None of the clocks I have built have worked first time. All of them need adjustment and debugging before they can be expected to run reliably for proper long periods of time. The latest design (R4) was intended to reduce the number of bugs and improve reliability, but here are some of the problems or types of problem that have had to be ironed out.
- Defective gears. On a couple of occasions faulty running has come down to using gears with damaged teeth. Both times I have been surprised by how small the damage has been and how difficult to find! The lesson is to use new rather than second hand gear wheels or at least to check each over rigorously.
- Axle setting. All of the clock’s gears rely on being set to achieve the lowest possible running friction afforded by Lego. Since this design is compact the gears are quite close to one another and it’s quite possible for them to clash if there’s any axial movement. So, I have a habit of setting each axle’s position very precisely by keeping its gears and bushes very close against the wall without gripping it so that there’s no axial movement but no gripping friction either.
- Gear wheel axial slippage. The highest torque axles, which are near to the drive spool and the rewind differential, can be prone to allowing their gears to ‘walk’ along the axle. This inevitably leads to fouling against other gears or the wall. I believe this is where the torque exceeds the gripping force that the gear exerts on the axle which allows the gear to walk along the axle. This led me to add mini 1/2 bushes into the centre of the lower rewind differential and recommend that the weight is limited to 350g-400g. But it is still necessary to choose gear/axle combinations from the parts supply which are as stiff as possible.
- Pendulum fulcrum rubbing. With R3, the pendulum fulcrum consisted of a pair of 8-tooth gear wheels rocking over gear racks. Depending on variation within the tolerances of the particular cog moulding this interface could lead to some friction in the swing, which is hard to spot but is enough to spoil the performance. To avoid this problem I tended to file down the gear rack very slightly on the o/s slopes of the teeth adjacent to the central groove. This problem has now been resolved with R4 which has a more simple fulcrum.
- Clock tilt (mounting angle). The escapement will only work when the clock is mounted precisely level. Finding this level can involve shifting the clock about its hook on the wall marking the working position in pencil on the wall. Sometimes this has taken a while, listening to the evenness of the tick and observing which side the escapement becomes stuck. The pitch angle needs to be set within a range of just +/- 1 degree and the perfect position isn’t always perfectly horizontal! I would also mention that the front-to-back tilt is also important as the axles and pendulum fulcrum will all float and bear lightly to the front or the back adding a fair bit of friction overall.
- Gear wear. After a few months of running the high-torque parts of the gearing tend to show wear, particularly the bevel gears inside the lower differential and the rewind gearing. Small patches of grey Lego dust appear. It may be necessary to change gears with fresh ones every so often.