sangi39 wrote: ↑15 Apr 2024 01:05
Finally have the
feeling of having my own space (no-one looking over my shoulder), so can make a start on what should end up being a handmade pendulum clock:
It's coming along. I
think the gear layout will work, and this should, I think, be what they look like face-on. I'm missing the second gear on the middle right, mirroring the one off to the side on the middle left, but that's I, to start, forgot to make it, and then decided to not make it, just in case I had to mess with the layout a bit, and couldn't keep the symmetry (looking at it, though, I should be able to fit it in, which is quite nice)
I think I've worked out the correct "depth" for each gear, relative to nearby ones, so that the overlapping bits that aren't where teeth should actually meet don't cause the entire mechanism to just jam, so next step is to actually test that out, and see what happens
EDIT: Just realised the "second" gear up in the top left is actually in the wrong place. The next gear along connects to the inner teeth, not the outer teeth
EDIT2 (29th April): Yeah, I had to lose the symmetry in the main gear train. The smaller driving gear attached to the "day of the week" gear (third from the bottom... I should probably label these on diagrams...) has to go at the back of the mounting to connect to the next gear down (second from the bottom), but that means that it (the "day of the week" gear) needs to be moved further away from the gear physically directly above it, which results in the intermediate gear in the gear train (the one physically off to the left) crossing over the central vertical axis, meaning that it can't have a gear directly mirroring it (the clearance is also pretty small around there as well, as it turns out, e.g. half a centimetre between a couple of gears, but that feels manageable)
What I
might do to kind of bring it back is have what originally the mirroring go directly behind where it was, and have it take the role of a drive gear with a weight attached, which would make it somewhat central within the gear train, which might be okay. It also means that the pendulum and the weight are both then behind the gear train and the displays, which feels sensible, in case they need fiddling with at all
EDIT3 (1st May): Think I've got something I like. As before, it'll be a case of building up a physical "test" version to make sure the dimensions work out (I mean, they should, it's a really minor change), and then try and make a version that
works (at least in principle). Main limitation is still tools, resources, and skills, but I think I'm getting a better feel for the clock
in general, and how is should all come together, which is a big step from a month ago when it was still just, effectively, a 2D diagram
EDIT4 (3rd May): So actually trying to produce the "depth" model, I realised that almost every single gear can be directly mounted on a single back board without any overlap (as long as the axles are narrow enough), and all but three gears can be held in place by a single front board (the three that can't all need holding place in the same plane, and have roughly similar surrounding environments, so holding them in place should be fairly easy)
The main problem came from the 12hour and 24 hour hands, with the former needing to be mounted above the latter. I'm fairly certain I can think of a way to do this, but this has lead me to abandoning a distinct 24hour gear, and switch wholly to a single 12hour gear
Strictly speaking, the gear that juts off to the left is a 24hour gear, turning once per day, so I could modify how I want days to be displayed down there, but up in the top right, it'll just be 12 hours, rotating twice per day, which should just be easier to make, probably easier to read since it's what people are actually used to