Davidz90

I wonder if it's possible to build something like this but with 8t gear. Did some experiments, I could do a single direction stepper no problem, but two-directional one eludes me. Also, rubber connectors are great for replacing rubberbands when flexible components are needed.

I agree, IMO it is a technic set, and one of the better ones recently. I'm working on a motorized version right now, although wing flapping takes a lot of power; that's why regular helicopters are somewhat more practical :P

Well, axle is "achse" in German. These two languages are surprisingly similiar, leading to many interesting misspellings on both sides. Another early review, here print alignment seems pretty ok:

If you are talking about Earth's axis precession, it is 1 rotation in 26000 years so basically stationary for all intents and purposes. So the part does rotate, but in the sense that it remains fixed in relation to the base, properly demonstrating seasons. I'm eager to get this set and do some improvements. Too bad about the print mismatch, and I don't understand why there are pins at all - axle going all the way through would hold two hemispheres just fine?

And good thing it isn't, the forum is already saturated with vehicles. It definitely has inspired me. I've also considered building a lathe, with a turntable-based chuck. In my case it would be even smaller, which could possibly allow it to work with wood and even soft metals like aluminium, despite limited power. The lack of motor power can be compensated for to some extent by adding a heavy flywheel and doing the cuts in small steps, allowing the lathe to spin up. Another possibly interesting thing to do is a drill press.

Wow, this is amazing! Surprisingly strong too. Very neat design of the chuck.

Thanks! I'm going to use it in science fairs etc. in my university, but contacting some science museum is a great idea too. One of my Youtube colleagues (Darrell Aldrich) did exactly that with his clock.

Some extra info on various parts/mechanisms: 1. The auto-rewinder uses a magnet to close the circuit. This ensures that electrodes actually stick together instead of being 0.01mm apart due to a random speck of dust, which was a major headache in trying to trigger motors reliably. 2. The chime mechanism uses a set of four custom cams (basically slope pieces attached to 4x4 round brick with electric tape) to encode the melody. While there are several ways to do this encoding in a "legit" way (for example, with threads), none of them were as compact as this. and here is an earlier prototype, with thread encoding and gravity powered (note the fan used for speed regulation): 3. To get the necessary musical notes, I had to cut the pipes to proper length. The math behind this is remarkably simple: 4. The calendar uses 1:365 gear reduction and a custom dial with all months and days, attached to a large ring gear.

Thanks!

Very impressive! I thought that the new parts will result in a very large gearboxes, but this is remarkably compact.

I'm proud to present my newest build, which took me almost half a year: Grandfather clock with 19 different functions, possibly the most complicated Lego clock in the world. 20240210_133209 by David_Z1, on Flickr Standing almost 2 meters tall, this pendulum clock was an engineering challenge on multiple aspects, but the biggest problem was how to power all 19 functions and how to handle a highly variable friction produced by them. The answer was to use several electric motors triggered at the right time. There's no electronics, just mechanical contacts. Escapement - the central part of the clock that powers the pendulum, is powered by a small dropping weight that is frequently lifted back up by electric motor. This ensures a very steady input power, and thus good accuracy; the mean error is less than 3 minutes/day (after a day of working, it is off by less than 3 minutes). This is possible due to the use of John Harrison's grasshopper escapement, which is the most accurate type of pendulum clock mechanism. The electrical system is based on custom DC motors fitted with RCA plugs: 20240210_163033 by David_Z1, on Flickr Above You can also see the Westminster chime mechanism - at every quarter, it plays a melody like Big Ben. The chimes are made from aluminium pipes, 70-110 cm long, length tuned to specific musical notes. Here's the list of functions: Schem_front by David_Z1, on Flickr Schem_back by David_Z1, on Flickr And here a video demonstrating all of them:

Amazing! Recently I got the Yamaha bike and I'm planning to do something similar with it.

Wow, so cool! Very relaxing to listen to.

Merry Christmas everyone! :)

I'm back with results. As expected, the earth stays still as long as it is connected with the central gear with 1:1 total gear ratio, with even number of gear ratios. Only then the two motions are fully separate. To quote the presentation in the last post ": take the sidereal year (in seconds) and divide by the sidereal day (in seconds) to compute the number of spins Earth makes in one orbit…366.2564". For 366.2564 target, I get (among others): 1/60*12/40*(16/28-1/40)=1/366.0131 1/24*1/20*(8/40+40/36)=1/366.1017 1/16*1/24*(40/36-1/16)=1/366.1987 8/140*1/24*(1/28+40/36)=1/366.2284 1/16*1/28*(36/28-1/16)=1/366.2482 8/60*8/140*(1/40+8/24)=1/366.2791 1/20*8/140*(8/20+20/36)=1/366.2791 8/140*1/16*(1/20+20/28)=1/366.3551

I decided to dig into this a bit more. Assuming that everything is geared from the crankshaft on a fixed base, so that all the rotations are completely separate and counted in relation to fixed base, we need to use sidereal day and sidereal years (rotations in relation to fixed stars). According to this: https://multiverse.ssl.berkeley.edu/Portals/0/CalendarInTheSky/Resources/Presentations/HowLongIsAYear.pdf?ver=Wt-kX9xEaM0fZve8OXxBfg%3d%3d the proper gear ratio for the earth is 366.2564. So 364.5 is a little worse than I thought. I'll make a mockup to test if the rotations are really completely separate and run my program with the new target value.

Exactly. To quote Wikipedia: The Gregorian calendar improves the approximation made by the Julian calendar by skipping three Julian leap days in every 400 years, giving an average year of 365.2425 mean solar days long. As of 2008, a mean solar day is about 86,400.002 SI seconds, i.e., about 24.0000006 hours So yes, my bad, in this orrery the earth is not gaining additional rotation during the year, but one full rotation of earth, relative to the base, is not a solar day - its a sidereral day. So indeed, Zerobricks had it correct.

I don't think this is the case here. Earth is locked to the input axle/hand crank, which is not rotating and instead it is locked to the stationary base. If we would disconnect and lock the earth input, then over the year the earth would remain stationary and not do one rotation. But thanks for bringing up an interesting video - Veritasium is a good channel (even though he sometimes makes some minor mistakes in physics-related content)

Yes, with one powered input and one locked input, the differential casing has half speed and double torque.

It's in my earlier post. Some of the many, many options: 1/60*(1/20+16/140)=1/365.21741/140*(1/12+12/40)=1/365.21741/140*(8/24+1/20)=1/365.2174 in general, differential output is average speed of inputs. So, for example, if we have two inputs, both geared down so for 1 rotation of input we have: A: 1/3 rotations B: 1/2 rotations then the differential output is (A+B)/2 = (1/3+1/2)/2 = (2/6+3/6)/2=(5/6)/2=5/12 By gearing output 2:1, we get a simple sum (A+B) instead of an average. If inputs rotate in opposite directions, we have a difference (A-B).

Sun is not solid and it's rotation period depends on latitude; 24.47 days on equator, 34 near the poles. Basically anything between 1:24 and 1:34 is ok, so 1:27 used here is fine.

Assuming that the earth is doing 4 rotations per second (looks like it in the video), on the equator they reach 50% of the speed of light

I was curious if the 364.5 is optimal since it is so accurate with limited number of gear ratios. So I've written a computer program and the results are: up to 4 gear ratios 1/16*1/16*28/40=1/365.7143 1/28*1/28*28/60=1/365.867 1/36*8/36*16/36=1/364.500 up to 4 gear ratios, with differential 1/60*(1/20+16/140)=1/365.2174 1/140*(1/12+12/40)=1/365.2174 1/140*(8/24+1/20)=1/365.2174 I selected only few best ones because the output contained several hundred redundant solutions, like 1:3 done in several ways, the same gear ratios in different order etc. The 364.5 is, indeed, the best option with 3 or 4 gear ratios and no differential. Ideally, we want 365.2425 days; differential options are really close to that (error of 36 minutes)

Whoa, that is way more accurate than I expected. Also, great reverse-engineering!

Here is one from Lego Ideas (as I recall, it got rejected in the past, this is its 2nd or 3rd attempt, I think?). Full solar system, ingenious but quite bulky. https://ideas.lego.com/projects/c600ea70-0e2d-4192-8bbc-79eecbc796e4