To test this feature, visit your live site.

Modification to moon phase clock pallet

I have made five of these clocks and generally they work very well. Beautiful design. I did have one clock that ran well for a couple months then began stopping after a few hours. Cleaned the bearings again and ran it free without the escapement for three days for "break in" but little improvement. Also tried replacing the pendulum bearings with more expensive parts that did run more free and made sure all the gears were turning easily and had some end play. But it would still stop. After studying the way the escapement gear interfaced with the tips of the pallet I tried rounding the corner of the pallet tips to transfer energy to the pallet more easily (see picture) and it has been running continuously for a period of weeks with no other changes. The pendulum now settles in with slightly more amplitude than before. I am now modifying the pallets on all of my clocks the same way. I am considering doing something similar to the escapement gear though it would be easier to simply modify the STL form rather than work on each gear tooth.

3 comments

3 Comments

Steve

Nov 05

Thanks for posting the fix that is working for your clock. All of the desired effect can likely be accomplished by your pallet modification without having to touch the escapement.

I have considered adding the same gentle start on the pallet or escapement of a few of my designs. The reason I resist the change is that the clock would become a recoil escapement instead of a deadbeat.

A recoil escapement is supposedly easier to get working by simply adding additional drive weight. However, it is also sensitive to rate changes when the drive weight changes or when friction changes causing the escapement to have less energy. A deadbeat escapement is significantly more stable and reliable after getting it to run.

I am curious about the escapement behavior on the clock that got the new pallet. When I have a clock with a low energy pendulum, it is usually one of two things. Sometimes, the bearings need to be cleaned again. Other times, the cause is a low escapement energy due to friction in the gear train. The pallet arms move past the release point but the escapement lags and just barely touches the pallet before the pallet moves out of the way. The fix in this case is to eliminate the gear train friction from many possible sources. It could be arbor friction, sticky gears, elephant foot, lack of end shake, frame sag, not enough drive weight, or something else. A thin layer of grease on the pinion teeth can sometimes help.

Steve

Dirk Nansen

Nov 08

Replying to

I had wondered why the escapement and pallet were designed that way. Always good to learn more on the details of clock design. It is curious why this one clock was being difficult. The other four of the same type I have finished work fine and this one ran fine for several weeks. The pendulum bearings were good enough for a 20 minute free test and the clock gears were cycled many times without the pallet so they would break in. The cycle times of fully wound to weight on the floor stabilized (about 2.5 minutes as I recall) and the escapement seemed to be moving strongly against the pallet when it was back in place. So the gear train seemed to be quite free and the pendulum also working well but it would just go a few hours and stop. And I could watch the pendulum gradually loose amplitude of swing until it stopped even though the escapement was pressing it. I had verified that all the gears had endplay and felt free to turn. It is using a full standard length, 3.2 inch weight shell. It is possible that this clock just suffers from relatively low quality of printing on some of the gears though they don't look bad. This was done on a much modified Ender 3 (dual Z drive, Microswiss direct drive extruder, SKR 2 MB and new belts and wheels) with the prints sliced on Orca. But some days the print quality even with a filament dryer is not as good as others especially with the silk copper filament I was using. Elephant foot is definitely an issue and I spent a lot of time trying to eliminate that with a lot of hand finishing (and changes in the slicer). Anyway, the escapement now transfers enough energy to keep it running smoothly and it seems to be keeping time well. There really isn't any reason to also modify the escapement gear since this slight change on the pallet seems to have fixed this clock, whatever its problem was. And certainly the others I have made run well with the original parts. Although it is in my nature to tinker with such things when I can.

Steve

Nov 08

Replying to

It's good to hear you got it working. PrusaSlicer has slowly been increasing the default elephant foot compensation from 0.12mm or 0.15mm up to 0.2mm. I sometimes bump it to 0.22 to reduce retractions on the first layer.

Some filaments do seem a bit stickier than others. There is always some sliding motion when the gears mesh. Sticky filament makes it worse. One thing that can help is reducing the interlocking of the layer line ridges. Start with small layer heights for the gears. 0.15mm is usually good. Even better are different layer heights for alternating gears. For example, 0.15mm for the even number gears and 0.17mm for the odd number gears. Then if the ridges line up at one location, they will not be able to line up at the next location. A thin coat of lithium grease on the pinions seems to help also.