Groups Feed

This diagram shows the gear topology in the clock.

There is a long unbroken chain of gears between the winding drum and the escapement. If the pendulum is swinging properly then gear 4 must be rotating at close to once per hour. The rate is determined by the pendulum length and it cannot run at 2/3 speed. Therefore, the issue must be between the friction clutch and the hour hand.

The friction clutch adds a loose connection between gear 4 and gear 4b to allow changing the time without affecting the pendulum. It does not need to be very strong. It only needs to be strong enough to overcome any downstream friction. Music wire (also known as spring steel or piano wire) creates enough pressure against the M3 arbor. Soft steel or brass wire might not apply enough pressure.

Gear 6 is the primary gear to check for friction. The printed gear 6 arbor must pass through the dial without binding. Also gear 6 must not be pressed against the back of the dial. Push and pull on the hour hand to make sure gear 6 has some end shake. Look for end shake in gear 5 as well.

One more thing to check is the gear 4b set screw. The friction clutch is circled below in green in the exploded diagram below. It loosely presses against the M3 arbor. Gear 4b tightly connects to the M3 arbor using a set screw, also circled below in green. Move the minute hand and make sure gear 4b moves at the same rate. There should be a slight resistance felt by the friction clutch.

Hopefully, this will give you some ideas to check. Please report back if you find the cause.

That is probably the correct material, but really expensive. The image showing the inside diameter of 1/2" is misleading. 1/2" EMT conduit has a cross sectional area of 1/2" with an ID of 0.622" and an OD of 0.706".

Any hardware store in the US will have it for a lower price. Home Depot has 5' lengths for $4.96 and 10' for $6.58.

I have noticed it as well. The effect is exaggerated by the ball limiter that allows the balls to creep forward instead of jumping by a full ball width each time a ball drops. This creeping motion gives the balls more chances to find a bump or a flat spot where they can rest.

The upper ball ramp has a 3 degree slope with the balls rolling along the 45 degree surfaces. The lower curves have a larger radius than the upper curves so the slope might be just under 3 degrees.

Possible fixes may include:

1) Sanding the 45 degree slopes anywhere a ball stops.

2) Removing all oil from the balls and any that may now be on the tracks.

3) Adding oil to the balls, although I suspect this might make it worse.

4) Redesigning the upper ball ramp with steeper slopes.

The first two options might solve the problem.

The last option might be the best long term alternative. The clock would either have to be made taller or one spiral loop would need to be removed. Removing one spiral is the least intrusive since only the upper ball ramp would need to change. It might limit the clock to a 30 hour runtime using 100 balls.

I must have hit the "publish" button while MMF was still processing the zip file. I uploaded it again and it looks to be present this time.

I really like the metallic colors.

I sometimes get one or two individual balls left behind. They all start moving if the last one moves. Two possible causes for a chain of non-moving balls are surface imperfections or sticky residue. Some of the balls I received had an oily coating that could make the balls stick together. Try wiping them in a towel to see if it improves.

Designing the ball ramps was a tradeoff between slope and overall clock size. The upper ball ramp tracks have around a 3 degree slope. Making them steeper would have made the clock even taller than it already is.

There are four stepper motor driven clocks to select from. All use the same algorithm with jumpers to select motor speeds since each clock has slightly different gear ratios between the motor and the second hand. They will all be accurate to about a minute per year. If you build your own controller, the assembly guide shows how to set up the Arduino IDE. There are also plenty of YouTube tutorials to help with the environment.

I printed it twice, both as the original SP4, as well as the new 4b, but both only ran for a month or so before they would only run for about 24-48 hours at a time before running out of steam. I suppose it is possible the wall mounting screws were not quite tight enough. The top one certainly was NOT tight enough to mount the 16 I just printed. I had to screw it in several turns more to get a good seal on the 16 against the wall. I guess I could give the sp4b another go with tighter screws.

The above photo shows you what I was referring to as what I wished the 16 frame supported. I had to use double-sided Velcro, but how much nicer would it be if the top had a flat portion that fit the handle, with raised edges for solid capture? Could even modify the face to provide a little shelf at the bottom and a slight bend to the face to support the winding portion.

The image in the assembly guide is the wireframe view from the front of the clock. I will turn it around in the next assembly guide release. It only goes together one way.