AVR-Chat

> 1.  Have complex servo motor controls with fast...
> 2.  Have motors to match, with low inertia rotors...
> 3.  Your drive profile will be complicated
> 4.  Absolute encoders are expensive
> 5.  Resolvers, instead of absolute encoders, are expensive...
> 6.  DC servo motors will also have a large number of windings...

Sounds bleak.  If I'd known it was that difficult I might not have 
tried, but one of my current projects uses a retrofitted 
azimuth/elevation mount that is driven by inexpensive Mitsumi and 
Johnson commutated DC motors, had no position feedback at all, and was 
originally commanded by a four-switch joystick.  It was impossible to 
accurately position.  Once I'd found a suitable way to encode each axis, 
I found it was not at all difficult to build a satisfactory az/el drive 
that is accurate to about 0.5 degree.  I needed none of what you advise, 
above.

An H-drive, operated in PWM antiphase at about 500Hz provides sufficient 
pulsed torque to turn these motors slowly, and trapezoidal dynamics 
(linear acceleration/deceleration ramps) to achieve high speed requires 
trivial code.

Still, DC motors and steppers are different, of course.  One cannot turn 
these DC motors 1.8 degree at the shaft, which a stepper can do 
intrinsically.  And if you intend to move a 7-ton forge, these won't 
suffice, of course.  But if the application calls for lightweight 
gear-reduced position and you have appropriate encoders, it is not 
difficult and works fine.


Tom