A little more design work over the weekend. The biggest challenge at this point is the motive force. The very best spring or weight-driven movements are accurate to several seconds to over a minute per month - and that's brass gears and jewel bearings, not wooden geared. Not good enough. Originally, I thought I'd do a synchronous motor, driven by the 110VAC line. While the frequency varies a little in the short term, over the long haul it is remarkably stable, providing a near-perfect time base. However, given all the complications, setting the clock in the event of a power failure would be a real task. So I looked into a low voltage synchronous scheme with a high accuracy battery backup. This is way too complicated, so I decided on a stepper driven design with battery backup. No convenient way to grab the AC waveform, so it'll be a temperature compensated crystal, with an accuracy of a few seconds per year. Good enough. The gearing has been defined as per the spreadsheet. I'm not completely sold on the star chart with an accuracy of one hour in 35 years. I can get to 1 minute in 167 years by adding another gearset, but don't know if that'll fit. The moon phase is good at 1 day in 122 years. The initial layout is attached. Shown are the pitch circles of the gears, each color coded to the spreadsheet. It does not include the perpetual calendar yet which I am reserving for the right side. So far 12 pinons and 12 gears.
At this point I am completely blown away by the watch makers of the 1700-1900s who did all of this and more without spreadsheets and CAD!
