At an accuracy of 10 seconds, one not only must worry about correct EOT adjustment and proper celestial alignment, but consider that all of the dials that I know, including Bill Gottesman's Renaissance Sundial, are built on strict geometric principles and ignore the refractive effect of the earth's atmosphere (a function of barometric pressure, humidity, and temperature ... as well as the most important term, sun's altitude angle that in turn depends on solar declination and dial latitude).
So we should probably compare sundial accuracy at approximately solar noon and compare either adjusted dial time (mechanically compensating for EOT) or calculated clock time (using EOT table or mathematical equation).
Years ago Bill demonstrated his Renaissance Sundial at the North American Sundial Society's annual conference. Using approximate celestial alignment and EOT adjustment in the hour before noon, the dial was reading within 10 seconds of clock time in the early afternoon. The fact that you could easily read the time to that precision and obtain that accuracy was mind boggling.
Regards Bob --------------------------------------------------- https://lists.uni-koeln.de/mailman/listinfo/sundial
