Thread: PWM
View Single Post
  #14   Spotlight this post!  
Unread 27-12-2002, 20:40
Lloyd Burns Lloyd Burns is offline
Registered User
FRC #1246 (Agincourt Robotics)
Team Role: Engineer
 
Join Date: Jun 2001
Rookie Year: 1997
Location: Toronto
Posts: 292
Lloyd Burns is an unknown quantity at this point
The PWM signal that is sent to the Victors and the servos is never seen by the motor.

In the Victor, the length of time of the signal pulse is counted out (most llikely) in the on-board CPU, which also directs the MOSFET bridge to turn on and off providing a varying-average dc, while supplying the motor with full (12V) voltage bursts of current which vary in their on-time. The motor cannot change speed as fast as the FETs turn on and off (200 kHz ?), but the current provides a burst of torque, which maintains the rotation, and the motor's turning-generated EMF keeps the voltage between M+ and M- at the average.

If the Victor fails to see a pulse in a reasonable amount of time, it will change to output corresponding to neutral.

The servos read the input pulse and probably count it out, then decide what the motor should do to make the pot inside the servo out put the analog of the pulse-length. They probably turn on a transistor bridge full time to make the motor turn for a large movement, some might even start pulsing as they got near, to prevent overshoot. When pot and signal agree, the motor is stopped.

On full value pulses controlling average values :
The servos are the descendants of escapements powered by rubber bands : the escapements were modofied relays, which were pulled by a tap on the control on the operator interface. This allowed the rubber band to turn a three- or four-eared wheel 120 or 90 degrees. With four ears, to make a plane or boat turn a certain number of degrees, you turned the rudder full in the required direction for a time, then flipped through to neutral (3 pulses), then if the turn were not enough, you'd repeat the previous. Proportional control needed right turn - left turn - right turn - left with more time on the way you wanted to turn.
With a three eared wheel, one wheel (0 degrees) was neutral control position, one (120 degrees) was right, and another (240 degrees) was left.

This is sort of the way the Victor output works. The vehicle with the escapement steering would not turn significantly with a pulse, but with several cycles, going more often to one side, a biased average, a turn could be accomplished.