Re: Stepper Acceleration in RWTS

[vladitx <vladitx@nucleusys.com>]

On Dec 29, 5:51 am, Sheldon Simms <wsspa...@gmail.com> wrote:

> Ok that's helpful, but I don't see any of this specified on any motor datasheets.

It's not the motor only, but the whole mechanical construction. Floppy
drive specs have track-to-track min seek times.

> Would one just determine this empirically?

That works, too. :) As mentioned, newer drives got faster at seek.

> Assuming that a later motors have higher enough torque/faster enough response,
> then wouldn't the motor turn to an energized phase and bounce around it for
> a bit until it settles, only to have to start from zero angular velocity again
> when the next phase is energized?

After the oscillations upon decelerating reduce, there are further
oscillations while "keeping" the position. That's why it's not
recommended to read/write with coils still on.

> If that's the case, then the whole idea of
> "accelerating" the motor is kind of silly even though the code may successfully
> seek to the correct track.

Probably. You can't control different torque motors and different
mechanic loads equally effective with the same fixed timing.
One of the reasons the floppy drives got more intelligent.

Since there is no current control and plenty of time, the torque is
high enough to do the job.

> Right. I wasn't suggesting putting tracks on a "virtual pole" (although I'm
> pretty sure that was actually done for various reasons)

Quarter (0.25) tracks.

> but just wondering
> about the one-two-one-two pattern for seeking in the light of the fact that
> most datasheets I've looked at assume a two-two-two-two pattern  for
> stepping.

Going through smaller steps, like the A-AB-B sequence, results in more
smooth travel.
Probably less mechanical stress, too.