Subject:
|
Re: r/c servos
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 20 Aug 1999 16:05:31 GMT
|
Viewed:
|
763 times
|
| |
|
|
John Barnes <barnes@sensors.com> wrote:
> If the pwm is really being controlled by an 8 bit counter, that might be
> fine for the servos. The current rcx output has a cycle time of 8mS and
> uses 1mS increments. If those are the top 3 bits, then the resolution is
> 5 bits worth or 0.03mS. This would set the average servo, which needs a
> pulse from about 0.7 - 2.0 mS for 180 degress to one of about 45 steps or
> every 5 degrees. That could be useful, but note that a geared down lego
> motor and the position sensor can do a lot better than that!
To get rid of the speculation here, the PWM is controlled by a programmable
timer that the code in ROM configures to generate an interrupt every 1 ms.
Output for each of 8 power levels is stored in a small table of waveforms,
with the 1 ms timer causing the next bit in the waveform table to be used
to modulate the output.
The resolution of the hardware timers in the RCX is quite fine, and since
it is possible to reprogram both the timers and the interrupt handlers, it
would certainly be possible to write a routine to output a precise pulse
stream.
-Kekoa
|
|
Message is in Reply To:
 | | Re: r/c servos
|
| If the pwm is really being controlled by an 8 bit counter, that might be fine for the servos. The current rcx output has a cycle time of 8mS and uses 1mS increments. If those are the top 3 bits, then the resolution is 5 bits worth or 0.03mS. This (...) (25 years ago, 20-Aug-99, to lugnet.robotics)
|
20 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
