reply re Harrison's timing method - #13 in Vol 176, Issue 44 digest
Ben Bradley stated > "Perhaps closer to your question: I recall in my
readings about clockmaker John Harrison (likely either in "The Quest
for Longitude” or Dava Sobel's "Longitude") that he would look from
the edge of his window at a particular star each night and note (while
counting the ticks he heard from his clock) the exact moment it would
disappear behind a nearby chimney, and knowing the Earth's rotation
takes four minutes and some (I forget) seconds off from a day, he used
this to calibrate and test the precision and accuracy of his long
clocks. It was suggested he could get within less than second with
this method."
From Sobel - Chapter 7 > "The Harrison brothers tested the accuracy of
their gridiron-grasshopper clocks against the regular motions of the
stars. The crosshairs of their homemade astronomical tracking
instrument, with which they pinpointed the stars' positions, consisted
of the border of a windowpane and the silhouette of the neighbor's
chimney stack. Night after night, they marked the clock hour when
given stars exited their field of view behind the chimney. From one
night to the next, because of the Earth's rotation, a star should
transit exactly 3 minutes, 56 seconds (of solar time) earlier than the
previous night. Any clock that can track this sidereal schedule proves
itself as perfect as God's magnificent clockwork.”
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Harrison would watch a single star (obviously the same star for
several nights) in the gap between the right vertical edge of a window
mullion and the left edge of his neighbors chimney. He would move his
eye so as to always keep the star in the gap. Eventually, the gap
closes to zero and the star ‘winks out’. At that point he would
verbally signal his assistant watching Harrison's clock pendulum tip
swinging against a degree scale below. Harrison’s grasshopper
escapement clocks had a very large amplitude (+/- several degrees)
compared to that used by precision clocks today, so it is said that
the assistant could record the results to a fraction of a second.
Jonathan Betts has a description of the method in his “Harrison”
published by the National Maritime Museum in 2007 - see attachment.
A pendulum clock is not required to verify the method - all that is
needed is a similar star sighting arrangement and a means to record
the time of the ‘wink out’ - preferably to a fraction of a second.
Subsequent night ‘wink out’ times should be 3 minutes, 56 seconds
apart. (Is that single value valid over a 400 years period?)
Bob Holmström
Editor
Horological Science Newsletter
It must be a sign of the dedication that Harrison applied to his work.
It is not as simple as the description first appears, this is England,
and the method presupposes that there are no clouds. It might be a week
or two before two nights occurred, when an unclouded night was followed
by another night within several days that was not clouded.
Similarly, with the longitude method, stars must be visible within a
short period of dawn or dusk, when the horizon is visible together with
the star. Sun sights are not so difficult.
GPS makes is so easy for us!
cheers,
Neville Michie
On 27 Mar 2019, at 09:48, Bob Holmstrom holmstro@gmail.com wrote:
Ben Bradley stated > "Perhaps closer to your question: I recall in my
readings about clockmaker John Harrison (likely either in "The Quest
for Longitude” or Dava Sobel's "Longitude") that he would look from
the edge of his window at a particular star each night and note (while
counting the ticks he heard from his clock) the exact moment it would
disappear behind a nearby chimney, and knowing the Earth's rotation
takes four minutes and some (I forget) seconds off from a day, he used
this to calibrate and test the precision and accuracy of his long
clocks. It was suggested he could get within less than second with
this method."
From Sobel - Chapter 7 > "The Harrison brothers tested the accuracy of
their gridiron-grasshopper clocks against the regular motions of the
stars. The crosshairs of their homemade astronomical tracking
instrument, with which they pinpointed the stars' positions, consisted
of the border of a windowpane and the silhouette of the neighbor's
chimney stack. Night after night, they marked the clock hour when
given stars exited their field of view behind the chimney. From one
night to the next, because of the Earth's rotation, a star should
transit exactly 3 minutes, 56 seconds (of solar time) earlier than the
previous night. Any clock that can track this sidereal schedule proves
itself as perfect as God's magnificent clockwork.”
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Harrison would watch a single star (obviously the same star for
several nights) in the gap between the right vertical edge of a window
mullion and the left edge of his neighbors chimney. He would move his
eye so as to always keep the star in the gap. Eventually, the gap
closes to zero and the star ‘winks out’. At that point he would
verbally signal his assistant watching Harrison's clock pendulum tip
swinging against a degree scale below. Harrison’s grasshopper
escapement clocks had a very large amplitude (+/- several degrees)
compared to that used by precision clocks today, so it is said that
the assistant could record the results to a fraction of a second.
Jonathan Betts has a description of the method in his “Harrison”
published by the National Maritime Museum in 2007 - see attachment.
A pendulum clock is not required to verify the method - all that is
needed is a similar star sighting arrangement and a means to record
the time of the ‘wink out’ - preferably to a fraction of a second.
Subsequent night ‘wink out’ times should be 3 minutes, 56 seconds
apart. (Is that single value valid over a 400 years period?)
Bob Holmström
Editor
Horological Science Newsletter
<Screen Shot 2019-03-26 at 3.11.43 PM3.png>_______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On 3/26/19 3:48 PM, Bob Holmstrom wrote:
Ben Bradley stated > "Perhaps closer to your question: I recall in my
readings about clockmaker John Harrison (likely either in "The Quest
for Longitude” or Dava Sobel's "Longitude") that he would look from
the edge of his window at a particular star each night and note (while
counting the ticks he heard from his clock) the exact moment it would
disappear behind a nearby chimney, and knowing the Earth's rotation
takes four minutes and some (I forget) seconds off from a day, he used
this to calibrate and test the precision and accuracy of his long
clocks. It was suggested he could get within less than second with
this method."
From Sobel - Chapter 7 > "The Harrison brothers tested the accuracy of
their gridiron-grasshopper clocks against the regular motions of the
stars. The crosshairs of their homemade astronomical tracking
instrument, with which they pinpointed the stars' positions, consisted
of the border of a windowpane and the silhouette of the neighbor's
chimney stack. Night after night, they marked the clock hour when
given stars exited their field of view behind the chimney. From one
night to the next, because of the Earth's rotation, a star should
transit exactly 3 minutes, 56 seconds (of solar time) earlier than the
previous night. Any clock that can track this sidereal schedule proves
itself as perfect as God's magnificent clockwork.”
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
A small telescope and a camera might work, lining up with the two edges
as a "fixed offset knife edges". It could also work in day time (you
can see Polaris in the day time with a 28x telescope with a 1" objective
- a surveyor's theodolite)
There's a collection of navigation papers from ION available on CD-ROM
and there's a fair amount of info in there about celestial trackers and
detectors.
On 3/26/19 4:27 PM, Neville Michie wrote:
It must be a sign of the dedication that Harrison applied to his work.
It is not as simple as the description first appears, this is England,
and the method presupposes that there are no clouds. It might be a week
or two before two nights occurred, when an unclouded night was followed
by another night within several days that was not clouded.
Similarly, with the longitude method, stars must be visible within a
short period of dawn or dusk, when the horizon is visible together with
the star. Sun sights are not so difficult.
You can use an artificial horizon (historically, a pool of mercury) and
shoot the star and its reflection, then divide by two. I've used water
reflection in a bowl to shoot the sun and moon, but those are really
bright. I've not tried a star. It would be really hard in an area with
background light because you'd have trouble dark adapting. You can
fairly easily shoot moon and star together, though, to get an angle
between star and moon's limb.
Maybe you could float some aluminized mylar on the water surface to help
the reflectivity.
GPS makes is so easy for us!
Until the batteries go dead.
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
The Danjon impersonal astrolabe is perhaps better suited to accurate measurements:
https://www.nzmuseums.co.nz/collections/3267/objects/3380/astrolabe
Bruce
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On Wed 2019-03-27T16:26:09+1300 Bruce Griffiths hath writ:
The Danjon impersonal astrolabe is perhaps better suited to accurate measurements:
https://www.nzmuseums.co.nz/collections/3267/objects/3380/astrolabe
Danjon became director of Observatoire de Paris (and thus also the
BIH) in 1945. In 1948 the ITU realized that it could not make
reasonable regulations about frequency in the absence of an expert on
timing, so ITU requested the IAU send a representative to Study Group
7, and Danjon was that representative. Danjon was head of the CCDS
from inception into the late 1960s.
The raw data about the clocks at Observatoire de Paris (which were
located in the catacombs for temperature stability) from the 1920s
into the quartz era are almost all published in Bulletin Horaire. The
early issues are all scanned and online at Harvard ADS. That is a
huge amount data on earth disciplined oscillators, a treasure chest of
descriptions and diagrams of early circuits and drum recording
devices, and a pile of dirty laundry about who made good and bad
decisions in international agreements about time.
--
Steve Allen sla@ucolick.org WGS-84 (GPS)
UCO/Lick Observatory--ISB 260 Natural Sciences II, Room 165 Lat +36.99855
1156 High Street Voice: +1 831 459 3046 Lng -122.06015
Santa Cruz, CA 95064 https://www.ucolick.org/~sla/ Hgt +250 m
These light curves for a star being occulted by the moon should give some idea of the effects of diffraction:
http://tdc-www.harvard.edu/occultations/moon/vb141occa.html
Bruce
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Of course the time scale will be much shorter for a star occulted by an edge etc on Earth.
The longer time scale for a lunar occultation is due to the slower relative angular motion of the moon with respect to the star than the motion of the Earth with respect to a star.
Bruce
On 27 March 2019 at 17:41 Bruce Griffiths bruce.griffiths@xtra.co.nz wrote:
These light curves for a star being occulted by the moon should give some idea of the effects of diffraction:
http://tdc-www.harvard.edu/occultations/moon/vb141occa.html
Bruce
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Hi Bruce:
Would the David White 60 Degree Pendulum Astrolabe also work?
https://prc68.com/I/PendulumAstrolabe.shtml
--
Have Fun,
Brooke Clarke
https://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
axioms:
- The extent to which you can fix or improve something will be limited by how well you understand how it works.
- Everybody, with no exceptions, holds false beliefs.
-------- Original Message --------
The Danjon impersonal astrolabe is perhaps better suited to accurate measurements:
https://www.nzmuseums.co.nz/collections/3267/objects/3380/astrolabe
Bruce
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
There is a large period literature on ³dialing² which not only included
sundials, but all sorts of ways to measure time from celestial objects
using angles. Discussions of trigonometry, surveying, navigation, and
³dyaling² in relationship were also quite common during the period. These
could be found in popular books, and sometimes in almanacs (some of which
had huge circulations).
The use of building features to reckon sidereal time is quite old. In the
6th century, Gregory of Tours wrote DE CURSU STELLARUM about sidereal
timekeeping, and the techniques were later simplified to use buildings
within a monastery in HOROLOGIUM STELLARE MONASTICUM.
The secondary sources on these two medieval works are:
Constable, Giles. 1975. Horologium Stellare Monasticum. Corpus
Consuetudinum Monasticarum, volume 6, pages 1-18.
McCluskey, Stephen. 1990. Gregory of Tours, Monastic Timekeeping and
Early Christian Attitudes to Astronomy. Isis, volume 81, pages 8-22.
Sobel gives the false impression that such knowledge was held by only a
few and that Harrison was a bit of a country bumpkin, but in fact, when
one considers that the inventor of the deadbeat escapement was from
Carlisle, and that gear-cutting machines for clocks were improved by
Hindley in York, this suggests that there was widespread interest in
clock-making in addition to the widespread interest in astronomy in the
North of England, and a look at all the books and pamplets published on
the topic suggests that the interest extended into Ireland and Scotland,
as well.
Basically, the sort of expertise to do what Harrison did using his window
and a neighbor¹s chimney was extremely widespread, and the literature
giving instructions is overwhelming in its size. I did a quick search on
astronomy and dialing in one of my go-to databases (18th Century
Collections online) and got 383 hits.
If you know French, you can go to gallica at
https://gallica.bnf.fr/accueil/en/content/accueil-en?mode=desktop and
search for works on astronomy and dialing, and find similar sources,
although my impression always has been that the English were more obsessed
with the topic in the 17th and 18th century than the French.
Best,
Kevin
--
Kevin K. Birth, Professor
Department of Anthropology
Queens College, City University of New York
65-30 Kissena Boulevard
Flushing, NY 11367
telephone: 718/997-5518
"Tempus est mundi instabilis motus, rerumque labentium cursus." --Hrabanus
Maurus
"We may live longer but we may be subject to peculiar contagion and
spiritual torpor or illiteracies of the imagination" --Wilson Harris
On 3/26/19, 10:48 PM, "time-nuts on behalf of Tom Van Baak"
<time-nuts-bounces@lists.febo.com on behalf of tvb@leapsecond.com> wrote:
EXTERNAL EMAIL: please report suspicious content to the ITS Help Desk.
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison¹s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and
amateur timekeeping. Surely, a couple of people on the list could 1)
attempt to verify the Harrison method, and 2) determine what the limits
of its accuracy are, say, with little effort vs. with hard work vs. with
extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night.
Any identifiable star that crosses your edge is a recordable timing event
that evening. So, in theory, if you measure N stars you get sqrt(N)
improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the
riddle, either by uncovering existing professional or amateur literature
or by actually trying this at home. It boils down to how accurately can
you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the
middle of the 20th century was always a form of "Earth Disciplined
Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept
accurate time short-term and star tracking (earth rotation) kept accurate
time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a
combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On Wed, Mar 27, 2019 at 2:00 AM jimlux jimlux@earthlink.net wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
A small telescope and a camera might work, lining up with the two edges
as a "fixed offset knife edges". It could also work in day time (you
can see Polaris in the day time with a 28x telescope with a 1" objective
- a surveyor's theodolite)
If I've calculated it correctly, I think that's about 1 pixel width on a
24megapixel APCS DSLR with a 55mm lens, which is easily achievable. It
would be nice to have 10 pixels to find the true centre, but since the
star's location will move predictably across the entire fov, it should be
possible to interpolate very effectively.
Relative to Harrison's method of star position measurement and timing, and
the
possible effects of diffraction thereof:
Hanbury Brown's book "The Intensity Interferometer" may provide some useful
insight about various error sources, even though in a different context,
that of
measurement of a star's angular size.
The opening chapter discusses some early attempts at angular size. Galileo
apparently made the first really serious attempt, based on how far behind a
cord of known diameter one must stand so that the cord barely blockes the
star's image in the eye. He found a size of 5 arcseconds for Vega, for
example.
Evidently previous estimates were in the two arc minute range.
Hanbury Brown mentions that a pretty detailed description of Galilieo's
methods
and attention to detail were given in Galileo's "Dialogue Concerning the Two
Chief World Systems".
It is interesting to note that Galileo's method still yielded a figure some
3 orders of
magnitude larger than Vega's currently accepted (as of 1974) angular size
of about
3 milli-arc seconds. Hanbury Brown attributed the discrepancy largely to
angular
scintillation from the Earth's atmosphere. It seems to me that this has to
be a noisy
factor in the Harrison method, too, and it would be interesting to see more
research
done on this point. It also seems to me that the diffraction issue would
be complicated
by the different distances to the two obstacles necessitated by Harrison's
approach.
Right now I'm beginning to think about whether I have the smarts and
computational
resources to do some simulations of this situation.
Hanbury Brown's book is basically about his invention of the Intensity
Interferometer,
which is an attempt to go beyond the capability of Michelson's stellar
interferometer.
The tale spans from HB's original late-night inspiration until the eventual
construction
and successful operation of a special telescope in Narrabri, Australia. It
is a fascinating
story winding through HB's own initial self doubts, scientific criticism of
the validity of
the concept, technical hurdles, bureaucratic hurdles like getting funding,
more technical
hurdles, etc.
Dana
On Wed, Mar 27, 2019 at 5:00 AM Brooke Clarke brooke@pacific.net wrote:
Hi Bruce:
Would the David White 60 Degree Pendulum Astrolabe also work?
https://prc68.com/I/PendulumAstrolabe.shtml
--
Have Fun,
Brooke Clarke
https://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
axioms:
- The extent to which you can fix or improve something will be limited by
how well you understand how it works. - Everybody, with no exceptions, holds false beliefs.
-------- Original Message --------
The Danjon impersonal astrolabe is perhaps better suited to accurate
measurements:
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy
and amateur timekeeping. Surely, a couple of people on the list could 1)
attempt to verify the Harrison method, and 2) determine what the limits of
its accuracy are, say, with little effort vs. with hard work vs. with
extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per
night. Any identifiable star that crosses your edge is a recordable timing
event that evening. So, in theory, if you measure N stars you get sqrt(N)
improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the
riddle, either by uncovering existing professional or amateur literature or
by actually trying this at home. It boils down to how accurately can you
measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the
middle of the 20th century was always a form of "Earth Disciplined
Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept
accurate time short-term and star tracking (earth rotation) kept accurate
time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a
combination of the two.
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Brooke
Yes but the accuracy would suffer due to observer related effects.
However when used with a CCD camera or equivalent the accuracy should improve somewhat much as adding a TV camera to a transit circle improved its accuracy. I had a personal tour of the USNO setup on Black- Birch/Altimarloch during their southern hemisphere campaign during the 1980's.
Bruce
On 27 March 2019 at 20:39 Brooke Clarke brooke@pacific.net wrote:
Hi Bruce:
Would the David White 60 Degree Pendulum Astrolabe also work?
https://prc68.com/I/PendulumAstrolabe.shtml
--
Have Fun,
Brooke Clarke
https://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
axioms:
- The extent to which you can fix or improve something will be limited by how well you understand how it works.
- Everybody, with no exceptions, holds false beliefs.
-------- Original Message --------
The Danjon impersonal astrolabe is perhaps better suited to accurate measurements:
https://www.nzmuseums.co.nz/collections/3267/objects/3380/astrolabe
Bruce
On 27 March 2019 at 15:48 Tom Van Baak tvb@LeapSecond.com wrote:
BobH wrote:
This would be an excellent project for time-nuts to verify. First, a
better explanation of John Harrison’s method is in order. A vertical
window edge is not sufficient - a second vertical reference at a
distance is required - Harrison used a chimney on a neighbor's house.
Agreed! The project is the perfect intersection of amateur astronomy and amateur timekeeping. Surely, a couple of people on the list could 1) attempt to verify the Harrison method, and 2) determine what the limits of its accuracy are, say, with little effort vs. with hard work vs. with extreme dedication.
JimL wrote:
To get 1 second accuracy, you need 360/86400 = 0.004 degree
measurements. That's 0.073 milliradian - 1 cm at 140 meter distance.
I'm not sure an "edge" is sharp enough (diffraction, etc.), although
your eye is pretty good at "deconvolving" the linear equivalent of an
Airy disk/rings.
Keep in mind too that one can take more than one star reading per night. Any identifiable star that crosses your edge is a recordable timing event that evening. So, in theory, if you measure N stars you get sqrt(N) improvement in accuracy per day.
I want to encourage anyone to study the problem and help solve the riddle, either by uncovering existing professional or amateur literature or by actually trying this at home. It boils down to how accurately can you measure earth rotation using the Harrison method.
To put this in time nuts context, precision timekeeping prior to the middle of the 20th century was always a form of "Earth Disciplined Oscillator". Not unlike a GPSDO, your observatory's pendulum clock kept accurate time short-term and star tracking (earth rotation) kept accurate time long-term. The ADEV's crossed just like a GPSDO.
The short-term ADEV of a really good pendulum clock is here:
http://leapsecond.com/pend/shortt/
The long-term ADEV of earth rotation is here:
http://leapsecond.com/museum/earth/
So the performance of a DIY earth disciplined oscillator would be a combination of the two.
/tvb
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
In message 236772484.9174006.1553757616865@webmail.xtra.co.nz, Bruce Griffith
s writes:
However when used with a CCD camera or equivalent the accuracy
should improve somewhat much as adding a TV camera to a transit
circle improved its accuracy.
You know ... there is an official time-nut way to do this.
You want is a chevron shaped 'Høg grid' because that is
objectively a very, very, very smart way of converting precise
time to precise geometry.
I don't know of any popular explanations, but look at page 10 here:
https://www.hs.uni-hamburg.de/DE/Ins/Bib/AG2012AK1.pdf
The illustration on page 10 shows the original concept (from 1925!):
By modulating the starlight with a non-uniform pattern, and sampling
the modulated light at high rate, the transit time of the star can
be determined on the order of the sampling frequency.
Notice that the photon detector does not need high geometric resolution,
I belive Strømberg, 17 year old at the time, used a simple photo-cell
or possibly a photo-multiplier.
Now, if you want to measure both coordinates, you move to the chevron
shaped grid illustrated on page 11, the "Høg grid".
You still get a precise measurement of the transit along the logitudal
axis, but the width of the signal now also tells you where the star
was on the transverse axis.
This is how the Perth 1970 catalog was made, and if not for a loose
bolt, it would have been the most precise catalog on both axis instead
of just one axis.
The Høg grid still leaves rotation as source of error, so look at
page 2 here:
https://pdfs.semanticscholar.org/27f4/16df19441874fcd3b1bf52c477c889ca8045.pdf
Imagine the light-curve you get when a star transits that slit system
in various directions, including, crucially, with a rotation[1].
About 12 years ago I did some ad-hoc experiements on my 5" telescope,
with various simple slit geometries, and it works a treat.
I made the slits by taping mylar tape on a neutral filter, and cut
slits with a scalpel and a steel ruler, the detector was a large
area PIN photo-diode from the junk box and a digital oscilloscope.
While you can prove the concept, as I did, with portable tripod
mount, to get usable data you have to bolt the telecope to a cubic
meter of concrete or bedrock.
Poul-Henning
[1] This becaue very important for the Hipparcos satelite which a
rocket failure left stranded in the parking orbit ... but they still
completed their science objectives.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Hi
I randomly came across:
Publications of the United States Naval Observatory
January 1, 1900
U.S. Government Printing Office
Which turns out to be a Google E-Book. It goes into some detail about just how the transit
data contained in it was obtained. For the (free) price it’s worth taking a look at.
Bob
On Mar 28, 2019, at 4:00 PM, Poul-Henning Kamp phk@phk.freebsd.dk wrote:
In message 236772484.9174006.1553757616865@webmail.xtra.co.nz, Bruce Griffith
s writes:
However when used with a CCD camera or equivalent the accuracy
should improve somewhat much as adding a TV camera to a transit
circle improved its accuracy.
You know ... there is an official time-nut way to do this.
You want is a chevron shaped 'Høg grid' because that is
objectively a very, very, very smart way of converting precise
time to precise geometry.
I don't know of any popular explanations, but look at page 10 here:
https://www.hs.uni-hamburg.de/DE/Ins/Bib/AG2012AK1.pdf
The illustration on page 10 shows the original concept (from 1925!):
By modulating the starlight with a non-uniform pattern, and sampling
the modulated light at high rate, the transit time of the star can
be determined on the order of the sampling frequency.
Notice that the photon detector does not need high geometric resolution,
I belive Strømberg, 17 year old at the time, used a simple photo-cell
or possibly a photo-multiplier.
Now, if you want to measure both coordinates, you move to the chevron
shaped grid illustrated on page 11, the "Høg grid".
You still get a precise measurement of the transit along the logitudal
axis, but the width of the signal now also tells you where the star
was on the transverse axis.
This is how the Perth 1970 catalog was made, and if not for a loose
bolt, it would have been the most precise catalog on both axis instead
of just one axis.
The Høg grid still leaves rotation as source of error, so look at
page 2 here:
https://pdfs.semanticscholar.org/27f4/16df19441874fcd3b1bf52c477c889ca8045.pdf
Imagine the light-curve you get when a star transits that slit system
in various directions, including, crucially, with a rotation[1].
About 12 years ago I did some ad-hoc experiements on my 5" telescope,
with various simple slit geometries, and it works a treat.
I made the slits by taping mylar tape on a neutral filter, and cut
slits with a scalpel and a steel ruler, the detector was a large
area PIN photo-diode from the junk box and a digital oscilloscope.
While you can prove the concept, as I did, with portable tripod
mount, to get usable data you have to bolt the telecope to a cubic
meter of concrete or bedrock.
Poul-Henning
[1] This becaue very important for the Hipparcos satelite which a
rocket failure left stranded in the parking orbit ... but they still
completed their science objectives.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
In principle one could time tag each individual photon with subnanosecond resolution.
Bruce
On 29 March 2019 at 09:00 Poul-Henning Kamp phk@phk.freebsd.dk wrote:
In message 236772484.9174006.1553757616865@webmail.xtra.co.nz, Bruce Griffith
s writes:
However when used with a CCD camera or equivalent the accuracy
should improve somewhat much as adding a TV camera to a transit
circle improved its accuracy.
You know ... there is an official time-nut way to do this.
You want is a chevron shaped 'Høg grid' because that is
objectively a very, very, very smart way of converting precise
time to precise geometry.
I don't know of any popular explanations, but look at page 10 here:
https://www.hs.uni-hamburg.de/DE/Ins/Bib/AG2012AK1.pdf
The illustration on page 10 shows the original concept (from 1925!):
By modulating the starlight with a non-uniform pattern, and sampling
the modulated light at high rate, the transit time of the star can
be determined on the order of the sampling frequency.
Notice that the photon detector does not need high geometric resolution,
I belive Strømberg, 17 year old at the time, used a simple photo-cell
or possibly a photo-multiplier.
Now, if you want to measure both coordinates, you move to the chevron
shaped grid illustrated on page 11, the "Høg grid".
You still get a precise measurement of the transit along the logitudal
axis, but the width of the signal now also tells you where the star
was on the transverse axis.
This is how the Perth 1970 catalog was made, and if not for a loose
bolt, it would have been the most precise catalog on both axis instead
of just one axis.
The Høg grid still leaves rotation as source of error, so look at
page 2 here:
https://pdfs.semanticscholar.org/27f4/16df19441874fcd3b1bf52c477c889ca8045.pdf
Imagine the light-curve you get when a star transits that slit system
in various directions, including, crucially, with a rotation[1].
About 12 years ago I did some ad-hoc experiements on my 5" telescope,
with various simple slit geometries, and it works a treat.
I made the slits by taping mylar tape on a neutral filter, and cut
slits with a scalpel and a steel ruler, the detector was a large
area PIN photo-diode from the junk box and a digital oscilloscope.
While you can prove the concept, as I did, with portable tripod
mount, to get usable data you have to bolt the telecope to a cubic
meter of concrete or bedrock.
Poul-Henning
[1] This becaue very important for the Hipparcos satelite which a
rocket failure left stranded in the parking orbit ... but they still
completed their science objectives.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.