Wow, you sure are having some bad luck with the weight shell. I consider this to be the least "magical" portion of the entire clock. All it does is hang there. The pulley does provide some separation to help guide the line as it is being wound. One thing to watch while winding is to make sure the line does not creep up the 45-degree slopes on the barrel. If it does creep up, it can collapse back to the center portion and the weight will quickly drop about 0.5". This could add some additional stress on the line. I can only recall one line breaking in 6 years of building clocks. It broke at the knot. There is probably a fishing line style knot that would be much stronger than a simple overhand knot. I use Spectra Power Pro Microfilament Line. My first spool was 65lb test. My current spool is 80lb since it was the same price as 65lb. My quick strength test on a single strand broke at the knot with much more than 20lb. My guess is somewhere around 50lb of force. I use moss green but cannot imagine that it is any different than red except for the color. The pulley should double the effective load carrying capacity. With the proper knot, a loop of 65lb line should be able to support a 130lb dead load. Any jerking motion would reduce this, but it should still be way above the expected 10lb load with some safety margin. I looked up some fishing line break tests. The first random web site at Fishing Line Strength Test & Chart | Sport Fishing Mag showed about 40 different brands of 30lb line. All broke between 26.5lb and 66.8lb loads when wet. They estimated 8-12% increase when using dry line. They mention an improved Bristol knot. I will look this up to see if it helps.

The weight shell was originally designed with thin walls to reduce print times below 20 hours on an older printer. I have at least a dozen weight shells without a single break, but I have never dropped one on a tile floor to fully test the indestructability. 😀 The newer faster printers have made me think about increasing the walls from 4 perimeters to 6. The primary purpose of the internal ribs on the weight shell are to provide support for long bridges. They build up using a minimum amount of material to maximize the internal volume. Any added strength is a bonus. PrusaSlicer integrates so well with Prusa printers that I have not switched to Orca Slicer. Scarf seams look like a huge improvement, and I am eagerly waiting for it to migrate into PrusaSlicer. Did you try using it on the gears? I am curious if a scarf seam on the gear teeth would stay smooth.

I have not tried it, but it should work.

Good to hear. I should expand the section in the manual about debugging this issue. The central arbor has the most parts stacked up, so it seems to have the highest probability of causing issues. Steve

My first guess is to double check the pre-assembly steps. The critical checks are the ones for gears 5 and 6. Also, look for end shake on the central arbor stack. There are a lot of components. If the slicer or printer "rounds up" on all the sizes, the entire stack can be too tall and will be pinched by the frame.

I just purchased an MK4 and was amazed at the 2.5X increase in printing speed, so I upgraded my two MK3S's to MK3.5's with about a 2X speed increase. I don't know if I will continue listing the print times. The amount of filament used is a good enough indicator of the approximate size and complexity of each part. All clocks are designed to work with springs from common ball point pens. The nominal pen spring size is around 0.4x4x25mm (0.4mm wire diameter, 4mm outside diameter, and 25mm overall length). Anything close to this size will usually work. The primary criteria are they must fit over a 3mm arbor on some of the clocks. The maximum outer diameter is around 5mm. The Moon Phase Clock parts kits have 0.5x5x25mm springs for the friction clutches and 0.3x4x25mm for the ratchet. This gives a bit more strength for the friction clutches and makes the ratchet quieter compared to standard pen springs. Steve

The Moon Phase Clock is my favorite design. I am eagerly waiting to see your customizations. The dial has a large blank canvas, and the bob can also be customized. STEP files are released for both parts. Let me know if you want any other STEP files. Steve

Great videos. Thanks for posting.

You can pause manually for the color change. There is a bit of flexibility if you are late on the ideal time to swap colors. The ivory color dial has 12 layers. It should still look good if you intercept anywhere below the 10th layer. The numbers have 4 layers and should get a color change within the first 2 layers, so it has 2 complete layers with color.

The design was completely optimized to be easier to build. They are not interchangeable.

There were multiple messages. One said you could do 202mm without modifying the limit switches. Another showed the full dial printing after modifying the switches. It seems like you have several options with possible solutions. Steve

That's awesome news. It looks like the part is touching the top edge and extending 3.2mm into the handle section. Great job making it work. After this part and the back frame, everything else will be easy. Steve

202mm would require a 99% scale factor, which is the same amount of effort as the 98% scale factor that allows you to use the original slicer that you are already familiar with. There is minimal risk with either scale factor.

Hi Alan, This must be frustrating for you to have a printer that is just a tiny bit too small. I design clocks to fill the build plate on a Prusa MK3/4 (250x210) or Ender 3 (220x220) printers since they are so popular. The Bambu Lab printers are 256x256mm, so they work as well. This clock might someday receive a 75% scale for a Prusa Mini or Bambu Lab A1 Mini. The dial might get an 80% or 85% scale but would still end up smaller than 180x180. I am not actively working on it right now and I have a lot of other projects ahead of it, so for now we should figure out how to get the existing design printed on your printer. The back frame and dial are structural components that need to support the weight shell. They cannot be split down the center. The best cut would be to shave off a 3.2mm sliver away from any critical components. The best open area on the dial would be a 45 degree cut between the 7 and 8. The back frame could cut away 1.6 from the top and bottom or both sides. My personal opinion is that the clock would run perfectly with the 98% scale factor. And it would look better also. Modifying the slicer settings to add the 4mm would also be a great option if you can get it to work. I do not see Qidi in the list of supported printers for either PrusaSlicer or Cura. It seems surprising since Qidi printers look to be incredibly fast for the price. The Qidi website says their slicer is based on Cura. Steve

Hi Alan, The moon phase clock back frame and dial are both 8" (203.2mm) circles. I ran a test print on my Prusa MK3.5 printer. The default print size is 250x210mm, but it prints the initial purge line at Y=-4, so this tells me the printer is actually capable of printing at least 214mm in the Y direction. Here are the changes required to print a 10mm x 213.2mm rectangle on my printer. I use PrusaSlicer in Advanced mode. Select Printer_Settings => General => Bed_Shape. A dialog box pops up and I changed the Y size from 210 to 214. The origin was changed from 0 to 4 because I know that the startup G-code will print the purge line at Y=-4. Here is the setting under Printer_Settings => Custom_G-code => Start G-code. Scroll down near the bottom and it shows the purge line at X=0 and Y=-4. Then I went to the main screen and added a 10mm x 213.2mm rectangle onto the plate. The skirt is turned off. It successfully printed without hitting any end stops. Other slicers should have similar abilities. If the initial purge line is outside the specified 270mmx200mm print size, then you know that the printer is physically capable of moving more than 200mm in the Y direction. Steve

8" (203.2mm)

Hi Alan, Yes, I think it can be done. There are only two components that will not fit, the dial and the back frame portion behind the dial. Both are 8" (203.2mm) circles. Here are a few options: Chopping both parts and gluing on a 3.2mm sliver is an option but will leave a visible scar on the dial. Many printers use a portion of the bed below the reported 270x200mm area to print the purge line. You might be able to tell the slicer that you have a 270x204mm print bed. The clock should easily be able to handle a 98% XY scale to get under the 200mm limit. Scale the frame, gears, and hands. Leave Z at 100% or else you will also need to reduce the arbor lengths by the same amount. Leave the pendulum and weight shell unscaled. All the arbor holes need to be drilled to fit even without scaling. The screws will be 2% tighter but should not be a problem. The bearings might be a tight fit. You may need to sand the bearing holes so they can go in loosely. The biggest risk is consistency in making sure ALL parts that need to be scaled are properly scaled. One gear at 100% might bind up. Steve

Thanks for posting and for the positive comments. The colors look great. Nice contrast and easy to see the hands.

I just updated the links at https://www.stevesclocks.com/sp6 to include better references pointing to the controller. Refresh the page and download the CNC Shield V4 Addendum showing the full details on how to build the controller using parts from Amazon. Eventually, I designed a smaller custom board. There is a link to Etsy for the new controller with options for bare boards or fully programmed and ready to plug in to the clock. Steve

Hi Alan, Stepper motors are perfect for driving a clock because if the motor is rated at 200 steps per revolution and you give it 200 steps then it will rotate exactly one revolution. A clock needs so little power that it is very difficult to overload the motor so it will miss a step. Any time error might be caused by the reference clock timing. That is why I started using a DS3231 precision time base in my "Silent Shield Clock Controller". It has an accuracy of around a minute per year. The SP6 clock was originally designed with a custom resistor based stepper motor driver. It needs to use a stepper motor with 30-ohm coils like the one you referenced. It also uses the built-in Arduino Nano time reference that needs to be manually calibrated for each clock. However, the old controller has been upgraded to the significantly easier to use and more accurate silent shield clock controller. I strongly recommend using the new controller. It uses the more common stepper motors with 2-3-ohm coils. One example is 17HE12-1204S. The key features are 2-3-ohm coils and a 4-pin straight header (usually black) on the controller end. A connector at the motor end is nice, but not required. The default algorithm for SP6 expects a 200-step motor, which is the most common. The crazy gear clocks are slightly quieter with 400-step motors, but also support 200-step motors. Steve