Back Up Next

Skeletonize calendar frames, other misc. items - June 2015

The photos below show the final steps in pinning the various levers and cams to their collets. This is a very tedious process. Every part must be perfectly positioned not only in relation to the rest of the surrounding parts on that same arbor which may contain several collets; each with its own lever, but also in relation to the type and strength of the various tensioning springs used throughout the calendar and how those components interact with their neighboring parts.

With the exception of the bevel wheel drive to the year counter, the entire calendar assembly can be disassembled and reassembled without regard to the positioning of adjacent components. There is no need for the synchronization of wheels, pinions and cams. The fact that Buchanan has taken the time to attach all of the cams and levers to their arbors with taper pins instead of allowing for later individual adjusting means that there is no apriority knowledge necessary of how the computer must operate. This is extremely important since the likelihood of anyone having this knowledge many years from now is slim. It should be noted that to fabricate the calendar in this way is much more difficult than to have had it made with the parts allowed to be adjusted upon their various arbors. Everything had to be perfectly aligned; the entire calculating operations debugged, checked and double checked before being permanently pinned.

I was concerned about the possibility of there being problems with adjustments of this device and Buchanan said that pinning, while very difficult to set up was the most stable way to make a complex set of levers as might otherwise be found in a quarter or minute repeater strike train. Our grande sonnerie, pull-repeat system comes to mind. It would be easier to make all of the levers with small adjustments allowing them to be moved in relation to each other, but as one can imagine, this opens the door to them coming out of adjustment. Unlike even a minute repeater mechanism of which there was many prior examples to refer to, this is a unique device. We do not think a reversible; third-order perpetual calendar has ever been made before so there would be no reference examples. Additionally this is as complex as a repeater. There are about 550 parts that comprise the entire calendar mechanism. Later we found that we also needed lever stops to keep the various levers from flopping out of position if the calendar mechanism was removed from the rest of the machine and turned sideways or upside-down. This is another unlikely scenario but one which we will ‘over-engineer’ for. Given the extreme complexity of this machine we have no other option but to over-engineer for every contingency; not unlike the methodology NASA uses in their space vehicles.


Here Buchanan is using the EDM machine to make the smallest taper holes for the pinning of the levers on their respective arbors. The diameter of the holes needed, (6.5 thousands of an inch), as well as the fact that the arbors have been hardened makes this too difficult to do with a conventional drill. A thin wire is soldered to a thicker holder which is then held in the lathe and spun between two fingers to make it straight and true. That wire is then connected to the copper electrode of the EDM machine, red arrow.


These photos show other various parts that have larger taper holes made in the conventional way. The last photo shows a very narrow five sided watchmaker’s broach used to give the parallel hole drilled by a conventional twist drill a tapered profile. It is spun very slowly in the lathe as it cuts the proper profile. This is what was done on the very best movements made in the past. For those who are reading this and have had experience working on clock movements, how many times have you suspected that the hole that has a taper pin is merely a straight sided, conventionally drilled hole? Did that pin go in the same distance when inserted in either end? Yes? Then the hole was not tapered. A properly tapered hole will allow that pin to go in further in one direction than the other. Furthermore, it simply ‘feels right’ when the pin inserted into a properly mating tapered hole.

Here I show a small adjustment problem where two adjacent levers are rubbing against each other. I point this out since this is exemplary of the countless number of instances that Buchanan has had to deal with these types a ‘bugs’. Notice the beautiful delicacy of the spoke and rim connection of the crenulated wheel in the lower right hand corner.

Next we discussed again how the calendar would be adjusted including the perpetual module, shown above. Basically the perpetual module has two sub-components that sit one on top of the other and which can be independently adjusted. Buchanan suggested installing a knurled knob just behind the readout hand for the date on the perpetual module and this will turn the outer sub-component, left hand component in photo. Turning this by one revolution will advance the 20 year cam. So five turns of the knob will click over the 100 year cam which is located just below the 400 year cam and this will advance that cam one quarter turn. This is done while the readout detent is engaged and holds the inner subcomponent stationary. The month tracking ring resides in the inner sub-component and this can be adjusted by lifting the readout detent and then simply spinning the readout hand. This will rotate the entire module. (See video later in this segment).

Here the small indicator hand for the 20 year cam is fabricated. Next the rim of the perpetual calendar module is drilled to accept that hand.

Here the hand and where it is attached is shown. The third photo shows the knurl knob fabricated to manually adjust this portion of the perpetual calendar module.

The knurl adjustment knob is attached to the calendar module and is shown in place in the third photo.


The video shows the two-step process to set the perpetual calendar date module. The first is where one uses the center knurl knob to spin the twenty year cam one revolution per turn of the knob, so 100 years is five turns, and if the clock were to lay dormant for 400 years then 20 turns would fully rotate the 400 year cam. Next rotating the entire assembly by the indicator hand while the calendar detent is disengaged, as is shown next will set the module to the correct month internally. Twelve turns comprises all of the possible months. These steps can be done in either forward or reverse. Notice how small the perpetual date module is.


The upper and lower bevel drive wheels are now installed and this completes the connection of the calendar remontoire drive to the rest of the celestial train.


The calendar unit appears to have filled in quite nicely!

Below are several videos showing the completed calendar system under test.


In these videos the calendar is stepped forward and then in reverse in real time by manually winding the remontoire that powers this system once daily. The remontoire under normal operating conditions is wound in very small increments and slowly throughout the day. Under the demo it is wound more quickly, within seconds, and thus one sees a bit movement in the system before the works are released and the calculations take place. The speed at which the components move is the actual speed as driven by the remontoire and will take place at midnight daily.


In the first video we see the calendar operating at maximum speed. This would be achieved only while the operator is cranking the celestial demonstration key on the machine at maximum speed. The calendar system is equipped with an 'over speed' safety clutch system to prevent damage to the remontoire and possibly the rest of the calendar mechanism from being overdriven.

In the second video we see how the celestial drive will be cranked by the machine's operator and the results within the calendar assembly. Notice how the hands also move with the demo crank. These are our world time dial hands. This dial shows the time at a few other cities in the world. But it also serves to allow us to bring back our demonstration once completed, back to within about ten minutes of where it began no matter how many years forward or reverse we go. The calendar allows us to see where we need to be to within a day. A ten minute accuracy is more than sufficient at the scales represented by celestial functions. Notice how the rest of the clock appears to alive and breathing during this demonstration.

Buchanan now begins to skeletonize the calendar assembly plates.


The front and rear main frames as well as a smaller sub frame are seen in the first photo along with scanned images. Next he front dial plate is shown mounted to the machine.


We see here the process used to design the decorative features for the plates. The open book has a detailed photo of the back side of an antique pocket watch showing the fancy, skeletonized balance cock. The initial design begins to germinate in the drawing in the lower left. The next drawing shows where all the current pivot and pillar holes are located. The object in the center is an outline of the front pillar of the time train to which the calendar is affixed. Here is where ideas are imagined.


The first drawing is the same as that shown on the table above. The initial design begins to take shape with simple curves that attempt to engage with as many of the hole locations as possible without contorting the curvilinear design. Next those lines have been decorated into the organic ivy and spur design used throughout this project. Here is where imagination meets artistry.


Now the fun begins! Once more Buchanan employs the trusty jeweler’s saw equipped with binoculars to delicately pierce the brass sheet and create his ornate design. The same saw that has cut all of the flat stock, from frames to many hundreds of wheel spokes on this project. One can see the inspiration of a pocket watch balance cock on this piece.

The completed decoration for the rear main frame. All of the ivy appears to be growing just like a real plant from a location at the 10 o’clock position from the center hub. There will be further hand filing and refining of the entire piece. A bit of explanation here is in order regarding the nomenclature used between what is called a plate and a frame. A conventional solid piece of brass used to hold the wheelworks of conventional clock work is called a plate. This type of movement is usually found in a cased clock made of any variety of materials ranging from wood to stone and where that clockwork assembly is never seen since it is hidden with the case. In a skeleton clock, or a clockwork movement that is not cased and is specifically made to be seen as an exposed entity and where that movement’s plates are decorated, usually by the process of piercing, or the cutting out of the plate into a fanciful design, the name is changed from a plate to a frame. This is appropriate since in conventional clockwork what holds the clock’s wheel works are indeed solid plates, but in a skeleton clock the plates become the “frame” which holds the wheel works within, not unlike a beautiful picture frame holds a painting. Here artistry is made real through hand-crafted machining skills.

The front frame also will serve as the mount for the enamel dials and their bezels. Unlike the rear frame, where one could draw the curvilinear lines to catch all of the pivot holes, here we are constrained by the outlines of the dial rings. Consequently, there are a few pivots caught by small ivy ‘growing’ from two of the dials as well as several ‘buds’ around their perimeters; once the dial bezels are in place those buds will nearly disappear.


In the first video the operator cranks the demonstration handle four turns to fully wind the remontoire spring. The mechanism is then tripped to power the calendar functions. In this video none of the internal levers representing the logic components are present. The right hand fly fan would be turning if the demo crank were being turned in the opposite direction. In the second video we see the complexity of the calendar mechanism upon the backdrop of the fretted out back frame.

The calendar now reassembled. Front view. Buchanan's hand gives one a sense of scale. Here hand-crafted machining skills result in a beautiful piece of kinetic art.

Rear view showing the triple frames that encompass the main calendar mechanism in front as well as the remontoire and overrun safety clutch to the lower rear of the assembly. 

Rear, upper three-quarter view.

In this photo the front frame has the cutouts for the dial work. The dial rings appear to be a bit larger wider than one might like, but later it will be seen that when the enamel dials and their bezels are mounted they not appear to be too wide.

Rear three quarter view.

All of the behind-the-dial work has filled out that area nicely.


The closely packed components. Here kinetic art is made manifest in this complex, fantasy machine.

Back Up Next