2025-01-22

Last modified: 2025-01-22 20:30:40

• Clock
• 500 error

Clock

The plan is to replace the 7-tooth lantern pinion with a 10-tooth ordinary pinion, and see if the drive weight can be reduced below 7x M20 nuts.

Well it's working with 7 nuts.

Working with 6 nuts.

Not working with 5 nuts.

I think 7 nuts would have been basically equivalent power usage to before, so it now using 6/7 as much power as it was, so another ~15% improvement.

To add another 10:1 gearing stage would take probably 4.5kg of drive weight, and would run the clock for probably 5 days? Rough estimate, could be wrong.

I think it would work with 5 nuts if the catching-arm shaft were balanced better, just runs out of momentum when the arm has to move "up".

I am going to make new versions of the catching arm block and remontoire middle gear, that take a micro bearing instead of using plain bearings. Oh, I only have one left! I've ordered some more. I don't know where all the others have gone, I thought I should have more than this.

I can nick one out of the pulley on the friction tester for now.

It is working, the remontoire winding is less reliable, and the amplitude is higher. So this could suggest that the extra weight of the bearing means effectively there is more remontoire weight? Or alternatively the pivots for the catching-arm shaft are not aligned properly which makes the winding worse, but the reduced friction makes for more amplitude?

I've removed the pinion that was acting as a remontoire weight, so now the weight is just the gear, bearing, and shaft. It is stable at about 205 degrees amplitude, and only 4x M20 nuts for drive weight.

So at the start of today it had 2.5x as much drive weight and ran for 10/7 as long, which means for a given runtime I think we are only using about 60% as much power now.

If it takes 5.75 seconds to pass one escape wheel tooth, and there are 37 teeth, then it takes 212.75 seconds for an escape wheel revolution. The remontoire is 1:1 so ignore it. The winding drum is 10:1 so 2127.5 seconds for a winding drum revolution, and we have enough string for about 14 revolutions, which is 29785 seconds, or about 8h16m. We probably have enough height to fit twice as many revolutions, so 16h32m, which is about 1/530 of a year. And the drive weight is 4 * 85g = 340g. So to run for a year would be 340 * 530 = 180kg. Still not workable, but getting closer.

I fully wound it and set it going at 20:22.

So, todo is:

• take off the catching-arm shaft, balance it, put it back
• make pallets that let the escape wheel tooth pass closer to the centre of the balance shaft (with new 3mm balance shaft)
• try out a hollow balance shaft running on taut wire

500 error

Why is clockwatcher giving a 500 Internal Server Error when you try to view historical data?

The log has lots of:

Error writing to database: table readings has no column named bmp390_temperature

The CREATE TABLE does add this column, so I guess just deleted readings.db and restart.

And historical data looks to be working again as well now.