sundial program
ne...@rtrussell.co.uk
> If he wants my QBasic source, I'll append it below. However, you did
> make some valuable changes, such as making the printing routine work
> with a modern printer.
Do you happen to have a 'precision' sundial program which takes
account of the thickness of the gnomon? The correction is non-
trivial, because you have to divide the dial into four quadrants and
shift them with respect to each other.
I'm particularly asking because I've made a surprising discovery about
my sundial (a relatively expensive, supposedly accurately calibrated,
model from Haddonstone). It has an unusual design of gnomon, where
the edge that you use to make the reading is actually the *underside*:
The effect of this is that the compensation for the thickness of the
gnomon must be made in the *opposite direction* from a conventional
(e.g. triangular) gnomon, however Haddonstone appear to be completely
unaware of this! It is therefore impossible to adjust the sundial
accurately: if it is set to be accurate in the morning it is about 20
minutes slow in the afternoon, and if set to be accurate in the
afternoon it is about 20 minutes fast in the morning.
If Haddonstone hadn't bothered with the gnomon thickness 'correction'
at all it would be considerably more accurate! Very annoying!!
Richard.
http://www.rtrussell.co.uk/
To reply by email change 'news' to my forename.
Fred McKenzie
<12e074b8-4c01-41b5...@q2g2000vbr.googlegroups.com>,
"ne...@rtrussell.co.uk" <ne...@rtrussell.co.uk> wrote:
> Do you happen to have a 'precision' sundial program which takes
> account of the thickness of the gnomon? The correction is non-
> trivial, because you have to divide the dial into four quadrants and
> shift them with respect to each other.
Richard-
I have a publication from the U. S. Department of Commerce dated 1933,
entitled, "Sundials". It gives a simple graphical method of
constructing a sundial. I don't think it is very sophisticated, since
it is for a vertical gnomon rather than one that is parallel to the
earth's axis.
With regard to the thickness of the gnomon, there is a morning half and
an afternoon half of the scale, separated by the thickness of the
gnomon. If your program can calculate an accurate half of the scale,
the other half would be the mirror image just spaced at the center.
Fred
Richard Russell
> With regard to the thickness of the gnomon, there is a morning half and
> an afternoon half of the scale, separated by the thickness of the
> gnomon.
No, I don't believe that's right. The compensation for the thickness
of the gnomon changes in direction at local 'midday' (shadow due
north) as you say, but it also changes when the shadow passes due east
and due west (which can certainly happen in the summer). So my
understanding is that you actually have to split the scale into *four*
(not two) pieces.
This page shows graphically what you need to do, but I would still be
interested in a program which does it automatically and accurately:
http://www.mysundial.ca/tsp/wide_gnomon.html
Note that, as I explained, with my (Haddonstone) sundial you don't
*separate* the morning and afternoon scales by the thickness of the
gnomon, you *move them together* by the thickness of the gnomon! The
above web page describes that situation, too (pity Haddonstone didn't
read it!).
william...@gmail.com
wrote:
> On Feb 26, 3:54 am, david.willi...@bayman.org (David Williams) wrote:
>
> > If he wants my QBasic source, I'll append it below. However, you did
> > make some valuable changes, such as making the printing routine work
> > with a modern printer.
>
> Do you happen to have a 'precision' sundial program which takes
> account of the thickness of the gnomon?
On the best sundials I have seen, the gnomon tapers to a knife-edge at
the top. This edge casts the edge of the shadow, except extremely
close to noon. Even this limitation would be avoided if the gnomon
tapered to a point at its tip.
dow
Richard Russell
> Even this limitation would be avoided if the gnomon
> tapered to a point at its tip.
I would have thought it would be very difficult to achieve adequate
strength and rigidity of the gnomon in that case. I see no objection
to having a 'thick' gnomon so long as the dial is correctly
compensated for the thickness (which is where we came in). If you
don't know of a program which can draw an accurate sundial with the
appropriate compensation, I might have to write one myself!
winston19842005
83dc885b-04cd-4ef2...@r13g2000vbr.googlegroups.com, "Richard
Russell" <ne...@rtrussell.co.uk> wrote:
> On Jun 9, 3:26�pm, williamsdavi...@gmail.com wrote:
>> Even this limitation would be avoided if the gnomon
>> tapered to a point at its tip.
>
> I would have thought it would be very difficult to achieve adequate
> strength and rigidity of the gnomon in that case. I see no objection
> to having a 'thick' gnomon so long as the dial is correctly
> compensated for the thickness (which is where we came in). If you
> don't know of a program which can draw an accurate sundial with the
> appropriate compensation, I might have to write one myself!
I am enjoying this thread, however I'd really like to know where to purchase
sundials? I'm not so much interested in their accuracy, more in their
artistic value.
I'm really surprised at how hard they are to find locally. I live in south
Georgia, US, and we have Hobby Lobby stores that have everything similar BUT
alas, no sundials...
When I was a kid in elementary school I made a sundial out of cardboard. I
remember reading about them in our encyclopedia. Fell in love with them, but
have never owned one. Always wanted one.
dow
> On Jun 9, 3:26 pm, williamsdavi...@gmail.com wrote:
>
> > Even this limitation would be avoided if the gnomon
> > tapered to a point at its tip.
>
> I would have thought it would be very difficult to achieve adequate
> strength and rigidity of the gnomon in that case. I see no objection
> to having a 'thick' gnomon so long as the dial is correctly
> compensated for the thickness (which is where we came in). If you
> don't know of a program which can draw an accurate sundial with the
> appropriate compensation, I might have to write one myself!
>
> To reply by email change 'news' to my forename.
Hmmm... Common table-knives are pretty strong and rigid. And they're
usually made of stainless steel, which would last indefinitely
outdoors.
One of the limitations of sundials is due to the non-zero size of the
sun in the sky. It's angular diameter is a bit more than half a
degree. Shadows, such as that of a gnomon, therefore have penumbral
areas half a degree wide. People usually perceive the edge of the
shadow to be not in the centre of the penumbra but close to its "dark"
edge. The reading of a sundial is herefore usually wrong by something
like a quarter of a degree, which corresponds to one minute of time.
In the morning, the reading is one minute ahead of true solar time, in
the afternoon behind.
Sure. Write a program the compensate for thick gnomons. You'll have to
include some variability, of course. Is the cross-section of the top
of the gnomon semicircular, or rectanglar, or what?
dow
dow
wrote:
> On 6/10/09 6:25 PM, in article
Around here (Toronto, Canada), sundials are sold in garden nurseries,
along with things like statues that can be put in gardens to beautify
them. Sundials are used for the same purpose.
I have bought a couple of them, but found that they were pretty
useless as timekeepers. The gnomons were not at the right angle for
this latitude, and the markings on the dials were badly laid out. I
got the impression that they were intended to be admired esthetically,
but not used to tell the time.
If you want a good sundial, you'll probably have to make it yourself.
You might, of course, be able to find an antique sundial, dating from
the time when they were used as practical timekeepers. But you may
have to buy the building that it's attached to!
dow
Fred McKenzie
winston19842005 <bjjl...@NOSPAMbellsouth.net> wrote:
> I'm really surprised at how hard they are to find locally.
Same here in Florida. I may have seen one in a garden store many years
ago, but it was probably for looks and not very functional.
Over 50 years ago, I had a Boy Scout pocket sundial. I don't know what
happened to it. I'll look for it next time I'm back at the old
homestead.
Searching the web, I came across a variety of sundials made of brass.
See <http://www.brasscompass.com>. The company is located in California.
Fred
Nick Cramer
> On Jun 10, 6:25=A0pm, Richard Russell <n...@rtrussell.co.uk> wrote:
> > On Jun 9, 3:26=A0pm, williamsdavi...@gmail.com wrote:
> >
> > > Even this limitation would be avoided if the gnomon
> > > tapered to a point at its tip.
> >
> > I would have thought it would be very difficult to achieve adequate
> > to having a 'thick' gnomon so long as the dial is correctly
> > don't know of a program which can draw an accurate sundial with the
> > appropriate compensation, I might have to write one myself!
> Hmmm... Common table-knives are pretty strong and rigid. And they're
> usually made of stainless steel, which would last indefinitely
I immediately thought of my wife's kitchen knives, too. ;-)
--
Nick, KI6VAV. Support severely wounded and disabled Veterans and their
families: https://www.woundedwarriorproject.org/ Thank a Veteran!
Support Our Troops: http://anymarine.com/ You are not forgotten.
Thanks ! ! ~Semper Fi~ USMC 1365061
Richard Russell
wrote:
> I am enjoying this thread, however I'd really like to know where to purchase
> sundials?
The company from which I bought mine (Haddonstone) do have a presence
in the US:
Haddonstone (USA) Ltd, 201 Heller Place, Bellmawr, NJ 08031
Telephone: 856 931 7011 Fax: 856 931 0040
Haddonstone (USA) Ltd , 32207 United Avenue, Pueblo, CO 81001
Telephone: 719 948 4554 Fax: 719 948 4285
However whilst their sundials are attractive (with a price to match) I
can hardly recommend them given the issue that started this thread.
Mind you, more customers compaining about the fault with their
sundials might provide more incentive to fix it!
Richard Russell
> Hmmm... Common table-knives are pretty strong and rigid.
I guess it comes down to a choice between functionality and
attractiveness. Sure, a thin stainless-steel gnomon might be strong
enough, but it wouldn't look very good with my dial.
> People usually perceive the edge of the shadow to be not in the
> centre of the penumbra but close to its "dark" edge.
Something else to compensate for in a 'precision' sundial, then!
Really, unless you're going to bother to take account of the Equation
Of Time whenever you read it, a conventional sundial is never going to
be 'accurate', but I do strongly object to having one that has been
compensated for the gnomon thickness *in the wrong direction*.
Richard.
http://www.rtrussell.co.uk/
dow
Here in Canada, we have a one-dollar coin that is made of some alloy
that is strong and wear-resistant. When the coins are new they look
like gold, but after a while their colour resembles bronze. I don't
know what the alloy is, but I'd guess it would be good for making
gnomons. Maybe the Royal Canadian Mint would tell you. However,
they're distracted right now because they seem to have lost a pile of
gold...
Good sundials have some sort of table or graph that shows the Equation
of Time, and instructions for using it. Also, of course, there has to
be a correction for longitude, unless the dial happens to be right on
the central meridian of the time-zone it's in.
This is, of course, the reason why good sundials ae hard to buy. For a
dial to tell good time, agreeing with a clock (after the Equation of
Time has been compensated for), it must be made for the exact location
in which it will be used, taking both latitude and longitude into
account. Manufacturers can't just churn out a whole lot of identical
dials and have them work well in different locations.
dow
dow
> Do you happen to have a 'precision' sundial program which takes
> account of the thickness of the gnomon? The correction is non-
> trivial, because you have to divide the dial into four quadrants and
> shift them with respect to each other.
I've thought of two more simple cases, besides that of having a knife-
edge at the top of the gnomon. One is if the top of the gnomon has a
rectangular cross-section, so there are two sharp edges, the western
one of which casts the edge of the gnomon's shadow in the morning, and
the eastern one in the afternoon. These can be thought of as two
linear gnomons, so the dial plate must consist of two halves shifted
apart. If the dial that my program draws is cut along its north-south
axis and the two halves are moved apart by a distance equal to the
thickness of the gnomon, the resulting dial will be compensated for
the gnomon's thickness.
The other case is if the gnomon is a narrow cylinder, mounted so its
axis is positioned where a linear gnomon would be, and the dial is
identical to the one my program draws. The shadow of the gnomon will
be a "stripe", and the person reading the time from the dial must
judge the position of the mid-line of the stripe. If the gnomon is
narrow, this should be possible quite accurately. A nice feature of
this arrangement is that it also compensates for the error produced by
the non-zero diameter of the sun in the sky!
We have been discussing only dials with a flat horizontal plate. There
are plenty of other types. One quite common one has the hour-marks on
a cylindrical surface mounted so the axis of the cylinder coincides
with the narrow rod that serves as the gnomon. This means that the
hour-marks are equally spaced, 15 degrees apart. This design has
several advantages. It works fine close to the equator, where the flat-
plate design is useless. Also, adjusters can easily be incoporated so
the same hardware can be used in different latitudes, and also to
compensate for the Equation of Time. The device can be adjusted every
few days, rotating the cylinder slightly about its axis. Once
adjusted, it shows the correct time directly, without any calculation
being required. The same adjustment also compensates for the device's
longitude relative to the prime meridian of the time zone. The
thickness of the gnomon and the diameter of the sun are compensated
for just by judging the mid-point of the shadow when the dial is read.
Fun stuff!
dow
Richard Russell
> One is if the top of the gnomon has a
> rectangular cross-section, so there are two sharp edges, the western
> one of which casts the edge of the gnomon's shadow in the morning, and
> the eastern one in the afternoon. These can be thought of as two
> linear gnomons, so the dial plate must consist of two halves shifted
> apart.
That's not what this site says:
http://www.mysundial.ca/tsp/wide_gnomon.html
There it says, for the case you describe, that you split the dial into
*four* (not two) parts; as I said previously (see Figure 3). It also
specifically deals with the case in which I am interested (Figure 6).
The article also mentions the case when the sundial is inside the
Arctic or Antarctic circle, so that it may be illuminated by the sun
for 24 hours a day during certain periods of the year. That situation
proves, I think, that your proposed dial plate consisting of "two
halves shifted apart" must be incorrect.
So I am currently using that site as my 'reference' for sundials with
thick gnomons. In particular Figure 6 illustrates the type of dial
that my Haddonstone sundial *should* have. I would certainly be very
interested if you believe it to be wrong.
dow
> To reply by email change 'news' to my forename.
No. On further consideration, I think it's right. There are basically
two possibilities, that the eastern edge of the top of the gnomon
casts the edge of the shadow, or the western one. Each possibility is
associated with its own set of hour lines on the plate. The western
edge casts the edge of the shadow between 6 am and noon, local solar
time. At noon, the sun passes through the plane of the gnomon, and the
eastern edge then becomes functional. But at 6 pm the sun passes
through the plane of the top of the gnomon, i.e. the plane including
the two edges. The eastern edge then passes into shade, and the
western edge again casts the edge of he gnomon's shadow. At midnight
and 6 am, if the sun is above the horizon, the edges switch again. So
the hour lines on the north-eastern part of the dial are continuous
with those on the south-western part, and ditto for the other two
parts. Horribly, there is a slight overlap at midnight.
Frankly, this fails to fascinate me. If I were building a sundial, I
would give the gnomon a knife-edge at the top, or I would use a
cylindrical surface for the hour markings.
The sundial I made in Chile used a piece of plastic I made by cutting
a cylindrical drum in half. I fixed a rod, the gnomon, along the axis
of the cylinder, and marked the hour lines from 6 am to 6 pm at
intervals of 15 degrees on the inside surface of the cylinder. Then
all I had to do was align the thing properly, and it worked. I suspect
it was, and maybe still is, the only functioning sundial in the
country.
I just watched a TV program about atomic clocks that can keep time to
better than one part in 10^16. And we're fussing over sundials....
dow
Richard Russell
> I just watched a TV program about atomic clocks that can keep time to
> better than one part in 10^16. And we're fussing over sundials....
Ah, well, funny you should say that. I have a large number of
'atomic' (i.e. radio-controlled) clocks in and around this house,
something like 12 at the last count! That includes two wristwatches
and one 'outdoor' clock visible from where the sundial is.
The trouble with these 'atomic' timepieces (all receiving the 60 kHz
MSF transmission, but there are several other transmitters around the
world) is inadequate error checking in their software. So whilst most
of the time they are precisely accurate (to a fraction of a second)
occasionally they can be ten minutes, or an hour, or several hours out
owing to RF interference. Basically you can never trust them unless
you have another clock to compare with.
This is of course a design fault, and not inherent in 'atomic clocks'
in general, but sundials don't suffer from that kind of problem!
Richard.
Ralph
"Richard Russell" <ne...@rtrussell.co.uk> wrote in message
news:d46916d6-0cff-46db...@a36g2000yqc.googlegroups.com...
--------------------------------------------------------
"Basically you can never trust them unless you have another clock to compare
with."
Actually you need three clocks. There is an old 18th century truism - "Never
go to sea with two clocks. Take one or three." lol
-ralph
dow
> "Richard Russell" <n...@rtrussell.co.uk> wrote in message
>
> news:d46916d6-0cff-46db...@a36g2000yqc.googlegroups.com...
> On Jun 14, 11:22 pm, dow <williamsdavi...@gmail.com> wrote:
>
> > I just watched a TV program about atomic clocks that can keep time to
> > better than one part in 10^16. And we're fussing over sundials....
>
> Ah, well, funny you should say that. I have a large number of
> 'atomic' (i.e. radio-controlled) clocks in and around this house,
> something like 12 at the last count! That includes two wristwatches
> and one 'outdoor' clock visible from where the sundial is.
>
> The trouble with these 'atomic' timepieces (all receiving the 60 kHz
> MSF transmission, but there are several other transmitters around the
> world) is inadequate error checking in their software. So whilst most
> of the time they are precisely accurate (to a fraction of a second)
> occasionally they can be ten minutes, or an hour, or several hours out
> owing to RF interference. Basically you can never trust them unless
> you have another clock to compare with.
>
> This is of course a design fault, and not inherent in 'atomic clocks'
> in general, but sundials don't suffer from that kind of problem!
>
> To reply by email change 'news' to my forename.
>
> --------------------------------------------------------
>
> "Basically you can never trust them unless you have another clock to compare
> with."
>
> Actually you need three clocks. There is an old 18th century truism - "Never
> go to sea with two clocks. Take one or three." lol
>
> -ralph
I recently noticed that my cellphone ("mobile" in the UK) interferes
with digital TV reception. I wonder what it would do to your clocks!
Sundials are also prone to interference, of course, by clouds.
Taking two chronometers to sea would not be useful. If they agree, the
second one is redundant. If they disagree, there is no way to tell
which is right.
It's common practice in research labs to do an experiment twice, to
guard against errors. If the two results disagree, the experiment is
done a third time. Hopefully, two of the three results will be in
agreement, and the erroneous one can be discarded.
dow
Richard Russell
> I recently noticed that my cellphone ("mobile" in the UK) interferes
> with digital TV reception. I wonder what it would do to your clocks!
There's rather a large difference between the near-microwave
frequencies that cellphones work at (I've always called them that, not
'mobiles') and the very low frequencies of the standard time and
frequency transmissions, so the likelihood of mutual interference is
small if the clock is competently designed. The main source of
interference to the 60 kHz signal is computer monitors (it's close to,
if not exactly the same as, common line-timebase frequencies).
> Sundials are also prone to interference, of course, by clouds.
A cloud cannot make a sundial read the 'wrong' time, it simply doesn't
read any time. This is a fundamental difference.
> Taking two chronometers to sea would not be useful.
That is an invalid analogy. What I am interested in knowing is
whether the 'atomic' clock is precisely right - which it usually will
be - or out by a relatively large amount (say at least ten minutes).
I don't need two other clocks to make that determination, I need only
one other clock that is known to be approximately correct, such as a
wristwatch or a regular quartz clock.
Richard.
dow
> On Jun 15, 3:48 pm, dow <williamsdavi...@gmail.com> wrote:
>
> > I recently noticed that my cellphone ("mobile" in the UK) interferes
> > with digital TV reception. I wonder what it would do to your clocks!
>
> There's rather a large difference between the near-microwave
> frequencies that cellphones work at (I've always called them that, not
> 'mobiles') and the very low frequencies of the standard time and
> frequency transmissions, so the likelihood of mutual interference is
> small if the clock is competently designed. The main source of
> interference to the 60 kHz signal is computer monitors (it's close to,
> if not exactly the same as, common line-timebase frequencies).
I'm not sure how my cellphone interferes with reception on my digital
TV. It happens no matter what TV channel I'm watching, so I suspect
the interference affects some internal process in the digital decoder.
I suppose it's conceivable that the phone could interfere with any
other digital device, but I've never noticed the interference anywhere
else.
> > Sundials are also prone to interference, of course, by clouds.
>
> A cloud cannot make a sundial read the 'wrong' time, it simply doesn't
> read any time. This is a fundamental difference.
Actually, weather conditions can sometimes make a sundial show the
wrong time. For example, scattering of light by high clouds of ice
crystals can produce an apparent image of the sun in the sky, far from
the sun's real position. If the real sun is hidden, the image can make
a sundial show the wrong time. It's rare, but it can happen. Much more
commonly, every day in fact, refraction of sunlight in the atmosphere
displaces the apparent position of the sun. Near sunrise and sunset,
this can make a sundial wrong by several minutes.
> > Taking two chronometers to sea would not be useful.
>
> That is an invalid analogy. What I am interested in knowing is
> whether the 'atomic' clock is precisely right - which it usually will
> be - or out by a relatively large amount (say at least ten minutes).
> I don't need two other clocks to make that determination, I need only
> one other clock that is known to be approximately correct, such as a
> wristwatch or a regular quartz clock.
But if the only clocks you had available were your atomic ones, you
could compare the readings of three of them to determine which was
wrong. Of course, this assumes that several clocks are unlikely to be
in error at the same time.
dow
Ralph
That is an invalid analogy. What I am interested in knowing is
whether the 'atomic' clock is precisely right - which it usually will
be - or out by a relatively large amount (say at least ten minutes).
I don't need two other clocks to make that determination, I need only
one other clock that is known to be approximately correct, such as a
wristwatch or a regular quartz clock.
=====================================
Stay on land.
-ralph
Richard Russell
> I'm not sure how my cellphone interferes with reception on
> my digital TV.
Usually it's due to overloading in one of the early RF stages.
Insufficient selectivity before the first non-linear device (e.g. FET
or bipolar transistor) can allow through frequencies well away from
those of the digital TV signal, resulting in cross-modulation (good
analogue RF filters are expensive, so are rarely fitted).
> It happens no matter what TV channel I'm watching, so I suspect
> the interference affects some internal process in the digital decoder.
Once it's past the ADC there should be little opportunity for
interference, unless the digital circuitry is very poorly designed.
> I suppose it's conceivable that the phone could interfere
> with any other digital device, but I've never noticed the
> interference anywhere else.
Actually it's far more likely that it will interfere with an
*analogue* device (which your 'digital' TV receiver is up until the
ADC). Any device using low-amplitude signals is vulnerable,
especially when screening and filtering are not up to scratch. That's
one reason why hospitals usually require you to switch off cellphones,
because equipment like EEG and ECG recorders can be affected.
> Actually, weather conditions can sometimes make a sundial
> show the wrong time.
A predictable response, but it's nevertheless a completely different
situation. As you pointed out yourself earlier in the thread, the
sun's angular diameter is only about half a degree, so it's most
unlikely that any 'false sun' will cast a comparably sharp shadow.
Even if it did, you'd only have to look up to notice that something
was out of the ordinary!
> Of course, this assumes that several clocks are unlikely to be
> in error at the same time.
Which would be a very silly assumption, given that they are all
receiving the same 60 kHz signal. If you're going to compare three
clocks, make sure they are genuinely independent!
Richard.
http://www.rtrussell.co.uk/
To reply by email, change 'news' to my forename.
dow
> Actually it's far more likely that it will interfere with an
> *analogue* device (which your 'digital' TV receiver is up until the
> ADC). Any device using low-amplitude signals is vulnerable,
> especially when screening and filtering are not up to scratch. That's
> one reason why hospitals usually require you to switch off cellphones,
> because equipment like EEG and ECG recorders can be affected.
I haven't noticed interference from the cellphone on any other
devices, analogue or digital. However, I haven't run any deliberate
tests. I noticed the interference on the TV by accident.
Incidentally, around here, hospitals no longer forbid the use of
cellphones, except sometimes in areas close to sensitive life-support
equipment.
> > Actually, weather conditions can sometimes make a sundial
> > show the wrong time.
>
> A predictable response, but it's nevertheless a completely different
> situation. As you pointed out yourself earlier in the thread, the
> sun's angular diameter is only about half a degree, so it's most
> unlikely that any 'false sun' will cast a comparably sharp shadow.
> Even if it did, you'd only have to look up to notice that something
> was out of the ordinary!
I have seen what looked very like two suns in the sky, and the only
obvious way to tell which was the real one was that it was brighter
than the other. I wasn't using a sundial at the time, but I think it
is plausible that, if the real sun were obscured by low cloud for
example, someone might take a false reading from a sundial without
realizing that anything was wrong.
Do you always check that the "sun" in the sky is the real one when you
use a sundial?
dow
dow
On Jun 1, 4:56 am, "n...@rtrussell.co.uk" <n...@rtrussell.co.uk>
wrote:
>
> I'm particularly asking because I've made a surprising discovery about
> my sundial (a relatively expensive, supposedly accurately calibrated,
> model from Haddonstone). It has an unusual design of gnomon, where
> the edge that you use to make the reading is actually the *underside*:
>
> http://tinyurl.com/mr6yg4
>
> The effect of this is that the compensation for the thickness of the
> gnomon must be made in the *opposite direction* from a conventional
> (e.g. triangular) gnomon, however Haddonstone appear to be completely
> unaware of this! It is therefore impossible to adjust the sundial
> accurately: if it is set to be accurate in the morning it is about 20
> minutes slow in the afternoon, and if set to be accurate in the
> afternoon it is about 20 minutes fast in the morning.
>
> If Haddonstone hadn't bothered with the gnomon thickness 'correction'
> at all it would be considerably more accurate! Very annoying!!
I took a look at the brochure from Haddonstone. I agree that the
thickness correction must be made in the opposite direction from usual
because the lower edge of the gnomon is used to tell the time.
Did you actually buy one of these things?! Or have you obtained more
information than is contained in the brochure?
I'd like to think that the lower edge of he gnomon is a knife-edge.
This might be a bit safer than having a sharp edge on the top. Fewer
kids might lose fingers. And no correction would be needed for the
thickness of the gnomon.
However, as far as I can see, all of these dials with flat plates are
going to be hopelessly inaccurate in many locations because they
aren't set up for the correct latitude. It looks to me like
Haddonstone makes two models, one for the U.K and the other for the
U.S.A. Probably, the British one is set for about 55 degrees north,
and thw American one for about 35 N. But, even ignoring Alaska and
Hawaii, the U.S. is about 25 degrees of latitude wide, from southern
Florida to the western part of the Canadian border. There are places
where the 35-degree sundial will be 10 or 15 degrees wrong, which will
have a catastrophic effect on the accuracy of the dial, especially at
approximately 9 am and 3 pm. The readings, I think, will be the better
part of an hour wrong. To a smaller extent, the British dial will be
wrong in southern England and northern Scotland. These inaccuracies
immensely outweigh those due to the thickness of the gnomon, the
diameter of the sun, atmospheric refraction, etc..
Basically, these sundials are toys, or maybe decorative sculptures.
They are not timepieces.
There are two sundials in the brochure that have the hour markings on
cylindrical surfaces: the armillary sphere and the "crescent" dial. In
the armillary sphere, the cylinder is in the form of a narrow band
around the equator. The crescent dial is essentially half of an
armillary sphere. It would be possible to make these sundials
adjustable for latitude, so they would keep good solar time at any
location. I can't see any means for making this adjustment in the
pictures in the brochure, but maybe it is hidden somehow for esthetic
reasons. If this adjustment is available, these two sundials are the
only ones I might consider buying, but not at the prices that the
brochure asks!
dow
Richard Russell
> I took a look at the brochure from Haddonstone. I agree that the
> thickness correction must be made in the opposite direction from usual
> because the lower edge of the gnomon is used to tell the time.
> Did you actually buy one of these things?!
Yes, we've had it for fifteen years (they've apparently been
manufacturing the same dial for 35 years) and I've been repeatedly
mystified that it can be set to be quite accurate in the morning, but
not in the afternoon, or vice versa. It's only recently that I
discovered why!
> I'd like to think that the lower edge of the gnomon is a knife-edge.
Certainly not. If it were, the problem that triggered this thread
wouldn't happen (wake up at the back!).
> These inaccuracies immensely outweigh those due to the thickness of
> the gnomon, the diameter of the sun, atmospheric refraction, etc.
No, they don't. None of those other inaccuracies causes a twenty-
minutes (or so) offset between the time indicated by the sundial in
the morning compared with the time indicated in the afternoon, which
is what happens with ours.
Obviously the *absolute* accuracy of a flat sundial, even if designed
specifically for the latitude and longitude at which it is installed,
isn't very good because of the Equation of Time. Being made for the
wrong lat/long exascerbates this, but nevertheless the effect is a
slowly-changing one during the course of the year. This is something
I expect and accept in a sundial. In no way (in my opinion) does that
excuse a design fault which causes a 20-minute discontinuous change to
the reading around noon!
> Basically, these sundials are toys, or maybe decorative sculptures.
Firstly, if they were designed to be only 'decorative' then they
shouldn't have been compensated for the thickness of the gnomon in the
first place. In that case I would have had no cause for complaint,
and in fact the dial would have been considerably more accurate as a
result.
Secondly, if intended to be only 'decorative' Haddonstone shouldn't
state on their site that the dial is "accurately calibrated to five-
minute intervals".
The fact that the dial is (incorrectly) compensated for the thickness
of the gnomon, and is described by the company as "accurately
calibrated", convinces me that I have good reason to be unhappy about
the faulty design.
dow
latitude different from that of your location, there is a simple way
to fix the problem. Suppose you are at 51 N and the dial is designed
for 55 N. (You can check the design latitude by measuring the slope of
the gnomon with a protractor.) If you set the plinth so it slopes 4
degrees to the north instead of being level, the dial will function
correctly. Of course, the thing will look a bit strange, but you may
have to sacrifice appearance for accuracy.
You said that if you set the dial to be accurate in the morning it is
about 20 minutes wrong in the afternoon. If this is due to the
thickness of the gnomon, it would have to be an very thick gnomon,
even with the correction being done the wrong way. I think it's likely
that a large part of this error is due to the dial being designed for
a different latitude, so you could improve the accuracy by sloping the
plinth.
dow
Richard Russell
> You said that if you set the dial to be accurate in the morning it is
> about 20 minutes wrong in the afternoon. If this is due to the
> thickness of the gnomon, it would have to be an very thick gnomon
The gnomon itself is about 3mm thick. The dial is compensated (in the
wrong direction!) for a gnomon 5mm thick. Therefore the overall
effect is like having an 8mm thick gnomon with an 'uncompensated'
dial. I think you'll find that is sufficient to explain at least a 10
minute shift between am and pm.
dow
> On Jun 18, 3:18 pm, dow <williamsdavi...@gmail.com> wrote:
>
> > You said that if you set the dial to be accurate in the morning it is
> > about 20 minutes wrong in the afternoon. If this is due to the
> > thickness of the gnomon, it would have to be an very thick gnomon
>
> The gnomon itself is about 3mm thick. The dial is compensated (in the
> wrong direction!) for a gnomon 5mm thick. Therefore the overall
> effect is like having an 8mm thick gnomon with an 'uncompensated'
> dial. I think you'll find that is sufficient to explain at least a 10
> minute shift between am and pm.
>
> To reply by email change 'news' to my forename.
It depends on the diameter of the dial, of course. If the dial were
immensely large, a few millimetres would make practically no
difference.
Suppose the distance from the tip of the gnomon to the point on the
dial where its shadow lands is 240 mm. So, seen from that point on the
dial, the angular thickness of the gnomon is 8/240 or 1/30 of a
radian. Just about two degrees. At 4 minutes per degree, that comes to
8 minutes. Okay. Ten minutes, near enough.
But you said that the difference between morning and afternoon is
about 20 minutes. Presumably, that's from observation. So there are 10
or 12 minutes that aren't accounted for.
Imagine a sundial that is at the exact latitude that it's designed
for, and the time is 3 pm. Obviously, the shadow of the gnomon will be
on the 3 pm line on the dial. (For simplicity, assume that the
Equation of Time is zero.) But suppose we now somehow rotate the whole
thing about a horizonal east-west axis by 4 degrees toward the south.
The sundial is now orientated as it would be if it were set up
conventionally at a location 4 degrees of latitude south of its
design latitude. Obviously, rotating the gnomon will cause its shadow
to shift, so it will no longer be on the 3 pm line. I haven't done the
exact math, but I believe the shadow will move around the dial by 4
degrees or a bit less. The hour lines on the dial are something like
15 degrees apart, depending on the latitude, so the sundial will, from
this cause, now be something like a quarter of an hour wrong. 10 or 12
minutes sounds plausible...
dow
Richard Russell
> Suppose the distance from the tip of the gnomon to the
> point on the dial where its shadow lands is 240 mm.
The actual distance is 150 mm (measured).
> So, seen from that point on the dial, the angular thickness
> of the gnomon is 8/240 or 1/30 of a radian.
So that's about 8/150 radians in practice.
> At 4 minutes per degree, that comes to
... just over 12 minutes.
> But you said that the difference between morning and afternoon is
> about 20 minutes.
That's not a precise measurement. It could be 15 minutes.
> Presumably, that's from observation. So there are 10
> or 12 minutes that aren't accounted for.
More like 3 minutes.
So the simple 'gnomon thickness' explanation is sufficient to account
for the bulk of the effect.
Richard.
George
This discussion has been long and detailed!
But.
How accurate do you want a sundial to be?
With the passing years, unless there is a (very) complicated adjustment
system,
will there not be an accumulating discrepancy?
george
"Richard Russell" <ne...@rtrussell.co.uk> wrote in message
news:e27974d4-62ca-4384...@o20g2000vbh.googlegroups.com...
Richard Russell
> How accurate do you want a sundial to be?
I'd like it not to suffer a 15-minute discontinuous jump at noon!
After some correspondence, Haddonstone have now removed the
'accurately calibrated' comment from their web site (hardly an ideal
solution, but a success of sorts).
> With the passing years, unless there is a (very) complicated
> adjustment system, will there not be an accumulating discrepancy?
From what cause? Don't leap-years and leap-seconds effectively
compensate for any "accumulating discrepancy" (at least in my
lifetime)?
George
Well, for me 15 mins. would be quite acceptable.
The difficulty in reading and the possibility of no shadow
make for a poor timepiece anyway, and so is an approx. indication
only?
But I was really thinking about the sun not being a good time-keeper, with
variations throughout the year. So why use it to get an accurate time?
Needing some guess work? Or considerable calculations?
Don't misunderstand me.
Searching for perfection is an ideal, but is it achievable in this field?
Never mind. If I could get a consistent accuracy of 1 or 2 mins. then
I would be well pleased!
Cheers!
george
"Richard Russell" <ne...@rtrussell.co.uk> wrote in message
news:4b87f7b4-1d03-42e5...@r2g2000yqm.googlegroups.com...
Richard Russell
> But I was really thinking about the sun not being a good time-
> keeper, with variations throughout the year.
Some sundials (not mine) have the Equation of Time marked on them,
normally in the form of a figure-of-eight curve, so as long as you
know the approximate date you can read off the correction factor.
With that adjustment, I see no reason why you shouldn't achieve an
accuracy of 2-3 minutes.
The dial on mine is engraved with sufficient detail to allow you to
read the time to better than 5 minutes, so (adjusting for the Equation
of Time) the design fault resulting in the 15-minute discontinuous
jump at noon is the dominant contribution to inaccuracy.
George
I got the impression that a much greater accuracy was being sought.
george
"Richard Russell" <ne...@rtrussell.co.uk> wrote in message
news:74d76d30-2344-4915...@d4g2000yqa.googlegroups.com...