Small geometries and object space interactions do not work (e.g. gears)

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Add a small, somewhat complex component or group (less than 1 cubic 
inch).
2. Object space interactions (gears) do not work as expected.
3.

What is the expected output? What do you see instead?
Upon RUN:
(1) Object should fall and/or be affected by controls.
(1) Object does not fall. Object not affected by controls.

(2) Gears rotate within each-others object spaces (meshing gear teeth).
(2) Gears separate so that diameters do not intersect.

What version of the product are you using? On what operating system?
VistaHP, SU7, SP2/3

Please provide any additional information below.
It appears small geometries do not work in SP as they also do not work in 
SU (known bug).

View attached file in SP2. RUN has no effect.
Ctrl-A then scale by 2. RUN has no effect.
Ctrl-A then scale by 2 (again). RUN works.

Also note how the gear objects will separate (repel) each-other so that 
they can rotate autonomously. Objects moving within each-other’s space do 
not work (this appears to be similar to the concave curve restriction). 
Extending this idea by creating “bearing” holes for gear shaft tops and 
bottoms to keep gears from moving causes erratic, unpredictable results.

Is there a way to allow mechanically accurate involute gear teeth to mesh?

Thank you.
Mike.

Original issue reported on code.google.com by pane...@gmail.com on 12 Jul 2009 at 2:25

Attachments:

• [gears 1.skp](https://storage.googleapis.com/google-code-attachments/sketchyphysics/issue-29/comment-0/gears 1.skp)

[GoogleCodeExporter]

I have discovered a work-around solution to point 2.

The gear shapes were "cylinder"; they are now "convexhull" which provides for 
clean, 
accurate meshing of involute spur gears without repulsion. A “powered” 
hinge or 
motor control must be joined, in opposite directions, to each gear. The most-
accurate representation comes from assigning identical "accel" and "damp" 
values to 
each hinge/motor control. Note also that a “driver” gear will not affect 
a “follower” gear – both gears must be powered. 

Original comment by pane...@gmail.com on 13 Jul 2009 at 3:00

[GoogleCodeExporter]

I vote this up, small geometries do not work.

Create a cube, 1" x 2" x 4". Group, set as box, and run. The cube will fall 
into the abyss as expected.

Now scale the cube by 0.5 and run. The cube will remain where it is.

Scale the cube back up by 2 and run. The cube will once again fall into the 
abyss.

Further experiments:

A 1 x 2 x 2 cube falls. (4 cu in)
A 1/2 x 2 x 2 cube does not fall. (2 cu in)
A 1 x 1 x 2 cube does not fall.  (2 cu in)
A 1 x 1 x 4 cube falls.  (4 cu in)
A 1 x 1 x 3 cube does not fall. (3 cu in)
A 1 x 1 x 3 3/8 cube does not fall (3 3/8 cu in)
A 1 x 1 x 3 3/4 cube falls. (3 3/4 cu in)
A 1 x 1 x 3 9/16 cube does not fall (3 9/16 cu in)

The cutoff appears to be somewhere between 3 and 4 cubic inches.

Original comment by autoph...@gmail.com on 12 Aug 2010 at 5:24

Attachments:

• small-geometry-bug.skp

[GoogleCodeExporter]

Has there been any updates on this issue? This makes it incredibly difficult to 
create a design (to scale). If there is an update or workaround would someone 
post it to this thread?

Original comment by DRWhit...@gmail.com on 29 Aug 2011 at 6:23

[GoogleCodeExporter]

I used the UI to change the density of the object from 0.2 to 2.0 and it seems 
to have resolved the issue

Original comment by Baus...@gmail.com on 7 Dec 2011 at 10:59

[GoogleCodeExporter]

I can confirm the small geometry error.  I can also confirm that density of the 
object has something to do with it.  Changing it to 2.0 seemed to help but only 
for certain geometry sizes.

Original comment by paybyl...@gmail.com on 11 Dec 2011 at 7:10

[GoogleCodeExporter]

Is this going to be changed in the near future?  Or at least addressed?

Original comment by adrianh....@gmail.com on 15 Oct 2012 at 3:24

[GoogleCodeExporter]

What seemed to work for me is setting "worldscale" (Top menu: 
Plugins>SketchyPhysics>Physics Settings) to 0.2.

Maybe someone can elucidate ? 

Original comment by binksdja...@gmail.com on 15 Oct 2012 at 1:12