Sandball Machine

Files, photos and discusions on Tickers and kinetic devices.
Post Reply
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4574
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Sandball Machine

Post by ArtF »

Hi All:

  Heres a quick video

https://www.youtube.com/watch?v=ojiINT6 ... e=youtu.be

  on a sand-ball machine I'm putting together from spare parts.  It takes only 4
gears ( Gearotic laser cut in my case) a few pieces of flat plastic, 2 nema 17 motors ( which
I had laying about), 1 arduino nano ( $1.00 from ebay) and 2 drivers ( $1.00 each from ebay).


Art

  I
User avatar
kit
Old Timer
Posts: 85
Joined: Sat Apr 02, 2016 3:51 am
Location: Tasmania
Contact:

Re: Sandball Machine

Post by kit »

Art,
I've only just noticed this on the forum. A great example of what you can throw together from the scrap box. My wife has plans for me to build a version of the Sysiphus tables being sold by Bruce Shapiro but I have no idea when this will get to the top of the projects list, though I might get the funding for a bigger CNC router if I decide one is 'needed' to build it.

I think I've got the basics of the mechanics worked out from the available videos online but have no idea what kind of software is needed to drive it. Obviously it's based on continuous repetition of simple repeated moves but where do you start with generating those moves? Is this based on an existing library or drawing program or do I have to start learning the mathematics of computer drawn vectors from scratch? What do I need to convert the computed moves into step and direction pulse trains? I'm assuming you didn't just rattle off 100,000 lines of code on a rainy afternoon for your table... Then again, for you that would be a piece of cake ;)

Kit
User avatar
tweakie
Old Timer
Posts: 170
Joined: Wed Dec 01, 2010 12:58 am

Re: Sandball Machine

Post by tweakie »

Hi Kit,

Don't forget that Auggie is already there for you. Although it has been designed for laser use there is no reason it cannot be used for rorary tool engraving, cutting, machining, etc.

Tweakie.
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4574
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Sandball Machine

Post by ArtF »

Hi Kitt:

>>Is this based on an existing library or drawing program or do I have to start learning the mathemati cs of computer drawn vectors from scratch? What do I need to convert the computed moves into step and direction pulse trains? I'm assuming you didn't just rattle off 100,000 lines of code on a rainy afternoon for your table... T


  The driver of that thing is a $1.00 arduino nano. It doesn't have much memory so I couldn't implement
a full planner. It took a couple of hours to code up. The Arduino I used is too weak to allow for a
Cartesian coordinate system, so I instead relied on a polar one. That makes it simple.

  The gear mechanism, if you look at it is pretty simple, but also allows for a polar type of equation that
will allow one to make patterns. (Ill post the code so you can use mine or modify it if you wish.)

  If you look at the attached photo, youll see how I decided to minimize the problem. Consider the
left stepper motor, it has one job. It rotates the magnet wheel ( magnets are attached to a stick glued to the top of the largest gear, they are arranged so that the magnet when rotated will exactly cross over the top
of the right hand gears shaft. If the left gear turns but the right doesnt, the arm with the large gear will rotate
as the right shaft has that arm on a bearing.

    The right gear, as it rotates, only rotates the large gear.

  Now, because of the gear ratios involved, if you rotate the right gear at a rate of 1/8 of the speed of the left,
the large gear will be swung around the outer radius and appear to not rotate on its own. It will maintain a
static position. The magnet on top of the table will now simply do circles around the outer perimeter.

  This extra speed on the left gear is necessary as the large gear will rotate as it goes around the stationary
right gear otherwise.

  If you rotate the right gear at 1 1/8 speed of the left, the magnet will appear to draw a straight line across
the table, thus making it look as if you have a Cartesian coordinate system and just drew a straight line.

  On mine I simply define one cycle as rotating the arm 1/2 full turn while rotating the outter gear 1 1/8 turn..thus a perfect straight line across.

  Now, if I randomly select a number and add it to the large gear on every cycle I get a repeatable pattern we can name as Random Pattern #n where n is the number of steps added. This means the ball will drive straight across plus a little bit. Figure out how much extra in terms of degrees and one can figure out the number of times to repeat to complete the spirograph and end at start.

  There are many ways to do this type of thing I suspect. I opted for a quick Spirograph only device
just because it allowed me to use the junk drawer and not write a full planner for it.  A buck for a cpu seemed
ideal, the drivers were also a dollar each.

I haven't finished it yet, but it does look kinda cool while running. I was curious as to how it was done
more than anything else.. originally picturing some sort of slip ring thing, I was surprised the whole thing
could be done with just 4 gears.

  I too saw Sisyphus but the videos didn't show me enough to explain how it worked.
(That drives me crazy and I have to figure it out...thus this prototype.)

  I still don't know how they do it, but I doubt its possible to do it in a more
mechanically efficient way or with fewer parts. But thats not to say my way is
the best way..just the only way I could figure to do it quickly.

    Let me know if you'd like the arduino code, but it isn't complete..more the
code wanderings of that afternoon as I tried to figure it out. I kinda think it'd
be better with a Cartesian coordinate system planner in it, but I cant fit that
in less than a Arduino due or a raspberry maybe. At this point planners hold
few mysteries for me, so I think I could fit one in a larger arduino but dont
know when Ill get to it.

If people wish to build this thing Ill do a better video explaining it and the
ups and downs of building it.. only takes a day or two to build it if you have
a laser engraver.. longer if you have to make the gears from wood.

  Noise for example is one variable, unless well supported, ( or glued on
a rubber bed) the bad can act as a drum amplifying noise. Larger tables
would probably be quieter.. takes a bit of futzing with ball size and magnet
strength as well as you can pull the ball right down to the plexi top and get
a grinding effect.

Art

 

   

   



 
 
Attachments
polar (2).jpg
User avatar
kit
Old Timer
Posts: 85
Joined: Sat Apr 02, 2016 3:51 am
Location: Tasmania
Contact:

Re: Sandball Machine

Post by kit »

Art,
Thanks for that explanation of a very cunning but simple mechanism which can easily be reproduced using gearotic and a CNC machine. Simple is always the best.
The link below includes a picture of the mechanism in the Sysiphus tables. It's a far more complex structure than yours, including a rack and pinion for true polar co-ordinate input to the two steppers though this may ease the calculations for converting x,y to polar for inputting text and graphics which these tables can do. It uses a Raspberry Pi as it's driver so a lot more potential complexity in the patterns.
Having just tried Vexx for the first time tonight I can see how we could reproduce all of that hardware as well. You never cease to amaze us all.

Tweakie,
I did wonder if driving the steppers might need interpretation of the source pattern code by an intermediate step not unlike MACH3, Auggie or LinuxCNC but the pattern generator is unlikely to output G-code! Then again...

I've no idea when I'll get to this project as there are enough half-finished things lying around the shed as it is, but I do have some spare little NEMA17 motors, EasyDrivers, Arduinos and a Raspberry Pi with nothing else to do. And a CNC router that keeps begging to cut something out.

https://makezine.com/2016/10/02/gorgeou ... e-shapiro/

Kit
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4574
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Sandball Machine

Post by ArtF »

Kirk:

Thanks for the link. I do think the rack and pinion would make the equations easier, but
if you had the power of a raspberry pi, Id think the additional computations necessary
for a dual polar wouldn't be too taxing. My simplification of the mechanism would seem
not a bad idea if one can do the code for the switch from dual polar to x,y Cartesian.

  I havent ordered a pi yet, but I see it runs win10 which makes it interesting to me..

Art
Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests