Stepper Motors and Twist in Shafts
Posted by
Ronald Anderson
on 2000-11-12 07:37:50 UTC
FYI
The problem here is that there is an assumption of torque. A motor
when asked to index one step will only generate torque, if the force
is resisted! The problem here it that there was no way to measure the
load and the twist at the same time. Had this been done then a good
experiment would have produced a twist measurement in the shaft in
accordance with the load measured.
To often data is faulty due to the lack of a carefully controlled
experiment setup and instrumentation capability.
To illustrate, as an extreme, a 350 chevy engine will put out 350 #/ft
of torque (or at least a good one will) If I were to attach a shaft
out the back of the motor and wing the throttle full, the shaft will
spin and follow the engine rpm, yet not twist much as the only load is
the inertia load of the shaft mass. Now if the shaft is attached to a
load (the vehicle axle) and the vehicle is also pulling a trailer
(more mass to accelerate) and now the throttle is opened fully, the
shaft will twist an amount based on its ability to resist the force
being driven through it.
Thus the torque wrench test previously mentioned in one of these posts
is more representative of the twist in a shaft when a force is
introduced and the force is being resisted. This a concept which must
be carefully applied when trying to critique system forces, torsional
deflection and the associated response of servos.
Keep on experimenting and you will see what I am illustrating here.
Sincerely,
Ron A.
The problem here is that there is an assumption of torque. A motor
when asked to index one step will only generate torque, if the force
is resisted! The problem here it that there was no way to measure the
load and the twist at the same time. Had this been done then a good
experiment would have produced a twist measurement in the shaft in
accordance with the load measured.
To often data is faulty due to the lack of a carefully controlled
experiment setup and instrumentation capability.
To illustrate, as an extreme, a 350 chevy engine will put out 350 #/ft
of torque (or at least a good one will) If I were to attach a shaft
out the back of the motor and wing the throttle full, the shaft will
spin and follow the engine rpm, yet not twist much as the only load is
the inertia load of the shaft mass. Now if the shaft is attached to a
load (the vehicle axle) and the vehicle is also pulling a trailer
(more mass to accelerate) and now the throttle is opened fully, the
shaft will twist an amount based on its ability to resist the force
being driven through it.
Thus the torque wrench test previously mentioned in one of these posts
is more representative of the twist in a shaft when a force is
introduced and the force is being resisted. This a concept which must
be carefully applied when trying to critique system forces, torsional
deflection and the associated response of servos.
Keep on experimenting and you will see what I am illustrating here.
Sincerely,
Ron A.
> Message: 5of
> Date: Sun, 12 Nov 2000 00:38:07 -0600
> From: Jon Elson <jmelson@...>
> Subject: Re: Torsion Test Results
>
>
>
> Mariss Freimanis wrote:
>
> > Hi,
> >
> > Ok, as promised, here's the results of my tests.
> >
> > The short answer is there is no significant difference which end
> > the leadscrew the encoder is on.rotational
>
> > The leadscrew simply did not twist enough to make any difference.
> > Pulsed torque 275 oz-in, 3/8" leadscrew, 30" travel.
>
> Fascinating! I'm amazed that you didn't see much twist under
> this setup. Now, one question - did you have a load on that
> leadscrew? (The significant difference between this test and the
> earlier one you reported was the big, heavy wrench on the end
> of the bar, which dramatically enhances the effect. The tiny
> inertia of the encoder doesn't compare.)
>
> Jon