AVR-Chat

> ... using timer1 to generate our pwm signal and [] one or the other 
of the OC1 pins to change direction...

There are two common PWM motor drive techniques, usually called 
Sign-Magnitude and Antiphase (or Push-Pull).  In my experience, these 
two methods of PWM motor speed control are very different and yield very 
different results.

The "Sign-Magnitude" method (i.e. direction/braking selection to the 
H-bridge phase control inputs, PWM speed control to the H-bridge enable 
  input) produces an exponential PWM/speed curve. Perhaps 60% speed 
change comes from only 15% duty cycle input change. Lower PWM 
frequencies produce better torque at the expense of noise chatter and 
pulsing motion, and very slow motion is possible due to armature 
hammering by pulses as narrow as 3% before stall. Except for those 
effects, low PWM
frequencies require no particular concern.

The "Locked Anti-Phase" method (i.e. PWM applied in push-pull (i.e. one 
side is inverted) to the H-bridge phase control inputs, on/off to the 
enable input) produces an almost-straight-line bipolar linear PWM/speed 
curve, with zero-speed at  50% PWM duty cycle. My device yields 30 
degrees/second at 100% PWM and 15 degrees/second at 75% PWM, -30 
degrees/second at 0%, and -15 degrees/second at 25% PWM, pragmatically 
perfect PWM/speed tracking. My motors stalled at about 18% PWM, drawing 
150mA average at 449Hz. Higher PWM frequencies provide higher stall 
speeds at less current; lower PWM frequencies produce lower stall 
speeds, but at greatly-increased current demand and motor heating. In my 
testing, for instance, 113Hz PWM overheated a motor at 36% PWM, yielding 
8 degrees/second motion. On the other hand, a 904Hz PWM frequency 
produced almost the same control as 449Hz, drawing 80mA instead of 
150mA, and stalling a little earlier.

Dynamic braking is possible with Sign-Magnitude, as is coasting. 
Anti-Phase, though, is essentially dynamically braked continually and 
offers no coasting except by disabling the H-bridge.


Tom