11
$\begingroup$

I had a horizontal sundial made specially for my latitude (approx. 38° N). The default correct way to orient a sundial (in the northern hemisphere) is with the gnomon pointing towards polar north. The caveat is that my latitude (approx. 122° W) causes "clock noon" (the time when a clock on my wall indicates noon) is approximately 8 minutes behind solar noon (the time when a sundial pointing polar north indicates noon), hence the time indicated by the sundial will never match clock time, not even on days when the Equation of Time (EoT) is zero.

If I were to instead orient the sundial such that the gnomon pointed at noon when it's noon according to my clock (which means the sundial would be rotated a few degrees counter-clockwise from polar north) on an EoT zero day, would that be "better" overall or make discrepancies worse overall? Or should I just orient the gnomon to polar north?

$\endgroup$
11
  • 2
    $\begingroup$ I had to check the Help page to see if this is on-topic, but I think it just scrapes in under the convoluted "unless" exemption to Earth Science questions. ;-) $\endgroup$ Commented Jul 18 at 4:17
  • 4
    $\begingroup$ @ChappoHasn'tForgotten I did do a search for "sundial" prior to posting my question and saw there were other questions about them, so I figured it was OK. $\endgroup$ Commented Jul 18 at 13:27
  • 3
    $\begingroup$ If you want it to read clock time, you have to have one specifically made for your longitude also. The gnomon has to be aligned with the Earth's axis, moving it 6 minutes for one time will not make it offset by 6 minutes at all times, the offset will get greater the longer the shadow is. Since your sundial is already made, one solution is to just have a static offset you always add to the equation of time, since you always have to adjust for the EoT anyway. $\endgroup$ Commented Jul 18 at 14:54
  • 1
    $\begingroup$ Slightly off topic, but probably helpful advice on orienting the sundial. Most tutorials you see involve a compass, or the North Star, etc. The easy way to orient the sundial is to just set it to the correct time (after accounting for the EoT and longitude offset). That method is both easier, and more accurate than the compass/star methods. $\endgroup$ Commented Jul 18 at 14:56
  • 2
    $\begingroup$ @ChappoHasn'tForgotten The Sundial is one of the oldest astronomical instruments. Questions about sundial theory and construction are generally on-topic here. $\endgroup$
    – PM 2Ring
    Commented Jul 19 at 8:31

2 Answers 2

20
$\begingroup$

A sundial needs to be aligned with the planet's axis of rotation, with the gnomon pointed in towards the nearest pole, for it to work correctly. The point of the shadow will then give you your Local Time (LT).

As you've discovered, LT only coincides with "clock time" (i.e. your time zone when there's no daylight saving) if your longitude exactly corresponds to the UTC offset for that time zone. This is somewhat rare. Each degree east or west of that exact UTC-offset longitude creates a 4-minute discrepancy between local time and clock time.

The problem is that if you align your gnomon so that local noon equals clock noon, it will no longer be aligned with the Earth's axis and the "mean time" curve imprinted on your sundial may no longer match the diurnal path of the gnomon's shadow. In effect, you've distorted the conic section. You can picture this distortion more easily if you imagine exaggerating the adjustment: say, 90 degrees (eg gnomon pointing due west).

A 1.5 degree distortion is probably trivial, but some would say that a 6-minute discrepancy between LT and clock time is equally trivial (especially when the EoT discrepancy can be much more than that).

A much more accurate solution is possible. If your sundial is correctly oriented, the point of the gnomon's shadow will closely follow the mean time line on the sundial's horizontal plane on "EoT zero day", with the discrepancy from "clock time" remaining more or less consistent throughout the day. Since the problem is not the device but the time labels, the solution is to change the labels. You could use a permanent marker to incorporate the six-minute delay, so that (for example) either the label for the noon mark on the sundial reads 11:54am, or a new mark labelled 12:00pm is placed at the 12:06pm point on the sundial.

$\endgroup$
5
  • 3
    $\begingroup$ Note that most permanent marker inks aren't very photo-stable (e.g. black Sharpie is not a good idea for garden plant labels as it will fade in a few weeks). So that would make it a good temporary test rather than a permanent modification. Except you can't trust that either on materials like brass, because they can leave a region that doesn't weather the same as the rest $\endgroup$
    – Chris H
    Commented Jul 18 at 10:48
  • 1
    $\begingroup$ Unless the dial has different scales for different seasons or months, writing static offset on it does not even make sense, does it? Solar noon oscillates up to 16 minutes around the clock noon during a year. $\endgroup$
    – Edheldil
    Commented Jul 18 at 12:04
  • 4
    $\begingroup$ The hour lines are static offsets already. Back when sundials were invented, timezones didn't exist, so none of this mattered. But with timezones, I think what @ChrisH is saying is that to make a "modern" sundial that takes timezones into account, not only do you need to draw the hour lines at offsets for a specific latitude, but you could rotate just the dial plate (leaving the gnomon pointing north) to compensate for your longitude and offset within your timezone from its 0 point. $\endgroup$ Commented Jul 18 at 13:36
  • 4
    $\begingroup$ I mean heck, time zones were only invented in the 1880s to make train scheduling less of a nightmare. Used to be every town just set their clocks by local noon and the difference didn't matter much unless you were trying to coordinate across long distances. $\endgroup$ Commented Jul 18 at 19:15
  • 1
    $\begingroup$ Historical trivia: telegraphs predated long-distance trains in Australia, so time zones were implemented specifically to suit inter-state telegraph commerce. $\endgroup$
    – david
    Commented Jul 19 at 4:16
10
$\begingroup$

There's another possible solution: tilt the sundial's dial plate so that it aligns with the earth surface at the "nominal longitude" for your timezone, while the gnomon still is aligned with the axis of rotation.

Why does that work?

If you use your sundial at a place where local time and clock time match (center meridian of a timezone), it shows the correct time. Of course, this does not magically know where you are on the globe, but just depends on the geometry of the dial plate and the gnomon relative to the sun and the axis of rotation.

So, wherever on Earth you mount your sundial, if you orient the gnomon and the dial plate exactly the way you'd do at the "nominal" location, it will show the same "sundial time". As the earth isn't flat, when installing the sundial in a different location, you'd need to compensate for the curvature of the surface.

How to find the correct orientation of the dial plate and gnomon?

You can calculate the necessary tilt, but that involves quite some trigonometry (see EDIT below).

Or you can calibrate it on a sunny day:

  • As you're living a bit west of your timezone's center meridian, by 2 degrees, lift the west end of the plate by 1.5 degrees. E.g. use wedges to temporarily hold that position. Then rotate the dial plate until it shows the correct clock time.
  • Check the accuracy at different times of day and adjust lift and rotation until you're satisfied.
  • The gnomon's edge (not the 12 o'clock line of the dial plate) should always point to the north and have an upward angle corresponding to your latitude of 38° N, if you can manage to measure that.
  • When you're satisfied, mount the sundial permanently.

Don't expect too much: the sun's diameter always causes a blurry shadow instead of a sharp edge. That means that you can't reliably read out times better than maybe two minutes.

EDIT (Here's the math as per OP's comment):

You want to rotate the sundial aound the gnomon's axis by the angle $\lambda$ that you have to compensate (i.e. 2° counterclockwise for your 2° west of the center meridian).

So, you can move the west end of the dial plate up and south, in a direction perpendicular to the gnomon's axis, by an amount so that the overall rotation is exactly what you want to compensate.

  • Let $\lambda_{rel}$ be the longitude offset (west = positive).
  • Let $\phi$ be your latitude (north = positive).
  • Use a fixed point at the east end of the dial plate.
  • Find its west counterpart. Let $L$ be the distance between these east and west points.
  • Lift the west point up by $L \cdot \sin \lambda_{rel} \cdot \cos \phi$.
  • Move the west point southwards by $L \cdot \sin \lambda_{rel} \cdot \sin \phi$ (i.e. rotate the dial plate)
  • Make sure the north-south axis of the dial plate is still horizontal (both ends at the same height). That is not universally correct, but for deviations of a few degrees it's a very good match.

If you're east of the nominal meridian, you'll get negative values for the west-end movement, implying you should move the west end down and north. But you can of course instead hold the west end, and move the east end up and south.

If you're on the southern hemisphere, you'll get negative move-south values, meaning to move north instead.

$\endgroup$
8
  • 1
    $\begingroup$ In my original post, I confused a number: yes, it's an 8-minute difference, not 6. (I fixed my post.) As for your suggestion, it's interesting. I have no problem using trigonometry. Can you point me at a source for an equation to calculate the tilt? $\endgroup$ Commented Jul 19 at 13:34
  • 1
    $\begingroup$ I assume all angles have to be converted to radians first, right? Assuming so, if I plug in the numbers 𝜆𝑟𝑒𝑙 = 2° (.0349 rad), 𝜙 = 37.7861° (.6595 rad), 𝐿 = 9.5", I get: Wup = .262", Ws = .203". Does that seem right? $\endgroup$ Commented Jul 19 at 18:31
  • 2
    $\begingroup$ +1 This is not "another possible solution". This is THE solution, all the other answers are "other possibles". $\endgroup$ Commented Jul 20 at 6:35
  • 1
    $\begingroup$ @PaulJ.Lucas Yes, that looks correct. And "all angles have to be converted to radians": that depends on your calculator, whether it supports degree-based trigonometry - pocket calculators usually do, programming languages normally not. $\endgroup$ Commented Jul 21 at 10:31
  • $\begingroup$ @RalfKleberhoff Your next to last bullet says to move the west point southwards, but in the paragraph that follows, you said "... instead of moving the west end down and north, ...". These seem to be contradictory. I think the bullet is wrong, i.e., the west point needs to be moved north when west of the meridian. Which is correct? $\endgroup$ Commented Jul 24 at 14:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .