I won’t pretend to any kind of subtlety or elegance in how I went about designing the gear train for the clock; it was sledgehammers all the way, baby, (in Excel, no less …) and definitely no KISS-ing!
As a first go, I simply made a set of worksheets which calculated all of the possible ratios from a single pair of cogs, with between 11 and 47 teeth, and then used some repetitions of the ratios until I got close to the one I was after.
Later, when it became clear that the series I needed wouldn’t fit easily with the way I needed them to work (eg: the tide indicator), I allowed myself to expand to using the chain as if it were a cog too. This meant I had more flexibility over the ratios (as I could choose … kind of …) the number of chain links to use, but also meant that the calculations and construction got a bit more challenging too. I’d given myself a 600mm x 600mm space to work in, with 200mm depth, so that constrained the size of some of the larger rings a little.
The time is shown by a minute-hand attached to the main driving shaft, which revolves (a little predictably?) at 1 rev/hr. This drives an hour hand via a 12:24, 14:28, 14:42 series to give the required 1:12 reduction of hour-cycles of the minute hand, to 12-hour cycles of the hour hand. The hour hand is attached to the 42 cog wheel on a free bearing surrounding the main central drive shaft. The minute hand is external to the casing, attached to the drive shaft in front of the perspex. At present there are no numbers as I’m still pondering how best to do them (if at all).
The moon position needed a ratio of 24.8412 hours per revolution: instead of the 24 hours that the sun takes to cycle relative to a fixed point on earth, the moon takes a little longer. The closest I could get to this with the integer ratios available with the limited bike cogs was 1:24.8447, from a series of minute hand drives the 12:24, 12:24, 14:25, 23:80. The 80 cog is created by glueing the bike chain around a big perspex disc (the moon-position ring), which contains a cutout to represent the position of the moon in the sky. Nothing too fancy here – above the axis is above the horizon (ish). Behind this cutout the moon phase disc is shown, as described in the next section.
This is the most complicated part of the clock (and just between us, I’m not convinced I’ve got it right, shhh!). Because the moon position is continuously changing, all of the calculations of phase have to be relative to that motion. I used a coaxial shaft: the minute-drive shaft runs through the centre of another drive shaft, onto which the 18 and 15 cogs are mounted. This means that they can rotate about the same axis as the main drive shaft, but they’re unaffected by that shaft’s motion. The moon phase is driven by a sequence from the 42 cog hour cog, as 42:42 (but reversed direction), 16:39, 15:18 to the coaxial drive shaft. The other end of the coaxial drive is connected directly to the 15 cog, which connects finally to the 11 cog mounted on the orange moon position wheel. This drives a yon-yang type design that sits behind the moon cutout to show the phase. The phase ring is turned predominantly by the moon position ring, with this combination providing the correction it needs to rotate correctly on top of that motion.
Auckland has two harbours, and the clock can predict the tides at both of them. Because both are driven by the same mechanism, it’s a simple matter of having a tide level indication, but building in an offset to account for the approximately 3-hour lag between the harbours. On the outside of the casing is a large ring of 104 chain spaces, with indications for high, low, ebb and flow tides on it. This hangs under its own weight from a small 11 cog at the top of the clock. The offset between the harbours will be achieved by positioning pointers on the top corners; Piha on the top left is approximately three hours earlier than Westhaven in the top right. Tide indications at other locations around the country can be added too. The 16 cog is on the main drive shaft (the minute hand), then 16:32, 11:17, 11:104 cogs on the large external ring.
So the works are built, the casing is made, but I’m still waiting for inspiration to strike about aesthetics and finishing. What kind of numbers? Are any needed? Should the back be transparent or mirrored? Should the sides be enclosed (to stop the dust getting in) or exposed (because that looks funkier and you can see the workings better)?
Last but not least I need to hook it up to a driver. The main central shaft will need to be driven at the speed of a minute-hand, and there’s plenty of space behind the clock to hook that up. The weight of the larger rotating parts has been balanced as best I can by adding large nuts at strategic places so that no part of the system is under too much tension at any time. This also means that there’s a fair amount of inertia to overcome, so I’ll probably add some weights to the driveshaft nudge it along.
Still a work in slow, steady, but (fingers crossed) well-timed progress. Watch this space!