Oct 16, 2026, 2:10 PMMath is amazing! Those brackets worked on the first try and the 900mm cross braces make the geocentric orrery look like clockwork alchemy diagram.
- Oct 16, 2026, 2:10 PM
Replies
- Oct 17, 2026, 7:27 AM
The larger multi-piece ring gears are more sqooshy than the single piece versions and this is making the mechanism jump rather than rotate smoothly. With a 2kg weight, the moon ring deflects from 66 to 61cm, about 92%. Going to try making a version that is more resistant.
- Oct 17, 2026, 9:48 AM
Doing a science to try to quantify how much flex different ring gear sections flex.
- Oct 17, 2026, 11:55 AM
That's better! Added some decorative webbing and now the deflection is about 1/8th what it was with the old ring gear design.
- Oct 17, 2026, 6:53 PM
Since I'll need to reprint the big ring gears to make them more rigid, I can also apply some lessons learned and redesign the gear train to eliminate the enormous 85 tooth zodiac gear.
- Oct 18, 2026, 7:22 AM
I do wish FreeCAD had a way to add text labels, other than on dimensions. It would be very useful to be able to annotate complex sketches to explain why things are laid out in certain ways.
- Oct 19, 2026, 8:54 AM
The printer skipped a step or something on a middle layer, leading to seven thousandths* of runout on this gear.
*: Of a meter.
- Oct 20, 2026, 10:19 AM
The additional structure and webbing on the ring gear really improves rigidity and makes the entire mechanism turn so much more smoothly than before. I'm still experimenting with how to display the moon phase on the ring, which is why each one is a little different (and the colors are random)
- Oct 21, 2026, 4:30 PM
I walk away from the printer for ten minutes and look what it does. "Printing completed" no it's not.
- Oct 22, 2026, 10:55 AM
FreeCAD's constraint solver helped me find the angles for this complex bracket and it worked on the first try (once the printer successfully produced one). Those aren't nice 30 degrees; they are 34.9 and 25.1 because of the 20mm offset in the axle spacing and the need to meet up with a T bracket on top.
- Oct 22, 2026, 11:27 AM
It has not been a good day for our printer.
- Oct 22, 2026, 11:28 AM
The one that worked fits quite nicely however!
- Oct 22, 2026, 6:22 PM
Maybe it's just bad filament or K factor settings? This gold pla seems to work fine.
- Oct 23, 2026, 12:33 PM
The new moon ring gear looks mighty fine and is much more rigid! It makes one rev every 29.53 days and has 28 spaces for shiny moon phase indicators to be installed
- Oct 23, 2026, 4:09 PM
Mr Pythagoras grudgingly allowed a test print of the 12-piece zodiac ring.
Oct 16, 2026, 2:23 PM@th do you want The Dark Crystal? Because this is how you get The Dark Crystal
Arthur Elsenaarartelse@mastodon.socialOct 16, 2026, 2:40 PM@th You're becoming a modern day Eise Eisinga. Not exactly the size yet, but getting there ;)
- Oct 16, 2026, 3:00 PM
@artelse Eisinga is my inspiration! The plan is to hang it in the living room like his planetarium.
Oct 16, 2026, 3:12 PMBravo! With your mad mathematical skills I would be unfazed if your encore project was a recreation of the Antikythera mechanism.
Ryanfullywoolly@mastodon.socialOct 16, 2026, 6:12 PM@th That is a thing of beauty!
rag. Gustavino BevilacquaGustavinoBevilacqua@mastodon.cisti.orgOct 17, 2026, 12:47 PMThe "decorative" parts of the steam era weren't there just for decoration.
Jeff Zugale - Z Factoryjeffzugale@mastodon.artOct 17, 2026, 7:28 PM@th just want you to know I’m learning a lot from this. Made my first ever gear train by the seat of my pants last year and it works great, but I do want to be more academic about it 😊👍
- Oct 17, 2026, 7:50 PM
@jeffzugale I'm just making it up as I go with the motto "measure twice, print thrice"
Frenshapefrenshape@beige.partyOct 18, 2026, 7:37 AM@th I've wanted that fairly often. The current release candidate has better text support, which is nice, but what I've been wanting is the ability to add a little comment that says "this does X because of Y"
Quixoticgeekquixoticgeek@v.stOct 19, 2026, 9:00 AM@th seven mm ?
- Oct 19, 2026, 9:04 AM
@quixoticgeek yeah, 7mm of runout means the part is totally scrapped. the two bearings don't line up very well either, so the axle barely even fits. you can see the discontinuity from the layer shift around 10mm down
- Quixoticgeekquixoticgeek@v.stOct 19, 2026, 9:12 AM
@th that's a lot of run out for what looks like a small layer shift.
- Oct 19, 2026, 9:25 AM
@quixoticgeek the gear is 20cm, which means that 7mm of runout at the edge of the gear is about 4 degrees off axis. the two bearings are 20mm apart, so I think that only requires 0.7mm misalignment (by similar triangles). The layer shift appears to be about 0.5mm of skew, so it's in the ballpark (plus I had to use a threaded rod rather than the actual 8mm axle, which adds at least 1mm of wobble).
- Quixoticgeekquixoticgeek@v.stOct 19, 2026, 9:31 AM
@th ah that would do it. Go up a bearing size, and bore out the center to fix the concentricity?
- Oct 19, 2026, 9:34 AM
@quixoticgeek just two problems: t a) I can't swing a 20cm gear on my tiny lathe's 50mm chuck and b) does boring 3d printed filament even work?
- Quixoticgeekquixoticgeek@v.stOct 19, 2026, 9:40 AM
@th clearly you need a bigger lathe... And yes.
Pseudo Nympseudonym@mastodon.onlineOct 21, 2026, 4:46 PMHow did you get a picture of my 3d printer?
I've found baking the filament to extract moisture usually helps the stringification fails.
PalmAndNeedlePalmAndNeedle@norden.socialOct 21, 2026, 5:34 PM@th easy. it's just some logic and memory strapped to a glue gun. You're basically talking down to a toaster.
- Oct 22, 2026, 11:25 AM
@th are these in a repo somewhere? I'd love to see how you're constructing these.
- Oct 22, 2026, 12:30 PM
@phooky I haven't figured out a good way to check the files in. the FCStd file is about 50 MB at this point so it doesn't play well with source code control. under the covers it is just a zip and I can see that the BRP rendered version of some of the large ring gears (288 teeth!) are like 20 MB each. I wonder if there is a way to manage the internal files individually to make git happier.
- Oct 22, 2026, 12:42 PM
@th man, why are these things so huge? [unzips FCStd] oh no
- Oct 22, 2026, 8:37 PM
@th When I've had problems I've been able to resolve it by drying the filament thoroughly or cleaning the nozzle
Ali MirjamaliAliMirjamali@mastodon.socialOct 22, 2026, 1:05 PM@th It is strange to see that this happens so frequently on your setup on a relatively premium printer (Bambu Labs)? Although a bed slinger rather than CoreXY, still one of the top of the line printers with linear bearings and high quality components. Is the material something fancy?
- Oct 22, 2026, 1:54 PM
@AliMirjamali it is a used A1 from a friend who upgraded to a larger unit, with nearly 2500 hours of print time. I'm not using fancy or even premium filament, and don't have a dryer, so this might be a material problem.
Oct 17, 2026, 2:58 PM@th at first I thought you were measuring the tape measure using a dial indicator, so for a second there I could feel my blood heating up.
- Oct 17, 2026, 3:19 PM
@mk the test was initially setup with the dial indicator and when the first printed piece deflected 8cm I realized I needed something with a bit longer range...
