2025-01-03

Last modified: 2025-01-04 09:22:34

Clock

I have moved the escape wheel slightly closer to the balance shaft, and adjusted the balance spring to take more weight off the pivots, and it is working much better now.

I have moved the remontoire weight inwards and the amplitude is about 245 degrees now, and the remontoire winds very reliably, with just 10x M20 nuts and no allen key. I'll let it run down and see the plot, and then try removing drive weight.

Weaknesses are:

• remontoire fly is installed backwards
• remontoire fly is not balanced
• remontoire weight does not have a micro bearing
• "temporary" winding drum bearing holder is too easy to move by accident
• needs a winding key, and something to prevent reverse-driving
• spiral on winding drum goes the other way to the direction that the string exits
• weight needs to fall inside the frame (does it?)
• strings from pulleys don't go in very straight lines
• string is not long enough
• too much friction in bearings
• balance pivot adjustment is too wobbly

And then obviously I want to try experimenting with a hollow balance shaft with a tight piece of wire up the middle, and try using a smaller diameter balance shaft so that the escape wheel teeth can pass closer to the centre, and put more teeth on the escape wheel.

Damn, the serial port got disconnected! Don't know why.

Seems broadly the same as before.

I've reset the serial port and am trying again. I expect what happened is I bumped the USB cable while mucking about with the clock.

The string slipped off the upper pulley once. I'm not sure whether this was caused by unwinding too fast because the escape wheel skipped teeth because the amplitude got too high.

I've moved the upper pulley to the side, opposide the winding drum, so that the string is closer to coming off straight instead of at an angle. Also slid the remontoire weight all the way in, to reduce the max. amplitude.

I found it only takes 8 nuts to keep the remontoire wound, so that is an improvement over the old frame. Probably because I'm using more proper bearings now instead of just running the shafts on bare plastic.

OK, let's say it's starting from 21:00, and I can use the "historical" view to get the plot tomorrow if it's still running when I go to bed.

There is something wrong with the clockwatcher. The serial port keeps giving a buffer overflow and I think the Pi side is getting out of sync. Not sure what is wrong.