Rattleback

[bergerud]

Here is a curious thing for a weekend project. It is called a rattleback. If you try and spin it, it will wobble and rattle until it changes direction and spins the other way. It will spin fine one way but not the other. (Another one of my student demonstrations.) It is a forgiving double sided carve since the sides do not need to line up exactly. If you have a nice small piece of exotic wood (denser is better), this is the perfect use for it. There are two versions, one nice and smooth and the other a "voodoo" version. (The pictures are of a smooth one made of rose wood and lacquered.) They are very easy to finish and make great little gifts.

This is my own mathematical surface that I created in a math program called Maple. Imported using the STL importer.

http://youtu.be/OHCc22r329Q

[brdad]

I have some of these in plastic, they are interesting. I often wondered how a larger one would work, now I can find out. Easy to see these are much classier in wood. Thanks for sharing.

[aokweld101]

what are the uses for the rattleback, never heard of it. is it just to look at?

[bergerud]

what are the uses for the rattleback, never heard of it. is it just to look at?

It is a 200 year old toy to amuse children. Its strange behavior still intrigues scholars today. Scientific papers are still published as people try to computer model the behavior. Can you explain why it does what it does?

[brdad]

I had some free time tonight so I carved the mask version of this tonight while cleaning up the basement shop. I used a scrap of maple and it came out nice, and I think I'll make a few more, probably the smooth ones but with darker wood and centerline names or windings of some sort. I think I'll also make a simple sled so I can carve them at 90° to cut down on a few minutes of carving time.

I questioned why the feather was not set to 0 to also help with carving time, and was surprised to see it actually estimated a longer carve time with 0 draft. But, carved at 90° setting the draft to 0 decreases the time, I'd guess because there are several passes on each side of the project across the X axis that serve no purpose than to carve the draft.

Anyway, thanks again for putting out this project.

[bergerud]

The feather is just to be easier on the machine. I do not like the bit plunging at the beginning and hitting a wall of wood on each side as it goes.

I like the idea to carve it sideways. I would worry a little about the line up. The x cannot be trusted to be as accurate as the y. Let me know how that goes.

[aokweld101]

Bergerud< [got the spelling right] ..lol...will oak wood work on this project?

[bergerud]

Sure, oak is fine.

[aokweld101]

thanks, I have some oak to try it out.

[brdad]

I've had carver's block since my wife and I got back from our 8000+ mile drive across country and back (so much to see! (http://s235.photobucket.com/user/brdad/library/NW%20Trip)), so I decided to try these at 90° on a sled yesterday. It worked very well, I was able to carve two at once from a scrap of bubinga. The software estimated 30 minutes carving time on normal with the feather off; I forgot to check the actual time but I think it was close. I also wish I had chosen best or optimal as they did take a bit more sanding than the first one I carved. There was a slight variance with the two sides in the X, but it sanded out easily and might not have been as prominent had I not used the normal setting.

I think my next attempt at these will be the smooth version with a conforming vector design of some sort on them. And I definitely will do them at 90°.

[bergerud]

Looks nice. Good idea to carve it sideways. I think I may try that. I would make the sled so the piece can be flipped without releasing the sled. That way the x line up does not depend on the board sensor measuring the board the same twice.

[rogermay]

I'm wanting to make a rattleback out of wood and would love to see your files. I've downloaded them, but can't see anything. I notice they're in mpc format, which I can't use as I've got a mac. Is there any other way of posting them?

[bergerud]

These files are for the Carvewright machine. If you had a CW machine it would carve it. What kind of file were you hoping for? You could download a 30 day trial of the CW software at:

http://www.carvewright.com/support-page/getting-started/update/

You can see the rattleback as a 3d model in this software.

[SteveNelson46]

.mpc files are used on a PC and a MAC. If you can't see the .mpc extension you need to turn extensions on. Either way it will still work if you have the Designer installed.

[ktjwilliams]

They are interesting and amusing .... Oh and I can explain there behavior but I'll never tell ...

[blhutchens]

Does anyone..... have an idea of how to make one of these that is a "rare celt" that will change rotation when spun from either direction and change directions several times during one spin?

[ktjwilliams]

Not me but I think this was Dan B's thing ...

[bergerud]

I know that one can make rattlebacks that rattle back both ways. My thought was that the "aggressiveness" of the rattle back would necessarily be less, and since I am proud of how aggressive my design is, I had not tried to make one.

I think a dual rattleback may have to be shorter so that both states of oscillation are similar. Maybe they only need to be commensurable. It could also be that the density of the material and not only just the shape is important.

When I get some time, I will have to research this. It is an interesting question.

[ladjr]

I have made them both ways. I gave the left spinning to a republican and left spinning to a democrat.
81661 81662

[DonCP]

Just made one of these out of scrap pine.
This thing is cool.
Works great!!

Thank you!

[bergerud]

An STL of the rattleback has been requested in the past by members and I had emailed it to them (since the file size was just a little too large to upload). I went back and slightly reduced the number of vertices to reduce the file size to upload. This is a shape plotted in a program called Maple from a formula I made up.

This is the smooth version. If you want a pattern on the top, you will have to add it yourself.

[oscarl48]

Dan, thank you. Beautiful design.

I may have to give this a try. I'm carving a cabinet door decoration this afternoon and if I have time I'll throw into the project line.

- Oscar

[brdad]

I am still making these when I get an urge to do so, the last ones out of hickory which looked nice. Usually I just give them away. I am still carving them sideways with good results and should make a dedicated sled for doing that.
As far as going both directions, I've considered making one double-sided - flip it over and it spins the other way. Not sure how the balance will be but easy enough to try.

[oscarl48]

Dan, a quick 21 minute carve and 5 minutes sanding in oak. I get an average of three turns in the reverse direction and occasionally up to 4.5 turns with mine. Very cool design.

[oscarl48]

Dan,

I have one of your rattlebacks on my desk and every so often I give it a twirl. Don't really know why it fascinates me so much. The other ones I've made I gave to friends with small kids.

But the spinning got me to thinking. These oddities can rattleback both in clockwise and counterclockwise directions.

So I am going to try my hand at making my own design. I made a quick model and then mirrored on axis to shift the center of balance. I am hoping since they are mirror images they will have opposite preferential spin direction.

I will throw into the machine tomorrow and see what happens. :D

Who need fancy math software when you can eyeball it. lol. Last words of really expensive experiments (hubble telescope, mars explorer, Solar Heliospheric Observatory, and many more).

Even if it doesn't work I now have a 3d model of a rattleback for cool rendered images.

[bergerud]

Let us know how that goes.

[oscarl48]

First attempt: complete failure.

I screwed up and re-shifted the center of mass back to the center when I was cleaning up the mesh without realizing it. So I pretty much carved a symmetrical pattern. It likes to spin clockwise perfectly and it likes to spin counter clockwise perfectly. If I put a twist on it, it would make a great propeller. :)

Back to the drawing board. This time I am not going to re-shift the center of mass.

[oscarl48]

Well, success. They work the way they should with off center of mass rotation. Preferable rotation is opposite of each other. After a sanding I applied some lacquer and waiting for them to dry.

Please let me know if anyone wants these patterns.

But honestly, Dan's pattern is more elegant and it is also a lot more aggressive than my version. You can get the same results from Dan's by mirroring the pattern so that you have one of each with different rotation preference if you want.

[bergerud]

I think your patterns are too offset. The offset only needs be subtle. Just think, all the action happens in a small neighborhood of the contact point.

[oscarl48]

You are right. My thought was to go with a gross starting point to set the outer boundary.

Since I am eyeballing it, it would be a series of tweaks until I got it just right visually while retaining the toy's aspect.

It was a fun little experiment but I think I have what I want from it. A simple understanding of the physics associated with rotational behavior in asymmetrical mass bodies.

The right one is from your pattern (it rocks). The left ones are my experiment and they have a clockwise and counter clockwise preferential rotation.