Servopod Controller
Servopod Controller
by subpilot
Mon Jul 17, 2006 3:25 am
I've been experimenting with my Servopod controller and it is showing great promise as a replacement for the RN controller. I am successfully driving the servos and getting position feedback back from them. I have also been able to talk to several different I2C based sensors like the SRF10.
Here's a link to the Servopod forum and a discussion about RN servo feedback.
http://www.newmicros.com/discussion/sho ... 7#post6017
The customer support from Newmicros is first class.
Here's a link to the Servopod forum and a discussion about RN servo feedback.
http://www.newmicros.com/discussion/sho ... 7#post6017
The customer support from Newmicros is first class.
by limor
Mon Jul 17, 2006 11:06 am
Very interesting approach with the diodes in parallel.
I assume the Pod would have to send the short-pulses to the servos (to trigger the position feedback) one servo after the other, waiting for each servo to respond before talking to the next servo.
This sounds very slow (few seconds for 17 servos..).
Hence this approach would probably not work if I wanted to continuously sample robot joints in order to activate some feedback control scheme.
Does your Pod support 17 PWM outputs or only 12?
Limor
I assume the Pod would have to send the short-pulses to the servos (to trigger the position feedback) one servo after the other, waiting for each servo to respond before talking to the next servo.
This sounds very slow (few seconds for 17 servos..).
Hence this approach would probably not work if I wanted to continuously sample robot joints in order to activate some feedback control scheme.
Does your Pod support 17 PWM outputs or only 12?
Limor
by subpilot
Mon Jul 17, 2006 11:48 am
If you want to get servo positions the maximum cycle time is going to be ~3mSecs / servo. Don't know why you would have to get positions from all servos all the time. Just get position data from the servos you have to service.
For that matter, I really don't see a need to get positions unless you are capturing a move like you would in Roboscript. You're commanding the servo to a known position at all other times. If you grab a position in mid move it's going to have some error since the servo drive is stopped and restarted during a position feedback command. I also think it will make the robot a bit jerky. If you really want dynamic feedback maybe it would make more sense to get the analog voltage out of the servo and feed it to a controller A2D pin.
The Servopod has 12 dedicated PWM output pins but also supports PWM I/O on the 15 Timer pins so there's no shortage of PWM channels. The PWM input functions make using a SRF4 ranger etc with PWM outputs a breeze.
For that matter, I really don't see a need to get positions unless you are capturing a move like you would in Roboscript. You're commanding the servo to a known position at all other times. If you grab a position in mid move it's going to have some error since the servo drive is stopped and restarted during a position feedback command. I also think it will make the robot a bit jerky. If you really want dynamic feedback maybe it would make more sense to get the analog voltage out of the servo and feed it to a controller A2D pin.
The Servopod has 12 dedicated PWM output pins but also supports PWM I/O on the 15 Timer pins so there's no shortage of PWM channels. The PWM input functions make using a SRF4 ranger etc with PWM outputs a breeze.
by subpilot
Mon Jul 17, 2006 12:02 pm
The Wibotix board has the exact same footprint as the Servopod so you can mount them back to back on a set of standoffs. That makes the package too big to fit under the standard RN back cover plate but it's not excessively huge. The Wibotix takes 300ma of 3.3V power so it should really be supplied by a DC/DC converter to save battery power. Both boards just have 3 terminal linear regs and I'm planning on adding a converter soon.
I haven't done anything with the Wibotix board other than establish a serial link and verify that I can loop thru the wireless connection via two terminal windows. Setup was pretty painless and it requires no screwing with for it to establish a connection once you power it up.
I haven't done anything with the Wibotix board other than establish a serial link and verify that I can loop thru the wireless connection via two terminal windows. Setup was pretty painless and it requires no screwing with for it to establish a connection once you power it up.