[python] Re: Steering, springs and damping (LONG)
- From: Ray Schümacher <mtb@xxxxxxx>
- To: python@xxxxxxxxxxxxx
- Date: Thu, 03 Mar 2005 07:47:15 -0800
At 09:06 PM 3/2/2005 -0600, you wrote:
>Thank you very much for your feedback on my previous post. I also want to
>extend
>a note of appreciation to Ray for his hard work on the stability codes.
Thank you, hopefully it leads to some truly useful design ideas
>... Such a scheme would require a downward-pointing
>triangular cam, fastened to the rear subframe with a spring, to press down upon
>an upward-pointing triangular cam on the front subframe.
It could also be done with a spring loaded roller on the front frame and a cam
shape on the BB shell, the opposite of a detent spring.
The authors which suggested the negative coefficient had the idea of a cammed
pivot, like on some bathroom stall doors. Their actual finding was that the
desired effect was proportional to weight, so the cam is ideal. Note that
downhilling increases weight on the python pivot.
>.... However, this is
>practically a textbook recipe for dynamical chaos--not something I'd like in a
>steering system. Better, then, to "passively" damp the motion by resisting
>inward movement, rather than by forcing outward movement.
Forcing outward movement is identical to less weight or self-centering, which
might be desirable in some regimes.
I was actually thinking of the dampers you buy for hatchbacks at the auto
store; they dampen one direction, and some push in the other as well.
>There is a directional, hydraulic bicycle-steering damper made by Tim Hopey,
>that has recieved positive reviews:
>http://store.hopey.org/
slick!
>Unfortunately, this restricts outward movement, and gives free rotation back to
>the center--exactly the opposite of what's needed here.
My reading of the PDFs lately has me thinking that bike stability has enough
degrees of freedom, and interaction, to be a chaos problem. I just read "Chaos
in the Solar System" by Ivars Peterson - very similar non-linear issues and
unpredictability. The author Hand uses 17 inputs and fourth-order quadratics,
and does not fully determine the problem. There could easily be regions in
certain designs where Hopey's thing might help. I think what makes bike
stability an interesting problem is that so much is not intuitive (besides
testing every day).
>As the negative damping constant
>is increased, the stable (or near-stable) region moves to higher speeds.
Remember, in my program, the pivot k value is a spring, not a damper. I have
not figured out yet where to insert a damping variable.
>It almost
>seems that two damping mechanisms would be needed--one for downhill riding,
>and another for everything else.
An on/off switch would be good.
One author, Ruina, suggested that some of the effects of the system (and those
mathematically related, apparently) are functions only of distance; therefore
higher d/t (speeds) cause shorter effects - high speed wobble. My Trek does it
>35km/h no-hands.
Humans can also only respond so fast (~5Hz?), so that anything higher is
uncontrolled and needs a damper (like Hopey). One bad resonance during that
time could ruin your day.
One of the interesting things I saw in the program was the effect of more mass
out front being able to make the python theoretically self-stable. It reduces
center-steer but mainly acts by the moment of inertia resisting high speed
rotation of the pivot.
Ray
============================================================
This is the Python Mailinglist at freelists.org
Listmaster: Juergen Mages jmages@xxxxxx
============================================================
Other related posts:
