I’ve just reinstalled the moon clock, after storing it away for a year during my renovations.   And it wasn’t working well.  Basically, the hands would stutter and ‘fibrillate’, meaning it just wasn’t keeping time.  I concluded that they were sticking, so got some proper watch oil [having been told by a clockmaker that bicycle oil was ruination for a clock, because it turns treacly after some time] and took the quartz movements apart.  They seemed to be quite clean, but there was no oil, perhaps the plastic components have little friction, but having got that far, I thought I would try some oil.  I made an oiler out of a piece of fairly thin garden wire, flattened at the end, and used it to carefully apply tiny drops of oil on the pivots and gears.

Testing the motors on the bench, they worked like a charm.  But when I assembled the whole thing, they would rotate beautifully but then start fibrillating and even running backwards!

I slowed the pulsing rate down but this didn’t help.  Then I realised that in a certain place, the long seconds hand touched the glass!   This caused it to spring backwards and then forwards repeatedly.  The answer was straightforward – to bend the seconds hand very slightly away from the glass, but not so far that it tangled with the other hands.  The moon phase disc also seemed to touch the edges of its housing at one point due to being very slightly eccentric.  I was able to pare away a very fine sliver from the edge.

Now, touch wood, it is keeping time beautifully. It was satisfying to watch it at moonrise this morning – the time hand clicked past the rise-time hand, the moonlight came on and the azimuth hand spun to the correct position in a fabulous little routine.  But it needs a bit more ‘soak testing’ before I’m sure.

When I have the energy, I will write this up into a complete project.  In the meantime, I have a wind-up clock to get working again.