Dave Bell wrote: > > John, > > > > If the gnomon is outside and the dial face is inside, with a pane of > > glass in between, isn't there a refraction problem? > > > > Mac > > There is a refraction offset introduced, but can't you compensate for > that? I suppose it would complicate the layout calculations somewhat, > since the offset varies with incidence angle... > > Dave
My recollection of the laws of physics leads me to believe that the effect of refraction is much simpler than you would suppose. It is exactly as though the glass were a thinner 'slice' of air. If you know the thickness (t) and refractive index (u) of the glass you just imagine the two surfaces were separated, not by a distance t filled with glass, but a distance t/u filled with air. (That's t divided by u.) So, the only practical effect for a normal vertical dial is that the gnomon must be positioned slightly differently relative to the dial. This can be checked by eye, as the gnomon must appear, when viewed through the glass, to be in line with the centre of the dialface (and vice versa). If the gnomon has a nodus, the scale of the dial must treat it as though its height is its actual distance from the glass plus t/u. For a double-glazed window, you can apply the same logic. As someone mentioned, sealed double-glazed units bow with temperature changes, so the gnomon should be fixed to the centre of the glass to minimise this effect. Hope this helps Chris -