Back Up Next

Finish tellurion, continue dial bezel work  - September  2019

This month Buchanan continues to wrap up the project. The Sun & Moon rise and set module is disassembled for finishing later. The tellurion is disassembled and then finished. The multiple dial bezels that serve as a decorative surround for the enamel dial work continues. Dial bezels may seem at first glance to be a relatively simple job, but it is difficult and exacting, especially as these must be made to look perfect surrounding hand-made dials that were designed and completed a decade earlier.


In this video Buchanan shows the disassembly of the tellurion. Note the sage advice, "As in most cases if you find it's impossible to proceed further, just keep removing whatever you can and it will generally become apparent what is holding up the whole procedure."

The tellurion is housed in a triple frame construction; these are the three curvilinear and jeweled parts on the right. The sickle-shaped part to the left is attached to the similarly shaped counterweight on the rotating tellurion module allowing for it to be poised resulting in smooth rotation and low power consumption.


These photos show a few of the parts within the module.


The first photo shows the large counter-poising weight along with other miscellaneous parts, next the numerous tiny screws in the Petri dish look like tiny bugs. These are being polished and readied for bluing.


The dial work and pointers are disassembled and made ready for final re-silvering and bluing.


The dials are first prepared with black dial wax, first photo. Next photo a finished dial complete with its knurled bezel, in this case the sidereal lunar month dial; the synodic lunar dial fits inside and is in the background.


The compliment of twelve tellurian dial work pieces after black wax filling of the engraving work and final French silvering. The main enamel dial and bezel were finished in June.


Shown above is the undercarriage of the Earth assembly. Here one finds the sidereal and synodic lunar dials. This strongly parallels the relationship between sidereal and mean solar time. In both instances the sidereal period is related the movement of a body, in this case the orbit of the Moon as it relates to the distant fixed stars. An explanation of these two functions was given in the November 2015 installment. The outer chapter ring is just 1.25" or 3.1 cm in diameter; on the scale of a wristwatch dial.

The ring below (the node ring) represents the slightly inclined orbit of the Moon around the Earth and is at a 5.14 to the ecliptic and rotates according to the Metonic cycle, a period of 19 years (235 lunar months). On just about every other tellurion if the maker even bothers to show the Moon moving on a slightly inclined ring, that ring is stationary, merely alluding to the moon's inclined orbit. In this example the fact that the ring rotates to the Metonic cycle allows for the prediction of solar and lunar eclipses.

The earth globe is surrounded by twice the compliment of longitudinal and latitudinal rings. One pair is fixed in relation to the Earth  and is related to the ecliptic, the second is related to the Earth as in a conventional tellurion but is movable from the knurl knob at the top and is used along with the Metonic node cycle ring in the solar and lunar eclipse demonstration. There is also a pair of brass and silver arcs indicating the solar and lunar horizons of sun and moon rise and set, brass for the solar and silver for the lunar horizons. This color scheme is consistent within the Sun & Moon rise and setting module dials located above the tellurion. Several videos are available showing the demonstration of a solar eclipse as well as other functions of the tellurion in the January 2016 installment. The Moon is a sphere of what else? Moonstone.

If one looks carefully, there is a tiny solid gold dot positioned on the globe where the clock will eventually reside.


This pair of photos show the main concentric tube array under the Sun. While we have often respected the methods and style of the past masters of horology, we do employ modern methods when useful. Here ball bearings are within the concentric tube set. This was a further imperative because unlike most tellurians, this one is mounted with the dial vertically oriented, which would cause a conventional concentric cannon tube set to be deformed under the force of gravity. This is further aggravated by the weight of the large rutilated citrine quartz Sun.

The completed tellurion assembly.


Predicting an eclipse is as follows: The node ring below and surrounding the Earth has a sector dial at the 5 o'clock position, 'Ascending Node' with a pointer, (the Descending Node pointer is opposite at the 11 o'clock position behind the earth globe; visible in the photo below); at the 3 and 9 o'clock positions are the 'Eclipse Season Window' dials for the solar and lunar eclipses, respectively. The alignment of the pointer on the node ring within the Eclipse window, and the Moon being directly between the Sun and Earth with its pointer at the beginning of the 'E' marking tag at the center of the Eclipse Window will indicate a solar or lunar eclipse event. For a solar eclipse the degree number that the node pointer registers within the Eclipse Season Window at the 9 o'clock position is then matched to the same number on the ecliptic latitude ring, then moving the Earth's latitude ring adjacent to that number, it is interpolated to the number on the Earth's latitude ring; once determined, the ring is then repositioned to align exactly behind the Moon. The position of the interpolated number is where on the Earth the eclipse will begin. Once the Moon's pointer transverses the the 'E' tag the eclipse event is over. This takes about one third the rotation of the Earth, eight hours) and careful observation along with the use of the twenty-four hour demonstration dial, or for more accuracy the demonstration winding square to the right which is one revolution every two hours, will allow the user see where and when the solar eclipse begins, the path it transverses over the earth's surface and where and when it ends. Substituting the Earth for the Moon's role next to the Sun  and using the Eclipse Season Window at the 3 o'clock position allows for the prediction of a lunar eclipse.

Notice the sickle poising weight needed to balance all of the machinery opposite. That weight is not simply a solid piece of brass but had to be hollowed out and filled with lead to achieve the needed mass to properly poise the armature. The use of high grade ceramic hybrid bearings is an imperative for the pivot point upon which the entire mechanism's mass rotates and can be used without the need for oil. The planets of Mercury and Venus are represented by Jasper and Serpentine.


In this video the finished tellurion is placed on a turntable. Everything but the enamel dial and bezel is present. The later is supported by the surrounding structure on the clock frame, not by the tellurion module itself. Notice the scintillations in the quartz Sun sphere. These are produced by the mineral, Rutile, embedded within when the quartz was formed over 100 million years ago and was chosen as an allegory for the Sun's rays.


The tellurian is installed into the rest of the machine, but without the dial work. It almost seems to disappear within the surrounding complexity. The rutilated citrine quartz Sun really shows off its color here along with the Moonstone, Jasper and Serpentine stones.

This video shows the disassembly of one of the more complex modules in the machine, perhaps only surpassed by the third-order perpetual calendar. I hope to never need these, but the set of disassembly videos that have been created are essential for personnel that may need to service these components in the next 50 or 100 years, depending upon how much usage the machine serves.

For me, this module contains one of the more interesting designs ever used in horology, the Antide Janvier variable, slant-wheel differential. I only know of three other clocks that have used this design in the past to correct for the equation of time or the various anomalies of the Moon's orbit. The reason is probably because the same functions can be done more easily with cams. For a further discussion of this component see the installment for November 2016.

Continue dial bezels

Janvier variable differential setting dials and name tags


The first photo shows a silvered dial with knurl bezel for the Great Anomaly setting dial and the initial fitting for the two enamel dials for both the Great Anomaly and Projection corrections for the Moon’s orbit. The second photo shows both variable differentials for the Moon’s orbital corrections, the left has the knurled bezel contrasting with the right yet to have its knurled bezel fitted.


The dial work continues; the bezels now properly finished and dress up the enamel dials nicely. Note the thin, curved bezel perimeter compared with the initial fitting in the first photo above.

Sun & Moon rise module and set twenty-four hour dial


The center 24 hour dial was made as a complete ring. At the time it was made ten years ago, we did not know exactly how it would be installed so we did not have it made as a semi-circular ring. In any event it is unlikely that it could have survived the kiln firing without warping if it were already cut. Now Buchanan has to cut the fragile, brittle dial, a nerve-wracking process. Fortunately we had two dials if a mistake was made.


The first photo shows the two cut-down dials. Next the good dial is shown from the rear as a perfect insert-fit into the lower dial seat. 


First the same dial is shown fitted into the lower dial seat from the front. Next the upper sector dial section is put into the machine and the lower dial bezel is checked for fit within the context of the surrounding material. 


The upper sector bezel is now fitted adjacent to the lower seat. 


This photo shows the lower dial bezel fitted over the lower dial seat. This gives a look like the lower bezel has melded to the parts of the sector bezel above it as it spans from the left before encountering the upper sector dial, then past the perimeter of the upper sector dial bezel and then as it covers the surface of the enamel twenty-four hour dial, then spanning across the lower section of the sector dial, and then back across the same set of sections until it exits the other side.


The first photo shows the inner bezel cutaways to accommodate the spokes of the equation calendar setting dial. Next the reverse side of one of the Janvier variable differential signs. The rim of the bezel is cutout around the fork that holds the dial back plate.

Demonstration, world time and thermometer dial


Left, the twenty-four hour world time and demo dial, right the thermometer dial. 

Equation of time calendar setting dial and strike select control dial


Left the equation of time setting dial, right the strike select control dial.

Pendulum beat plate dials


Above is a video showing the creation of the beat plate bezels. Here we see them cut from the brass blank on the mill. Next the bezels separated from the blank.


The enamel dials for these are handmade and are not an exact fit. Making a complex shape is far harder than a constant diameter dial ring. Buchanan could have made the job far easier by fitting the bezels to dials, but then they would not be a perfect match. So ensues more nerve-wracking grinding of the fragile and brittle dial material to achieve a perfect fit.


The first photo shows the ground dial to fit the bezel, next the view from the front after the bezel is finished and the rear dial plate fitted over the recessed enamel dial work with the brass back mounting plate.  


This photo shows the two enamel dials superimposed on each other and one can see how they are not an exact match. Again, the easy way out would have been to make the thin bezels to fit, but they would then not be perfect pair. The perimeters are ground until they are a good fit to the matched bezels.


The beat plates are complete. Look at the interesting distortion of the machine as reflected upon the pendulum sphere.


See how beautifully the end of the beat plate bezel tapers to a sharp point and perfectly contains the enamel dial within. Next the pair are installed. Looking carefully in the background one can see some of the machine’s main components waiting for their turn. On the upper shelf the planisphere, left, orrery right; lower shelf, tellurion.

Grande sonnerie, petite sonnerie pull/repeat button


The Grande and petite sonnerie pull/repeat button.


The interesting thing to see is the fact that Buchanan made this button an openable screw-down box. Clearly not the easiest way to have done this, the threading of the two halves could have been totally avoided with a press-fit design. But it is still very cool. Perhaps there is a small thing I could hide in there?

Orrery dial


In these two photos Buchanan first makes a metal disk to act as a stabilizing plug to secure the dial as well as the blank for making the inner bezel. The blank outer bezel has also been secured. The dial is not the actual enamel dial, but a plastic mockup. The enamel dial is far too fragile to use in this operation.  


The outer bezel undergoes decorative turning. Look at the prodigious swarf in the second photo.


The outer bezel is finished. Note in these photos the difference in the detailing of the side of this bezel as opposed to the other dial bezel work. Here there is a raised and wider outer edge past the rear of the bezel and some turning below that surface. Other dials have a flat outer surface with the decorative turning inside the outer perimeter. The reasoning for this is the fact that this dial is horizontally mounted and the bezel will be seen edge-on and so the extra detailing is needed. A flat perimeter would have looked unfinished and bulky from that viewing angle.


In both photos the outer bezel is done. Compare the inner bezel in the first and second photos, the former before turning and the latter afterward. Keep in mind that the outer bezel is over 14 inches and the inner 12 inches (35 cm and 30 cm), and this contributes to the greater difficulty in the making of this bezel to be be perfectly flat.


The orrery bezel is finished. 

One can see the turned edge of the orrery bezel, imagine if the edge was a simple flat surface.

Parts count to date: 5983

Back Up Next