How hard is to build a processor?

[-jg]

On Feb 27, 6:47 pm, Joe Pfeiffer <pfeif...@cs.nmsu.edu> wrote:

The clock/calendar I hope to build over the next year or so will be
solar.  The shadow of a post uniquely determines both date and time, if
you look at both angle and length....

Sounds like a cool idea.
Would need this sort of correction :
http://www.swanstrom.net/sundial/gnomon.htm

which means it could take a couple of days to 'train',
in order to be certain.

For the best precision, you'd probably do a simple sliding-data-
match, where maybe the last 7? days of readings, are moved along to a
point of least-errors.
(and maybe a cloudy-day default, where it just increments the day ?

I don't see why you think you can't get 'correct day'
precision out of this ?

Pushing the actual-time precision is likely to be
more challenging ?
-jg

 

[Paul Keinanen]

On Mon, 8 Mar 2010 14:20:46 -0800 (PST), -jg <jim.granville@gmail.com>
wrote:

On Feb 27, 6:47 pm, Joe Pfeiffer <pfeif...@cs.nmsu.edu> wrote:
The clock/calendar I hope to build over the next year or so will be
solar.  The shadow of a post uniquely determines both date and time, if
you look at both angle and length....

Sounds like a cool idea.
Would need this sort of correction :
http://www.swanstrom.net/sundial/gnomon.htm

which means it could take a couple of days to 'train',
in order to be certain.

For the best precision, you'd probably do a simple sliding-data-
match, where maybe the last 7? days of readings, are moved along to a
point of least-errors.
(and maybe a cloudy-day default, where it just increments the day ?

I don't see why you think you can't get 'correct day'
precision out of this ?

Possibly in March or September, but absolutely not in July or
December, just look at the graph.

A 7 day period would be sufficient to figure out, if it is March or
September by checking the direction of the solar movement.

Pushing the actual-time precision is likely to be
more challenging ?

That is trivial. The apparent solar diameter is 30 arc minutes and
since the sun moves 15 degrees each hour, it only takes 2 minutes to
move its own diameter, thus the actual noon can be determined with
much better precision.

In order to get the solar mean time noon, you also need to know the
approximate date to apply the equation of time.

 

[Jasen Betts]

On 2010-03-08, rickman <gnuarm@gmail.com> wrote:

On Mar 6, 1:43 am, Jasen Betts <ja...@xnet.co.nz> wrote:
On 2010-02-27, D Yuniskis <not.going.to...@seen.com> wrote:

Hi Joe,

Joe Pfeiffer wrote:
The clock/calendar I hope to build over the next year or so will be
solar.  The shadow of a post uniquely determines both date and time, if
you look at both angle and length....

Hmmm... is that (really) true?

yes.

Or, don't you end up with
*two* date,times for each angle,length?  

that may happen for some dates of some years

Actually, it happens for every day of every year other than the
solstices.

The solstice is not a day it is an instant,
and it does not happen same the date and wall time every year.

most years have no days equidistant from the solstice.

therefore noon (or any other hour) on most days will duplicate the same
elevation of the sun above the horzon.


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

 

[Jasen Betts]

On 2010-03-08, D Yuniskis <not.going.to.be@seen.com> wrote:

rickman wrote:
On Mar 6, 1:43 am, Jasen Betts <ja...@xnet.co.nz> wrote:
On 2010-02-27, D Yuniskis <not.going.to...@seen.com> wrote:

Hi Joe,
Joe Pfeiffer wrote:
The clock/calendar I hope to build over the next year or so will be
solar. The shadow of a post uniquely determines both date and time, if
you look at both angle and length....
Hmmm... is that (really) true?
yes.

Or, don't you end up with
*two* date,times for each angle,length?
that may happen for some dates of some years

Actually, it happens for every day of every year other than the
solstices. The two solstices (actually a day or two on either side
depending of the season) has the lowest or highest path across the
sky, so no other day will have quite in that same path. But every
time of every other day (excluding a few seconds at the start and end
of the day when one day has sunshine and the other does not) will
match a time of two days, between spring to fall and one between fall
to spring. The path of the sun may not be the same on those two days,
but each point will map to two different times and days.

So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

 

[D Yuniskis]

Jasen Betts wrote:

So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)? Probably just watch for a longer period of time?

 

[Walter Banks]

rickman wrote:

I can't say this for certain, but I believe the combination of length
of day and elevation of the sun at the zenith is a unique combination
for each day of the year and latitude. So I think you can get your
latitude the same day.

It is close but not exact. The earth's orbit is not an exact number of
days. For the calendar's purpose we accumulate errors and adjust
the calendar. These adjustments are every 4 years and sometimes
on the century. There are other errors that have an impact on the
observations depending on the required accuracy.

I saw a sundial on a beach near Kobe Japan that had elaborate
error correcting instructions that was probably good to a second
after ten minutes of calculations. There were a lot of factors involved
it accounted for earths orbital period

But I'm not sure you don't have the same two
day ambiguity. Otherwise I think the combination is unique. Even the
North-South issue can be resolved because of the eccentricity of the
Earth's orbit making things a little different in the two
hemispheres. But you may also be foiled beyond the artic/anartic
circles where the sun never sets. Then you only get one parameter,
the elevation at the zenith. But you might be able to make up for
that by measuring the time between the sun at due east and due west...
other than at the poles where there is no east or west... ;^)

Above the arctic circle the sun 24 hour path is tilted but there are other
factors that are significant. For a couple weeks around June 21 the
sun never sets as far as 80 miles or so south of the arctic circle.
Most of this is due to the optic effects of the atmosphere. Even above
the arctic circle actual and observed position of the sun has significant
differences.

Regards,

w..
--
Walter Banks
Byte Craft Limited
http://www.bytecraft.com

 

[rickman]

On Mar 9, 12:20 pm, D Yuniskis <not.going.to...@seen.com> wrote:

Jasen Betts wrote:
So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)?  Probably just watch for a longer period of time?

I can't say this for certain, but I believe the combination of length
of day and elevation of the sun at the zenith is a unique combination
for each day of the year and latitude. So I think you can get your
latitude the same day. But I'm not sure you don't have the same two
day ambiguity. Otherwise I think the combination is unique. Even the
North-South issue can be resolved because of the eccentricity of the
Earth's orbit making things a little different in the two
hemispheres. But you may also be foiled beyond the artic/anartic
circles where the sun never sets. Then you only get one parameter,
the elevation at the zenith. But you might be able to make up for
that by measuring the time between the sun at due east and due west...
other than at the poles where there is no east or west... ;^)

Rick

 

[D Yuniskis]

Hi Walter,

Walter Banks wrote:

I saw a sundial on a beach near Kobe Japan that had elaborate
error correcting instructions that was probably good to a second
after ten minutes of calculations. There were a lot of factors involved
it accounted for earths orbital period

I always thought a cool hack would be a motorized sundial.
(i.e., the motorization being a cleverly hidden aspect)
E.g., with nice, evenly spaced markings -- and a motor to
rotate the whole assembly such that the shadow fell
"where it should" (on this nicely marked indicator).

It;s the sort of thing that would elicit comment *only*
from someone who *knew* it was "quite impossible" to
work as it *suggests* it works...

(obviously, I like things that mess with people's heads :> )

 

[-jg]

On Mar 10, 6:20 am, D Yuniskis <not.going.to...@seen.com> wrote:

Jasen Betts wrote:
Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)?  Probably just watch for a longer period of time?

The fish-hook here is in the careful wording of
"and sufficiently good instruments, and the right data and skills,
yes"

So a smarter question, could be what is practical ?
- and using what measurement systems ?

I found this revealing page, which has real datapoints,
and a practical location (ie less than ideal)

http://www.austintek.com/astro/analemma/analemma.html

Most revealing are the nice dots-on-the-door
http://www.austintek.com/astro/analemma/images/4215.door_from_inside_rotate..jpg

and the red arcs, are snapshots of the actual path,
~4wks - note they include a dot on alternating sides
of the analemma, as the 12 arcs interlace.

This site below shows the analemma actually moves yr-yr, so that's
more data to track

http://myweb.tiscali.co.uk/moonkmft/Articles/EquationOfTime.html

-jg

 

[Paul Keinanen]

On Tue, 09 Mar 2010 10:20:08 -0700, D Yuniskis
<not.going.to.be@seen.com> wrote:

Jasen Betts wrote:
So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)? Probably just watch for a longer period of time?

By observing when the sum passes the meridian, one cloud free week in
the spring and one in the autumn should give a quite good resolution
for the latitude, provided that some internal time reference is
capable of measuring the number of days between the measuring periods
with at least +/-12 hour accuracy. During one week long period the sun
moves south and on he other it moves north.

Of course, there is the north/south hemisphere ambiguity, but with
additional sensors to the left and right of the meridian line should
help solve this ambiguity. After all, in order to detect meridian
passing you would have to align the device towards true north.

A camera with at least 150 degree field of view pointing directly
upwards towards zenith, should be able to detect the orientation,
latitude, date and local solar time within a year of observations.

 

[Joe Pfeiffer]

Paul Keinanen <keinanen@sci.fi> writes:

On Tue, 09 Mar 2010 10:20:08 -0700, D Yuniskis
not.going.to.be@seen.com> wrote:

Jasen Betts wrote:
So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)? Probably just watch for a longer period of time?

By observing when the sum passes the meridian, one cloud free week in
the spring and one in the autumn should give a quite good resolution
for the latitude, provided that some internal time reference is
capable of measuring the number of days between the measuring periods
with at least +/-12 hour accuracy. During one week long period the sun
moves south and on he other it moves north.

Of course, there is the north/south hemisphere ambiguity, but with
additional sensors to the left and right of the meridian line should
help solve this ambiguity. After all, in order to detect meridian
passing you would have to align the device towards true north.

A camera with at least 150 degree field of view pointing directly
upwards towards zenith, should be able to detect the orientation,
latitude, date and local solar time within a year of observations.

The hardware for this is in the current issue of Circuit Cellar.

Different programming needed....
--
As we enjoy great advantages from the inventions of others, we should
be glad of an opportunity to serve others by any invention of ours;
and this we should do freely and generously. (Benjamin Franklin)

 

[Jasen Betts]

On 2010-03-09, D Yuniskis <not.going.to.be@seen.com> wrote:

Jasen Betts wrote:
So (thinking in terms of a *truly* unique hack), if you *watched*
the motion over the course of a particular day (e.g., 'yesterday'),
could you *uniquely* determine that day?

Given lattitude and longitude (or equivalent) and sufficiently good
instruments, and the right data and skills, yes.

So, a *device* that watched these things could deduce date/time (?)
How much more would it have to do to deduce location (or, at least,
latitude)? Probably just watch for a longer period of time?

6 months would be sufficient,
probably shorter periods too.

the earths's axial wobble, and orbital precession, are probably going
to make it impossible do it in less than a week.

OTOH if you can see the stars and planets at night that would help a
lot...




--- news://freenews.netfront.net/ - complaints: news@netfront.net ---