To improve the accuracy, I would integrate several measurements. There is no reason a sampled measurement at only one time needs to be made. On Fri, 30 Dec 2016 09:18:17 -0800, you wrote: >Hi Anders: > >That's something I've thought about for decades using an optical system. A few years ago I looked at it again and >found that astronomical "seeing" limits the accuracy. So the accuracy achieved by a spaceborne "Stellar compass" will >be much better than a ground based observation. A radio based observation might work since the atmosphere would not be >a factor. >http://www.prc68.com/I/StellarTime.shtml

On Fri, Dec 30, 2016 at 2:42 AM, Attila Kinali <attila@kinali.ch> wrote: > On Fri, 30 Dec 2016 10:59:03 +0200 > Anders Wallin <anders.e.e.wallin@gmail.com> wrote: > >> out of curiosity, are there any amateur/semi-pro experiments that can >> measure the length of the solar or sidereal day to sub-millisecond >> resolution? Yes. It is not hard at all to measure the Earth's rotational period, if all you needs is "sub millisecond" It would get harder if you cared about nanoseconds. I worked on an amateur project with some others and while measuring the Earth was not the goal we had to know the Earth's rational period to do the work. The project was about stellar photometry. But I leave that part out..... Basically what we did was mount a camera made out of a small CCD sensor and a 135mm f/2.8 camera lens salvaged from an old 35mm film camera. The camera was fixed to the roof of my garage. (This was THE big cost saving feature: The camera could not move. The mount as fixed at one location in the sky forever, right at the equator) I placed it in one end of a long wood crate and it looked up at the equator through a square hole on the upper end of the box. The box provided some protection from the elements and provided a lot of light shielding. To measure Earth's rotation all you need to do in know exactly when you took an image and to have a GOOD catalog of star locations. Let's say your image captures 200 stars. They are rather blurry and each covers maybe 5 pixels but even so you compute the centroid of each "gaussian blob" and then do a least squares fit of all those centroids to the astrometric catalog. The catalog is "good" to several milliacrseconds and with hundreds of centroids you can figure out were the camera was printed to a few "MAS" (Milli Arc Seconds). We took many images every clear night for several years. Hardware cost today is "not much" and you can use salvaged camera equipment Almost all of the software is available for free. Certainly matching stares to catalog images is. Yes the lens has geometric distortion and the CCD is likely not exactly 90 degrees to the optical axis but the software models this. This is possible because millions of star positions are known to insane levels of accuracy and if they appear in the "wrong" place in your image you can bet the cause is geometric distortion in your camera, especially after seeing the same error in hundreds of images. We used narrow filters to limit the image to just one "color" so the chromatic aberration in the optics i not an issue, filters are cheap. As part of our processing we time-tagged each image and also recored where the optical xis was pointed at. So you'd need a small telescope or big camera lens and a camera that can be triggered by a computer and software. Not really expensive. I'd invest in the best used optics you can and get a monochrome camera. Some people in the past century used transit telescopes to manually measure the time a star crossed a hairline in an eyepiece. Then the next night to observe the same star again. Now you know the length of the day (after you reduce the data) Put you can measure a dozen stars every night and take an average. In concept it is very simple. But today we can measure a tens of thousands of stars per day from a suburban roof top. Almost all other methods of measuring the Earth's rotation do not collect enough data. You need tens or hundreds of thousands of data points. if you want to know the sidereal period to Time Nut standards -- Chris Albertson Redondo Beach, California

  On Friday, 30 December 2016 11:43 PM, Attila Kinali

          Attila Kinali

Yes, the below is basically correct. But you save a ton of time and get better results if you simply bolt the telescope down to the Earth so that it can't move at all. The aim point just needs to be "close" and then later you determine where it is aimed. If you are only measuring period you don't need a surveyed location. If measuring absolutely time you do. Using a fixed mount is what makes this affordable by amateurs. Epoxy the camera to a fixed masonry building. This removes an unknown and dramatically simplifies the processing and also saves most of the cost witch is always the mechanical stuff. One package of JB Weld epoxy replaces thousands of dollars of motors and encoders and precision gears. With a fixed mount camera you have two kinds of "tine", that observed by the camera and a second from your GPSDO. If they diverge then you deduce that it must be the Earth's rotation that changed. But maybe you wonder of maybe the camera moved or some effect you forgot to remove. So it is but to have some buddies running the same setup in different cities around the world and check that you all see the same results. That is what we did. It is FAR EASIER to do this kind of replication when the setup is very inexpensive. Today you could build a camera for a LOT less then we did. I'm thinking of a surplus used lens from a 35mm film camera. A 250mm lens or so and a 3D printed plastic part that holds this to a cheap point and shoot camera. We used epoxy to held the lens to the camera, it meant you'd never be abler to take it apart again but it was going on a roof top, rain and all. On Fri, Dec 30, 2016 at 3:16 AM, Ilia Platone <info@iliaplatone.com> wrote: > Bruce, > > I think that you refer on prjects like Astrometry plate solving. I think one > should got a reference to get a time reference instead of scope "pointing" > reference, so, once one's got local coordinates in encoder positions, for > example the values of the north pole with an alt/az mounting, can use a > sub/arcsec plate solver to obtain good sidereal timing reference. using two > encoders helps much. > > The problem can be visibility of the reference points, however. > > Best Regards, > > Ilia. > > > > On 12/30/16 10:59, Bruce Griffiths wrote: >> >> Attila >> Lookup "Stellar compass" as used for determining space probe attitude.Can >> also be used to determine the direction of the centre of an image of a field >> of bright stars.Subarcsecond accuracy is fairly routine.Pattern recognition >> techniques combined with measures of the relative brightness of the stars is >> used to identify them.Subpixel accuracy in determining the location of the >> stellar image centroids is also routine. >> There is at least one US PhD thesis on such stellar compass techniques.A >> stellar compass technique has been used to determine the pointing direction >> of small portable telescopes without requiring precision axis encoders etc. >> Bruce >> >> On Friday, 30 December 2016 11:43 PM, Attila Kinali >> <attila@kinali.ch> wrote: >> >> On Fri, 30 Dec 2016 10:59:03 +0200 >> Anders Wallin <anders.e.e.wallin@gmail.com> wrote: >> >>> out of curiosity, are there any amateur/semi-pro experiments that can >>> measure the length of the solar or sidereal day to sub-millisecond >>> resolution? >>> To reproduce data like this: >>> >>> https://upload.wikimedia.org/wikipedia/commons/5/5b/Deviation_of_day_length_from_SI_day.svg >>> >>> Something in the sky that goes "ping" every day - detected with a >>> pointing >>> accuracy of < 1ms/24h or <0.01 arc-seconds (!?). Or perhaps two >>> satellite-dishes pointed at the sun and >>> noise-correlation/interferometry?? >> >> I don't know of any such experiment already performed, but I am not up >> to date on what's going on in the hobby astronomy community. >> >> I am not sure whether sub-milisecond resolution is feasible, but >> I think the "easiest" method would be to do a "modern" version of >> an meridian telescope: >> >> Using a camera fix mounted (ie not moving and if possible vibration >> isolated) >> on a pedestal pointed at the sky, approximately looking south. A simple >> webcam would be probably enough for first experiments, as long as you get >> a good picture of the stars. A good compact camera which allows to use >> a remote shutter with a proper lens and exposure control should be better. >> Probably the best resource here are the people/websites that deal with >> book scanning, as they tend to automate the whole picture taking process. >> Using magic lantern (http://magiclantern.fm) with Canon cameras might >> give additional features needed for the task. >> >> >From the pictures taken, calculate the positions of the stars (by fitting >> circles onto the bright pixels) and figure out which star is which (using >> astronomical list of stars). For this step there is a plethora of open >> source >> astronomical software available, but I don't know how well they fit the >> task >> of figuring out what the position of the stars relative to the camera >> reference >> frame. After that, it's just some simple math of calculating the >> difference >> between the position of the stars and where you would have expecteded them >> at the time when the picture has been taken. >> >> Some usefull software projects are: >> http://astro.corlan.net/gcx/ >> http://www.clearskyinstitute.com/xephem/ >> http://starlink.eao.hawaii.edu/starlink >> http://astro.corlan.net/avsomat/index.html >> http://rhodesmill.org/pyephem/ >> >> HTH >> >> Attila Kinali >> > > -- > Ilia Platone > via Ferrara 54 > 47841 > Cattolica (RN), Italy > Cell +39 349 1075999 > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. -- Chris Albertson Redondo Beach, California

 On Friday, 30 December 2016 11:43 PM, Attila Kinali

         Attila Kinali

HI > On Dec 30, 2016, at 4:53 PM, Chris Albertson <albertson.chris@gmail.com> wrote: > > Yes, the below is basically correct. But you save a ton of time and > get better results if you simply bolt the telescope down to the Earth > so that it can't move at all. The aim point just needs to be "close" > and then later you determine where it is aimed. If you are only > measuring period you don't need a surveyed location. If measuring > absolutely time you do. > > Using a fixed mount is what makes this affordable by amateurs. Epoxy > the camera to a fixed masonry building. This removes an unknown and > dramatically simplifies the processing and also saves most of the cost > witch is always the mechanical stuff. One package of JB Weld epoxy > replaces thousands of dollars of motors and encoders and precision > gears. > > With a fixed mount camera you have two kinds of "tine", that observed > by the camera and a second from your GPSDO. If they diverge then you > deduce that it must be the Earth's rotation that changed. But maybe > you wonder of maybe the camera moved or some effect you forgot to > remove. So it is but to have some buddies running the same setup in > different cities around the world and check that you all see the same > results. That is what we did. It is FAR EASIER to do this kind of > replication when the setup is very inexpensive. > > Today you could build a camera for a LOT less then we did. I'm > thinking of a surplus used lens from a 35mm film camera. A 250mm lens > or so and a 3D printed plastic part that holds this to a cheap point > and shoot camera. If by 3D print you mean a filament machine, pretty much all of the useful filaments will soffen / deform / melt if left in the hot sun for a day. The few that are more temperature tolerant require an unusual printer to work with and are a real pain to work with. One alternative would be a 3D printed mould and a metal casting process. It would take more than a few steps to get to metal. You also would need to clean up the casting before it was of any use. Bob > We used epoxy to held the lens to the camera, it > meant you'd never be abler to take it apart again but it was going on > a roof top, rain and all. > > > On Fri, Dec 30, 2016 at 3:16 AM, Ilia Platone <info@iliaplatone.com> wrote: >> Bruce, >> >> I think that you refer on prjects like Astrometry plate solving. I think one >> should got a reference to get a time reference instead of scope "pointing" >> reference, so, once one's got local coordinates in encoder positions, for >> example the values of the north pole with an alt/az mounting, can use a >> sub/arcsec plate solver to obtain good sidereal timing reference. using two >> encoders helps much. >> >> The problem can be visibility of the reference points, however. >> >> Best Regards, >> >> Ilia. >> >> >> >> On 12/30/16 10:59, Bruce Griffiths wrote: >>> >>> Attila >>> Lookup "Stellar compass" as used for determining space probe attitude.Can >>> also be used to determine the direction of the centre of an image of a field >>> of bright stars.Subarcsecond accuracy is fairly routine.Pattern recognition >>> techniques combined with measures of the relative brightness of the stars is >>> used to identify them.Subpixel accuracy in determining the location of the >>> stellar image centroids is also routine. >>> There is at least one US PhD thesis on such stellar compass techniques.A >>> stellar compass technique has been used to determine the pointing direction >>> of small portable telescopes without requiring precision axis encoders etc. >>> Bruce >>> >>> On Friday, 30 December 2016 11:43 PM, Attila Kinali >>> <attila@kinali.ch> wrote: >>> >>> On Fri, 30 Dec 2016 10:59:03 +0200 >>> Anders Wallin <anders.e.e.wallin@gmail.com> wrote: >>> >>>> out of curiosity, are there any amateur/semi-pro experiments that can >>>> measure the length of the solar or sidereal day to sub-millisecond >>>> resolution? >>>> To reproduce data like this: >>>> >>>> https://upload.wikimedia.org/wikipedia/commons/5/5b/Deviation_of_day_length_from_SI_day.svg >>>> >>>> Something in the sky that goes "ping" every day - detected with a >>>> pointing >>>> accuracy of < 1ms/24h or <0.01 arc-seconds (!?). Or perhaps two >>>> satellite-dishes pointed at the sun and >>>> noise-correlation/interferometry?? >>> >>> I don't know of any such experiment already performed, but I am not up >>> to date on what's going on in the hobby astronomy community. >>> >>> I am not sure whether sub-milisecond resolution is feasible, but >>> I think the "easiest" method would be to do a "modern" version of >>> an meridian telescope: >>> >>> Using a camera fix mounted (ie not moving and if possible vibration >>> isolated) >>> on a pedestal pointed at the sky, approximately looking south. A simple >>> webcam would be probably enough for first experiments, as long as you get >>> a good picture of the stars. A good compact camera which allows to use >>> a remote shutter with a proper lens and exposure control should be better. >>> Probably the best resource here are the people/websites that deal with >>> book scanning, as they tend to automate the whole picture taking process. >>> Using magic lantern (http://magiclantern.fm) with Canon cameras might >>> give additional features needed for the task. >>> >>>> From the pictures taken, calculate the positions of the stars (by fitting >>> circles onto the bright pixels) and figure out which star is which (using >>> astronomical list of stars). For this step there is a plethora of open >>> source >>> astronomical software available, but I don't know how well they fit the >>> task >>> of figuring out what the position of the stars relative to the camera >>> reference >>> frame. After that, it's just some simple math of calculating the >>> difference >>> between the position of the stars and where you would have expecteded them >>> at the time when the picture has been taken. >>> >>> Some usefull software projects are: >>> http://astro.corlan.net/gcx/ >>> http://www.clearskyinstitute.com/xephem/ >>> http://starlink.eao.hawaii.edu/starlink >>> http://astro.corlan.net/avsomat/index.html >>> http://rhodesmill.org/pyephem/ >>> >>> HTH >>> >>> Attila Kinali >>> >> >> -- >> Ilia Platone >> via Ferrara 54 >> 47841 >> Cattolica (RN), Italy >> Cell +39 349 1075999 >> >> _______________________________________________ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > > > -- > > Chris Albertson > Redondo Beach, California > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there.