Hello Helmut,

Following your suggestion, I checked further into the error in the seasonal
marker. You are correct. There is a better date/declination to base it on
than the solstice. There seems to be a minimum error at Declination = 20
degrees. Have a look at the attached chart. The error is reduced by about
20% using the declination of 20 degrees rather than 23.44 as the base. I do
not know why. The new curve of the error shows more but smaller humps. Sigma
of 1.5 cm on the declination table of +/-156 cm is pretty close.

This analysis is based on the calculating standard deviation, sigma. I
calculated the position on the zodiac date line for the analemmatic dial and
then the position calculated from the seasonal marker distances for various
base declinations. I summed the square of the differences, divided by the
number of points and took the square root. This example is based on the
design for Mike Deamicis-Roberts at latitude 36.9. This error analysis is
incomplete but it does confirm your suspicions.

Mike, I recommend that you move the season marker on your dial out by 3.5
cm, from 2.720 meters to 2.755 as this cuts the error by ~20%.

The calculation for the season marker position is as given before but use
declination of 20 degree to calculate the azimuth and the position on the
zodiac table. Then solve for the season marker position based on the
geometry of the triangle.

Roger Bailey
Walking Shadow Designs
N 51  W 115


----- Original Message -----
From: "Roger Bailey" <[EMAIL PROTECTED]>
To: "Sonderegger Helmut" <[EMAIL PROTECTED]>
Sent: Monday, January 14, 2002 5:14 AM
Subject: RE: Seasonal Sunrise Marker


> Hi Helmut,
>
> You make a good point here. Let me look at it and get back to you.
Different
> ways of looking at it give different results. This is the value of this
> mailing list. It allows us to share ideas and develop better ones.
>
> Roger Bailey




Attachment converted: Macintosh HD:SMerror.pdf (PDF /CARO) (0003AC15)

Reply via email to