. 

 

Jon

[astrogeek64]

There was a NP-101 for sale, $2000, used but in great condition at a recent star party........ It was killing me.

[Mike W]

There was a NP-101 for sale, $2000, used but in great condition at a recent star party........ It was killing me.

Well I'm glad you lived through it, next time you may not be so lucky!

Edited by Mike W, 08 July 2017 - 11:15 AM.

[Bonco]

So what's going to happen if his results DO NOT validate Einstein? Maybe next time he'll use a bigger AP!...Just joking, please don't take this seriously. What a great project.

Bill

[Mike W]

I believe in 1973 this experiment was done with an 11% error and Dr. Bruns is trying to narrow that down to 1%, something that could not be done at the time with the equipment available. Also from what I read, and did no know, the accuracy of the 1919 measurements have been called into question, So I will be following this experiment very closely!

Also note that since 1919 there have been many other methods by modern astrophysics that don't contradict this theory such as:

1.Gravitational redshift. (By HST)

2.Gravitational lensing

3.Light travel time delay

4.The Equivalence principle

5.Binary pulsars

6.Frame dragging tests

(And I'm sure there are many more) 

Edited by Mike W, 08 July 2017 - 06:40 PM.

[leexeen]

Wondering what constellation  the sun will be when the eclipse happens. Looks like the initial verification was done when the sun was in constellation Taurus where there were plenty stars to measure which make the comparison easy.

[Mike W]

Any modern computer generated star chart should show you.

[Cotts]

  The predicted displacement at the edge of the sun is less than 2 arc seconds.  The difficulty is that stars this close to the sun are not visible due to the bright corona.  At a distance far enough from the sun the displacement drops to less than an arc second.  

 

A Four inch telescope isn't going to be enough, it seems.

 

I also wonder about the methodology.  Perhaps take an image of the field without the sun and then with the sun there during the eclipse.  Compare images.   Would this work?

 

here's an interesting article:

 

http://www.newtonphy.../appendix2.html

 

The seeing in the daytime will likely deeply bury any displacement.....

 

Dave

[Mike W]

Please read the above article before discounting the experiment, being an optical physicist I'm sure Dr. Bruns has taken all this into account.

[Cotts]

I've done a little number crunching.  The airy disc diameter for a 101mm scope is 1.25", an important stat. In perfect, perfect seeing the stellar image will be a circle 1.25" in diameter.  

 

At prime focus of an NP 101 with the FLI Microline 8051 camera (5.5micron pixels) and the stated (in the Televue blog) goal of achieving 0.02 pixel accuracy of measurement:

 

Image scale, 2.06" per pixel.  So  0.02 pixel = 0.0412"  

 

the limiting factor at this focal length is that the image is badly undersampled - the airy disc is only half the size of the pixel. This is an uncertainty of nearly 1 whole arcsecond.  There is no way of determining where on any given pixel the star actually is.  The 0.02 pixel accuracy goal is unachievable.

 

The way to deal with this problem is to make the focal length longer so that the scopes airy disc will be as large as at least two pixels.  Using a 4x barlow the image scale becomes 0.51" per pixel and the 0.02 pixel accuracy level becomes 0.0103".

 

0.01" measurement accuracy is a very tight standard, even for large, professional observatories.  This small angle is 0.008 - 8 one thousandths - of the diameter of the Airy disc.   If the seeing is perfect.   Which it never is.  

 

I wish Dr. Bruns the best of luck with the study.

 

Dave

[dbruns]

Hi all,

 

I just found put about this thread, and am happy to answer any questions!  That is how I learn, too.

 

More details on my experiment can be found on my web page at www.stellarproducts.com (then click on the eclipsed sun image).  The technical paper presented at the SAS in 2016 is still pretty accurate, and the SAS talk (downloadable on the same web page) does explain more about the pixel problem.  I'll be presenting a longer speech in Casper at Astrocon on August 17.

 

This is still a difficult experiment, but I've completed some simulated results (everything except the sun's gravity) down to about 2% relative error.  Daytime turbulence is, in fact, the biggest factor, and that is why I am setting up at 8000' altitude in Wyoming.  Seeing is typically good near sunset, so too, hopefully, during an eclipse.  My twilight experiments resulted in stellar FWHM about 1.4 pixels, or about 2.9 arcsec.  Finding the centroid to about 0.02 pixels is required, but this isn't too hard with modern software.  The turbulence has to be averaged over lots of frames and deflections averaged over all the stars.  If the seeing is too good, then there is a pixel problem.  I've done tradeoffs for a wide variety of telescopes and cameras, and the Tele Vue NP101is and the FLI ML8051 camera are the best combination of resolution and field of view.

 

I'll be using the UCAC5 star catalog, so that will be more accurate than more images 6-month from the eclipse.

 

Astronomers in the 1980's used quasars and radio telescopes to measure Einstein's deflections down to 0.01% - this project is simply a fun project in my retirement.

 

I hope everyone has clear skies on August 21!

[Cotts]

I read your web page and the downloadable article.   It's nice to see that my 'back of the envelope' number crunching mostly agreed with the numbers you presented therein.

 

I come from a background of measuring visual binary stars via video lucky imaging.  Your multiple exposures to average out the seeing and using software to calculate centroids is virtually the same methodology.  Where we differ and the source of my worry is that using the NP-101 at only 540 mm focal length will result in a rather large undersampling of the stellar images and reduced ability to calculate centroid positions to the accuracy you desire.

 

By the way, what software are you using for centroid calculation?  I have used AstroimageJ with great success - it gives centroids to a rather optimistic 6 decimal places of a pixel!!  Of course I only use three decimals for calculations  (six decimal places of a 5 micron pixel is on the order of magnitude of larger organic molecules!)  and report only 2 decimal places..

 

Of course, you are juggling the need for the widest field possible with the need for the greatest centroid calculation accuracy.  Also, the shorter focal length allows you to get useful images of fainter stars.

 

When I measure binary stars, lately with 180mm and 200mm telescopes, I have routinely used  focal lengths around f/30 or about 3000mm.  At this focal length and with this aperture separations to 0.01" accuracy can be measured and published.   My colleagues and fellow amateur binary star enthusiasts  (see http://jdso.org) also use such long focal lengths and 0.01" reporting accuracy is just about the norm.  Very few, if any, of the amateurs submitting measures resulting from video lucky imaging to the JDSO use telescopes smaller than 200mm and focal lengths less than f/15.  

 

I recognize that if you were to use f/15 or more with the NP-101 you wouldn't even get the entire eclipsed sun in the field of view........

 

You are hoping for similar accuracy with a telescope with half the aperture and a very short focal length. For field of view reasons you really have no choice in this - I understand fully.  I very much hope that you do achieve this success (and prove me wrong! ) and I look forward to seeing your results.  

 

Dave

[astrogeek64]

 

There was a NP-101 for sale, $2000, used but in great condition at a recent star party........ It was killing me.

Well I'm glad you lived through it, next time you may not be so lucky!

 

This telescope tried to kill me...............

[Mike W]

I'm glad you and Dave could get together Dr. Bruns and best of luck with your experiment, it'll be chilly at 8000'! I've hiked Mt. Windy (about 8000') in the Cascades in july and we camped at 6000' and snow was in the air when we climbed out of the tent in the morning!

[dbruns]

Hi Cotts,

 

Regarding the software to find centroids, I use MaximDL, Astrometrica, and Visual Pinpoint.  I did try AstroJ, but it was not accurate enough.  If I remember correctly, AstroJ uses Astrometry.net, and that only fits quadratic curves to the peaks of the star images.  The other software mentioned above uses gaussian fits to the entire curve.  I did lots of tests on all of the software, include Prism, and settled on the three mentioned above.

 

If I have clouds next month, I hope that by 2024, larger cameras will be available with faster downloads (by then, USB7.0 ?).  Then I'll get a larger aperture and longer focal length!

 

Hope this helps,

Don

[Cotts]

Don, I use AstroJ with my stacked images of double stars simply to get the centroid location in (x,y) coordinates.  There is no connection to astronomy.net or any internet at all.   All I really need is the separation and position angle.

 

In your experiment you need centroids from a plate solver for the pre-eclipse and actual eclipse images to look for differences.  I guess the software you have chosen, based upon research, is the best.

 

I am very keen to hear of your results...

 

Dave

[dbruns]

Hi Dave,

 

I'll know within a few hours after totality, if I was able to capture lots of stars.  I won't have any detailed conclusions for at least a month; I need time to do all the analysis and take a nice vacation, too!  I'll be posting my results on my web page, writing a technical paper, and talking at NEAF2018, but I can add to this thread, too.

 

Don

[Peter Besenbruch]

I'll know within a few hours after totality, if I was able to capture lots of stars.  I won't have any detailed conclusions for at least a month; I need time to do all the analysis and take a nice vacation, too!  I'll be posting my results on my web page, writing a technical paper, and talking at NEAF2018, but I can add to this thread, too.

Yes, link to your Web page from here, when your paper is ready.

[dbruns]

I haven't updated this thread since July, but all my analysis has been completed.  I had perfect focus, perfect weather, good seeing, perfect exposure and perfect timing in Wyoming.  Everything worked!  My results have been posted on my website at stellarproducts.com, and my final paper has been accepted for publication.  There is also a note on the Sky&Tel News site.  Thanks to my sponsors (Tele Vue, Fingerlakes, and Software Bisque) for providing the right equipment!  My final deflection coefficient matches the theoretical value to much better than 0.1%. I got 1.7512 arcseconds.  Best results ever, the first time in 100 years that the experiment worked without problems.

[North of Sixty]

Bravo. Thanks you so much for the update. 

[Axunator]

Extremely cool! Congratulations! 

[Aeternam]

Just wow. Congratulations!

 

This experiment is a testament to the quality of currently available consumer gear. What a great time to live in.

[Erik Bakker]

A wonderful achievement indeed, just stunning results. Congratulations on all things coming together in your well prepared experiment and analysis