Gina

Been looking again at the gear train from minutes wheel to escape wheel. My first attempt was dividing the required ratio of 60:1 for minutes to seconds into the obvious 10:1 and 6:1 but the ratios can be 8:1 and 7.5:1 with gear ratios of 80:10 and 75:10 with an 8:1 ratio being easier to produce than 10:1. The intermediate gear can have 10t and 75t and still be mod 2.25 to match the other gears so I would only need to redesign the one gear - result!! I think this looks noticeably better.

Gone for halfway between - 6mm steps.

Cam was too big so reduced the step size from 8mm to 4mm. Might change it again as 4mm steps seem a bit small.

Thanks Dave. I'd forgotten about the E1 driver so yes the Duex2 will do. The frame does flex which is why 4 mount points are required.

Miniature bearings.

Very interesting, Dave ? I'm using 4 threaded rods and just one motor on my Giant printer and it does struggle a bit. I need supports on all four corners because I'm moving a frame up and down with the XY axes on it rather than the bed. This frame is not rigid enough to work with 3 rods. It would be interesting to try 4 independent motors but I haven't investigated if the Duet could handle this. I would need 3 extra motors which would mean the Duex5.

Refined the cam deign.

This is a screenshot of the cam model that controls the striking.

The yellow minutes wheel (often called the "great wheel" in clock parlance) will need connecting to the minutes arbour/shaft with a slipping clutch so that the time can be set by turning the minute hand. Another gear attached to the minutes shaft will drive the intermediate wheel which in turn will drive the hour wheel. The hour shaft will be a sleeve on the minutes shaft with the hour hand attached. Also on the hour shaft (or maybe part of the hour wheel) will be a cam which will determine how many strikes to perform. Another cam on the minutes shaft, gives the striking mechanism "warning" and then actual time the strikes begin.

Another photo of the gears roughly laid out on the acrylic sheet. The ratchet wheel actually has 32 teeth not 45 - that was what was there before. The auto-winding stepper motor will fit in the lower left-hand corner and have a ratchet pawl attached. Another pawl will go just above the drive pawl to stop the ratchet wheel going backwards as the drive pawl returns. I haven't shown the anchor escapement in the top in this view. The escape wheel shaft will have the seconds hand on the end and a seconds dial will be attached to the 12 o'clock mark of the main dial. To reduce cutter, this clock will not have numbers beside the dial. The space top left will have the auto-regulator drive and top right the moon dial drive gears. Down the LHS will be the striking mechanism with the chain drive for it at the bottom. I don't think there will be much spare space!

We can calculate the time the clock would run for in the event of a power cut stopping auto-winding :- Circumference of main drive sprocket is about 350mm. Allowable drop of the weight is about 1.5m and with pulley corresponds to 3m of chain over the main drive sprocket. Number of turns of drive sprocket = 1500/350 = 4.3 approx. equating to 4x4.3 = about 17 hours. So a power cut lasting up to several hours would not be a problem. The auto-regulation system should have set the timing pretty accurately so the clock should keep time for the duration.

Calculations for the auto-winding system :- As shown the chain sprocket/gear to minutes wheel is 4:1 so the drive sprocket (pink & black) will rotate once every 4 hours. Drive sprocket has 18 teeth and auto-winding sprocket 8 teeth giving a ratio of 9:4 so auto-wind sprocket will turn once in 4x4/9 = 16/9 hours. Number of seconds for one revolution of auto-wind sprocket = 3600x16/9 = 400x16 = 6400. If this has a shaft with a ratchet with 32 teeth, each tooth would need a push every 200 seconds. A stepper motor would need 200 steps per revolution so that makes 200 steps in 200s or 1 step every second. A good result. Instead of a 45t gear on the auto-wind sprocket there will be a 32t ratchet wheel, pushed forward one tooth with each revolution of the stepper motor.

Proper drive chain gear added (red). (The curvature is not the clock but due to wide angle lens.)

This photo shows the probable gear arrangement. The upper white, green and yellow gears are the new ones, printed in PLA. The others are the original gears and sprockets (the latter in pink and blue).

Glowing Moon

Printed the globe from STL files on Thingiverse. Took a long time - lots of fine detail.

Too early for postman here - usually around 11am or so...

Photo of latest X carriage.

A standard longcase clock is regulated by turning a threaded thumbwheel on the end of the pendulum rod to raise or lower the bob. The top is suspended by a strip of spring steel held in a slot and another way of regulating the clock is by lifting the spring in the slot and changing the pendulum length that way. With this method the slot is fixed but the suspension point is moved up or down. The suspension point can be on the end of a lever which is controlled by a stepper motor and threaded rod. This means the clock can be regulated from the Arduino.

Printing a test block in PLA - 100mm x 100mm x 10mm. This is first layer and half of second printing pretty much perfectly.

Having given this more thought I am satisfied that reducing the gears to 90% of original size is well worth a try. I can do this in Slic3r without having to redesign the gears except for the escape wheel which will need redesigning because I want to retain the 100mm diameter to match up with the already made anchor escapement. I have the two acrylic sheets for the new design so can shortly drill holes and add bearings ready for testing once the new gears are printed. I can test the running without the auto-rewind mechanism with just chain and weight. The problem before was that the escape wheel was not concentric with its axle which was the fault of an earlier 3D printer - now replaced. Following jobs are to design the auto-winding gear train and drive to the moon globe dial plus a striking mechanism. Also on my mind is some way of providing auto-regulation of the clock timing. Having had experience of control systems I think I can probably devise a control system to accomplish this, taking the timing from a Real Time Clock chip (already controlling a stepper motor via Arduino for the auto-winding). I'm thinking in terms of a two stage arrangement with first stage getting the timing almost right and then the second employing a Phase Lock Loop to synchronise the pendulum with the RTC.

That looks like some wonderful device! ??

Dreadful weather here

Crumbs!! The parts cost more than that - even the cheapest you can get!! Looks far to good to be true! My mantra is "If it looks too good to be true it probably is!"