Back Up Next

Orrery, finish Saturn gearbox, begin Saturn armature drive wheel set  - August 2018



As explained before, space in the planet gearbox assemblies is very tight. Buchanan writes: Today I ground down ten 5mm diameter jewels to 2mm diameter for the bearings for the cluster gear.  

The jewel is glued onto a 1mm diameter spigot and then spun round in two ball bearings mounted in the T shape aluminium block (first photo). As diameter is very critical, I twisted a long coil spring onto the end of my “grinding arbour”. This gave me a drive with no up or down force. Using this method I could grind the diameter to within a 0.0001 of an inch. I made a few spares as well. Tomorrow I hope to start cutting gears.



The first two wheels are cut. In the next photo the tilt is fixed at the correct 34 to the eccentric orbit bracket and the rough eccentric gearbox allowing for an elliptical orbit is seen. A setting clutch is also needed to be devised that will fit within a 1.5mm space for the moons.




Buchanan writes: First cluster gear mounted. Level one. I needed to shorten the jewels as this gear is long enough to fit two jewels into. I will now start to work my way down the two sets of gears. I think I will be able to thin down the plate thickness as well before we start to fret a design out.

The second photo shows the accuracy achieved by the measurement of the micrometer. This incredible amount of jewelling work is the result of our design decision to not have any conventional pivots; that is no steel pivots in brass pivot holes. Only jewels or ball bearings for the holes.


Even at this size this small wheel is fully spoked out with six arms and weighs in at just 0.5 g (0.176 oz.) 


Buchanan writes: I have the first gear finished and working on the next set that drives it. I worked through the relevant angular velocities of the gear tubes and found some nice numbers.







Difference in

moons in rev./day

Days to make 1 rev. in

relation the next moon




100 years




































Just for comparison, a second hand makes 52,560 revolutions/year, 5,256,000 in 100 years. A minute hand makes 8,760 in a year and 876,000 revolutions in 100 years.

I write: I’m confused by your chart. The difference between moon in revs/day is identical for moon 1 and 2, how is this possible? Also in revs/year both moons 1 and 2 are the same, how can that be if the revs/day are 0.52 and 0.37? Am I missing something here?

Buchanan writes: It does look confusing, that is why I sent it to you. I am looking relative speeds, or the difference in speed, between each tube to see how much each tube rubs, or will wear the next. The first column is how many days it takes for one obit around Saturn. In the second column, I convert it to how much of a revolution per 24 hours. Then I subtract the moon 1 from moon 2, moon 2 from moon 3 and so on, in column 2, to get the difference in portions of a revolution per day between each moon. Then I converted it into round numbers for comparison, So moon 1 takes 1.9 days for an orbit. Moon 2 takes 6.6 days to be lapped or overtaken by  moon 1. Moon 3 takes 6.6 days to be overtaken by moon 2

So for moon tube 1 to revolve inside moon tube 2 for one revolution takes 6.6 days, and, moon tube 3 takes 6.6 days to revolve around moon tube 2 because moon tube 2 is already spinning,


In the first photo we see the trusty jeweler’s saw spoking out a wheel. Next the staking tool is used to secure the pinion to the wheel.


Buchanan writes: Fitting all the wheels on just two arbours is not as easy as I thought. I have had to recut two of 6 wheels so far to make them run smoothly. In photo 306 (last photo) I am testing the next set in the frames to make sure that they run freely. I had to cut the teeth in the smaller of the two wheels another 0.1mm deeper as they were just stiff. Level two.




Buchanan writes: We now have 4 sets of gears out of 9! Level four. Photo 1 is preparing the blank. Photo 2, teeth cut. Photo 3, testing; meshing. We got it right this time. Photo 4, staking the gear onto the pinion. Last two photos, the finished result.

I write in response: I really like the last photo. Again your work surpasses those who would (as did Hahn) skip the spokes. Once the plates are fretted out this dainty subsystem will be a gem.


Buchanan writes: Here is level 5. Just more of the same good things. I am really impatient to see what this will look like when the frames are skeletonised.

It is interesting to see how much photography has improved since you first posted that page on the Dent. It looks even better. (Here Buchanan refers to a Dent exhibition clock I purchased from an auction of the objects in The Time Museum, Rockford, Illinois in 2002. I had updated the poor photography that I had done at the time with better images. There were several ideas I had used from this clock for this project including bronze colored metal for the larger wheels, alternate click action for the click springs, vertical arbor fly fans for the remontoire, and most important the Wagner style gravity-driven rocking frame remontoire.)


Buchanan writes: Level 6 is complete. These photos (first and second) show the temporary bushes I make as soon as I have cur the gears, so as to check the depthing. The bushes in the test are only to replace a collet or a jewel to run it on the arbour to test the depthing.



Here we have the smallest wheel in the orrery so far. Look back on the orrery we used as our design model made by Philip Hahn and see how nearly all of the wheels he made were left as solid disks, only the idler wheels were spoked, and these as a simplified four spoke design.


Notice the jewelling on the ends. Buchanan writes: Level 8 is complete. One more layer to connect to the bevel gears and this set of gears will be complete. It is nearly monotonous, if I wasn’t watching the assembly build up.

In this photo the pair of gear trains are all unravelled from a stacked pair to a lineal view. 


Buchanan writes: Here is the gearing finished. Next is the dial and pillars in the rough and then the moon arms and clutches. I have to fit 5 clutches to fit into 8.75mm of height, from the top of the inner moon arm to the bottom of the outer moon arm.



Buchanan writes: Here is a photo of the start of the moon arms, the blanks, marking out and drilling the angled hole for the moon wire, (first three photos). I have some thinner spring wire on order. What I had was too thick. It may delay me a little as it must be spring wire and also be straight and blue well. I may use a bit of temporary wire to be able to place the Saturn gearbox properly on the arm so that I can proceed with the next stage.

The last photo shows an exploded view of the tiny clutch system for each planet. The disk with the hole in it is the rough blank for the moon mount; the hole is where the wire will be inserted as the stalk for the moon. That disk will later be milled down to a decorative arm. The two small rings are the compression clutch spring and clutch face ring. Next is a threaded flange which screws onto the end of the moon tube.


Here the parts are positioned on the moon mounting disk, the second photo showing a match head for scale.


In the first photo the compression spring clutch is inserted into the ringed recess in the moon mount. One can see the spring rise above the edge near the 6 o'clock area and the clutch ring next to it. Next the clutch ring is place on top of the compression spring. Later a screw on flange piece will compress and hold the spring clutch assembly together. There will be five of these, one for each of Saturn’s moons.


These illustrations are the artwork for the top and bottom plates of the Saturn assembly. Buchanan now turns to the artistry of dressing and decorating of the Saturn gearbox, the fun part of this process.



The plates for Saturn are now skeletonized. The first photo shows the separable ring bearing used to stabilize the five concentric moon tubes.



The Saturn assembly contains nearly half the same number of parts as the pocket watch and illustrates the scale of the planetary gearboxes. The last two photos show the lower and upper Saturn frames.

Here is a short video of a manually guided machining operation  on one side of the Saturn eccentric bracket.  It was quicker to do it this way than to write a CNC program.



These four photos show the eccentric gearbox being refined. Notice the bevel drive wheel to match Saturn's 34 planetary tilt to the ecliptic. Most other makers, including Hahn simply used regular wheels meshing at this angle. Yes that way works, but it is not the best way.



The dial ring for Saturn is fabricated. The first photo shows the advantages of using computer aided design for engraving. The second photo shows the computer aided machining. The engraved blank as originally engraved is shown with the initial black wax to show the detail, third photo. It initially appears a bit crude to our standards. The engraving is created using a fine tapered tool. But Buchanan purposefully engraves the dial deeper than necessary so as to be able to later mill the surface of the dial revealing the finer detail a few micro-millimeters below the original engraved surface from the very tip of the tool to give the fine detail in the last photo.


The frame pillars are now decoratively turned.

The Saturn assembly is complete except for the plain dial ring which will be replaced with the engraved piece shown previously. 

Notice the ferrules at the base of each moon stalk, another incredible attention to detail that Buchanan lends to his creations. The moons as well as Saturn are still just plastic mockups at this point.

The Saturn gearbox rotates Saturn and revolves five moons. This is the first component in the orrery for the astro-skeleton clock. It is based on a design by Philipp Hahn, 1780. Only five moons were known at that time.


Buchanan now turns to the armature and drive train of wheels from the hub of the orrery to the Saturn eccentric gearbox which in turn will drive the Saturn gearbox. 


These photos show the tight clearances facing the outermost planet. The most critical area is on the rear side, first photo, between the Saturn assembly and the rear support bracket for the pendulum degree arc dials (beat plate). Look closely, the outermost pillar supporting the dial ring was removed to make the clearance.


Buchanan writes: (First photo) is the basic design for the Saturn main arm. The reason for this is that I need to know where to plant the gear arbours. There are still add ons for dial mounts and so on, to be added to the design.  

This design works in our favour. In (the second photo), you will see that the next small gear out from the main arbour, is set to one side, pen is pointing to it, to enable its bearing to be in the centre of the curved arm. This makes the next gear somewhat larger in diameter , it will be over 4 1/2 inches in diameter, the plastic ring gives some idea of its size. Now imagine another gear the same size but below the arm and staggered because I put the small gear next to Saturn.



The other(four)photos show a double sleeve gear that transfers motion from Jupiter to Saturn.

I decided that I need to make all the Saturn gears before I move on to Jupiter, as they affect quite a few other dimensions. The next assembly is the 4 gear coaxial shaft assembly that allows the feed to Saturn’s moons to get under the eccentric orbit gears and also the Saturn arm (last photo).



Buchanan writes: Today I started on the lowest set of gears on the whole orrery. It is the fixed gear on the main armature and its drive that keeps Saturn facing the same way all the time, (first photo). Then I made a bearing bracket. This supports the outer lower gear, (next three photos).


This drives through the middle of the assembly (first photo above), the assembly brings the drive to Saturn’s moons through the plate. And we take the orbit drive up through the middle of the assembly (second photo). Tomorrow will be a tidying up day for this section before I move on to the drive out to the end of the arm.



One can see that as the wheels are fitted space is tight. In the first photo the bearing collet flange had to be countersunk to accommodate the adjacent wheel. The second and third photos show the wheel assembly to date. Next 130 screws are made, 100, 1.4mm and 30, 0.09mm, another busy day.

The orrery is beginning to take shape.

Back Up Next