Dear All,
One of the most interesting (to me!) side issues about the
Hawkshead dial was noted by Patrick Powers:
> On the matter of the dial being designed as a declining
> dial yet also being canted out, it may be of interest
> that there are only 16 dials known to the BSS Register
> which have this property.
I wonder whether the Register rules specify a minimum
"cant"? If not, I shall, sotto voce, reveal a carefully
guarded trade secret:
Almost all vertical dials are canted.
Here's why...
When you have a real client who wants a real sundial on
a real wall you do three things early on:
1. Assume the wall is vertical (this is rash)
2. You note the latitude (this is easy)
3. You estimate the declination (this is difficult)
There then follow months of excitement and frustration
(in equal measure) but you end up with a beautiful dial,
complete with gnomon, ready to fix on the wall.
With modern workshop techniques you have in your hands a
dial that it just about perfect for a vertical wall at the
noted latitude and with the ESTIMATED declination.
You then go back to the site and discover the wall isn't
quite vertical and, via a helpful sun, you find out that
your estimate of declination is a bit wrong. If you have
been careful, it should be correct to about a quarter of a
degree. Even so, if your dial is 1m wide that will mean
packing out one side or the other a little over 4mm.
Of course, you have to pack out the top or bottom if the
wall isn't quite vertical. Necessarily you do all this
when the client isn't looking and keep quiet about it!
And this, dear reader, is why most vertical dials are
canted, even if only a little bit!
Armchair diallists will now say:
1. Hey, surely you can do better than a quarter
of a degree, and...
2. What about painting directly onto the wall where
there is no scope for last-minute canting?
The real-life problem about doing better than a quarter
of a degree is that real walls are actually far from
flat. They have bumps and dips and undulations and a
horizontal line drawn on a wall will not be straight.
The deviations can be well over a quarter of a degree.
If you hold a 2m straight-edge horizontally against a
wall you will find that it nestles against a couple of
peaks that stand a little proud and you typically find
dips of 5mm or more between these peaks.
This is particularly true of old brick walls and, alas,
even of 21st century walls made of stone blocks. If
the wall is rendered you find gentle undulations.
If you hold a 1m square piece of truly-flat slate
against a real wall, it will typically rest against
three local peaks. These define a plane of course
but it may deviate significantly from the best-fit
vertical plane which is what an expensive surveyor
will come up with.
Another real-life problem is that the client, at
the very last minute, will say: "I think it would
look better a little higher up". You then find
you are resting against three different peaks.
In some ways cutting directly into a wall (or
painting on it) is actually easier. There you
really can use the best-fit vertical plane and
ALLOW for the deviations as you are setting the
furniture out.
Essentially, you regard the wall as being sectioned
into little squares and treat each square as having
its own separate nodus height. In a bad case, these
separate heights may deviate +/- 5mm from the nodus
height relative to the best fit vertical plane.
Other practitioners will no doubt tell their own
stories. I once had a client who suggested a quite
different wall on the day of fixing!
It's all good fun in the end!
Frank King
Cambridge, U.K.
---------------------------------------------------
https://lists.uni-koeln.de/mailman/listinfo/sundial
