Mike Pauza
08-16-2003, 04:02 PM
Hey everybody! I think I just discovered that "MD sometimes massages it's parameters". That may be part of the reason MD seems so unpredictable. Maybe some of you can help verify or refute this.
A spring is supposed to deform linearly right? If you up the mass by 10X, its supposed to stretch 10X as much, and if you up the spring constant (the stiffness) by 10X it's supposed to stretch 1/10 as much...right? Well guess what...that's exactly what happens...sometimes. I ran about 50 simulations and found that for my setup (see below) the sim behaved as expected as long as:
STIFFNESS > 10000 * sqrt(Weight) .....approximately
For example...for an object Weight of 1, MD would start "massaging" my stiffness settings (not what I would expect) unless I had it set over about 10,000. If MD deemed my stiffness setting was "too low", it would beef it up a bit internally. Sometimes this internal fudging would yield the same results even when I changed the stiffness by a factor of 1000! That's freaky isn't it???
I ran a number of simulations with a single two point polygon. The poly was modeled in the y direction with a length of 1 meter and given the spring attributes, the top point (really a 1 point poly) was set to be fixed, and the bottom point (another 1 point poly) was given the mass attributes. Basically I made a scale. :) I ran a few more sims with different scenes and got the same sort or results.
Has anyone else had similiar experiences? If MD really does internally mess around with our parameters it would explain a lot.
-Mike Pauza
A spring is supposed to deform linearly right? If you up the mass by 10X, its supposed to stretch 10X as much, and if you up the spring constant (the stiffness) by 10X it's supposed to stretch 1/10 as much...right? Well guess what...that's exactly what happens...sometimes. I ran about 50 simulations and found that for my setup (see below) the sim behaved as expected as long as:
STIFFNESS > 10000 * sqrt(Weight) .....approximately
For example...for an object Weight of 1, MD would start "massaging" my stiffness settings (not what I would expect) unless I had it set over about 10,000. If MD deemed my stiffness setting was "too low", it would beef it up a bit internally. Sometimes this internal fudging would yield the same results even when I changed the stiffness by a factor of 1000! That's freaky isn't it???
I ran a number of simulations with a single two point polygon. The poly was modeled in the y direction with a length of 1 meter and given the spring attributes, the top point (really a 1 point poly) was set to be fixed, and the bottom point (another 1 point poly) was given the mass attributes. Basically I made a scale. :) I ran a few more sims with different scenes and got the same sort or results.
Has anyone else had similiar experiences? If MD really does internally mess around with our parameters it would explain a lot.
-Mike Pauza