by ArtF » Mon Mar 06, 2023 3:56 pm
Bill:
I think the most important thing is the 90 degree angle of lock surfaces, even a small amount of angle can start a slide of two
surfaces that should be locked due to that 90 degree angle. When I looked at the newtons cradle for example, I found
part of the difficulty is that the newton in reality is adjusted so that when the initial ball hits the stack of 3 or 4 in their
row, the point of contact must be when its string is exactly at 90 degrees and the balls all touch on a spherical surface
at 90 degrees. The force transfers from ball to ball at that 90 degree vector losing basically nothing.. BUT , even a small
angular offset makes the force turn to a vector that causes chaotic motion of the following balls making the forces drop
off quickly. (I know you know all this ), but to be considered is that the drawing of a circle is really small triangles making
a too small to see aliased surface that can, if not REAL accurate to point of contact make that vector swing in a tangential
relationship to the desired force. This tan function can strip power fast or make things slide.
The solution I think is to ensure 90's are 90's even if that takes changing from a spherical contact point to one more
square depending on escapement type. I'll be playing about with that soon to see how to make such situations better,
I'm just fixing some oddball items I dont like in the extrusions before I try out some escapements to see how
they react. B2.0 should be out soon with any luck.
Art
Bill:
I think the most important thing is the 90 degree angle of lock surfaces, even a small amount of angle can start a slide of two
surfaces that should be locked due to that 90 degree angle. When I looked at the newtons cradle for example, I found
part of the difficulty is that the newton in reality is adjusted so that when the initial ball hits the stack of 3 or 4 in their
row, the point of contact must be when its string is exactly at 90 degrees and the balls all touch on a spherical surface
at 90 degrees. The force transfers from ball to ball at that 90 degree vector losing basically nothing.. BUT , even a small
angular offset makes the force turn to a vector that causes chaotic motion of the following balls making the forces drop
off quickly. (I know you know all this ), but to be considered is that the drawing of a circle is really small triangles making
a too small to see aliased surface that can, if not REAL accurate to point of contact make that vector swing in a tangential
relationship to the desired force. This tan function can strip power fast or make things slide.
The solution I think is to ensure 90's are 90's even if that takes changing from a spherical contact point to one more
square depending on escapement type. I'll be playing about with that soon to see how to make such situations better,
I'm just fixing some oddball items I dont like in the extrusions before I try out some escapements to see how
they react. B2.0 should be out soon with any luck.
Art