doctorow — 2014-09-05T14:00:22-04:00 — #1
jorpho — 2014-09-05T14:07:16-04:00 — #2
It looks both sufficiently nifty and straightforward to construct that I'm actually kind of tempted to look into it further. This might be the first DIY project I've seen posted here for which I could say that.
sockdoll — 2014-09-05T14:57:31-04:00 — #3
One of those electrically-powered pseudo-perpetual motion machines - I like it! I also really like the intricate shapes and precision you can get with laser wood cutting and engraving.
Definitely a wonderful thing for happy mutants, thanks!
nixiebunny — 2014-09-05T15:01:56-04:00 — #4
This would make a good CAD project - figure out how to design it for yourself.
shaddack — 2014-09-05T15:07:40-04:00 — #5
And add a microcontroller that will adjust the precise angle (and therefore ball speed) of the ball-meander plate so the time is not drifting too badly.
jandrese — 2014-09-05T15:47:24-04:00 — #6
At first I was impressed at the low price of the clock. Then I saw that the price was only for the plans.
The writeup blames dust for the inaccuracy of the clocks, but that's just one factor. There is a limit to the precision you can achieve when cutting wood, and a limit to the accuracy as well. Both factors will add error over time to the clock and drive it off. This is actually true of all clocks, but is especially pronounced in a device like this where you have a lot of physical forces at work.
nixiebunny — 2014-09-05T17:45:36-04:00 — #7
This clock could be deadly accurate if the motor drive was triggered not by the ball itself but by a microcontroller (or any time divider) every N seconds.
shaddack — 2014-09-05T18:33:43-04:00 — #8
Yes. But that could lead to the ball reversing direction while still on the way, or waiting in the end position. Which, if the plate's own period is not sufficiently matched, would look not-so-good.
I propose a little different approach, where by varying the plate angle the ball speed is controlled so its mean duration finely matches the desired one. That way the clock will still drift a bit here and there, but the per-cycle drifts will cancel each other over time. Measure the time of ball travel, if too fast use lower angle next time, if too slow increase the angle. A bog-standard PID could do the job.
westcarleton — 2014-09-06T06:32:06-04:00 — #9
No need to do that; use a stepper motor and adjust the number of pulses the stepper gets. A very cool design (though over-priced), but how many people have a laser cutter lying around?
christopher — 2014-09-06T08:05:40-04:00 — #10
Hi guys, I’m the designer of this clock. The clock is a typical scroll saw project.
I don’t know, why the author of the article said, it was lasercut.
I made my clock with my self build CNC Router „Solidis“, but most people build my projects with a scroll saw.
You all right. With micro controllers or stepper motors you can build a clock which is much more
accurate. But my intention was to design a clock which is entertaining and a bit funny.
Inaccuracy at the cutting process is not really a problem at this clock because the ball rolls down always the same
way. So the inaccuracies are also always the same.
I like the idea of shaddack. At the moment you can regulate the angle also, but manually.
You should see a rolling ball clock as a piece of art with entertainment value
shaddack — 2014-09-06T10:00:54-04:00 — #11
I think that people are lazy and assume that lasercutting is the way when the parts for a project are too complex and the manual scroll-sawing would take too much time. The scroll-saw also explains why the cut edges aren't blackened, a telltale sign of laser (I thought they were just sanded down).
A visually interesting variant could be the same but lasered from transparent acrylic, optionally of different colors. With the sensing wires done using optical fiber instead (or very thin wires hidden with some optical trick); the optical method would allow using a glass (or other non-conductive) ball. (A cheat would be sensing it via a camera from a distance.) That does not solve the issue of the motor, which cannot be transparent. But that could be cheated using a water reservoir and valves and using water as the counterweight to actuate the movement of the ball plate. (If the amount of water regulates the plate angle, the timing of valves could be also used for controlling the period of the ball roll and maintaining accuracy. Or the valves can be mechanical, run by the plate inclined to the end position, and the angles optionally adjusted by servos controlling the end-stops.) The water can be cycled around using a small membrane or peristaltic pump. (Or just using a rainwater tank with enough supply, replenished with rains.)
Keep the ball rolling!
shaddack — 2014-09-06T10:03:21-04:00 — #12
Me, for example. In the age of $600-800 K40-III machines and their clones, the amount of excuses to not have one went rather low. Even the low-end one, with all its quirks and imperfections, is a very nice thing to have.
..now, someone please get an ytterbium fiber laser into this price range, so sheetmetal can be cut too...! Perhaps even direct laser engraving of circuitboards?
doctorow — 2014-09-10T14:00:33-04:00 — #13
This topic was automatically closed after 5 days. New replies are no longer allowed.