this is my first attempt at using a FRC gyro sensor with vex. my goal was to get it to stay in a straight position, even if the ground beneath it rotated. the bumper switch on the back resets the gyro because it slowly increases in value (im not sure why) and the switch allows me to reset that value.
pictures:
video:
http://s37.photobucket.com/albums/e82/pacoliketaco/?action=view¤t=DSCN3692.flv
sweet…a tiny tiny bit off but who cares!?
Very cool … you’ve got a single axis gyro stabilized platform. Not quite ready for a camera yet:) A higher feedback gain could help the startup offset when changing direction. What gain are you using …pwm/rate?
The gyro drift is typical…I’ve had a chance to look at the FRC specs and you are up against at least the vex a/d 5mv noise (.4 deg/s) and the gyro noise density of .05 deg/s/sqrt(Hz). Assuming 40 Hz bandwidth this gives .32 deg/s noise. So you are looking at an rss value of around .5deg/s random noise plus the null drift. You can help this by doing multiple a/d reads per update and using the average value. Also, if you haven’t already done a null bias cancellation at initialization , this could cause a big drift if your gyro in not at the nominal 2.5 v.
The gyro is also g sensitive (.2 deg/s/g) so if your gyro is not perfectly centered about the rotational axis, there is a centripetal acceleration =off_center_dist*(rate)**2/g . It looks like your gyro is centered well and your rates are low so this should be negligible.
I would like to see how well you can follow a straight line over a 20 ft distance with the gyro guiding you. If you have a forward path integration on the error signal then any residual null drift will cause you to veer off, but it is a good test. The integration is necessary to offset the motor static friction offsets. You would of course have to eliminate the initial null offset with several seconds of averaging before the run.
the intent of this robot was to show to my team’s programmers that easyc pro is not a joke, and it is actually very nice. but also i was learning how to program and work with a gyro, and recently i also set up a full pneumatics cirvuit with electronics and everything, since we are hoping to use them on this years’ robot.
This is really cool, i am going to try and get some sensors like this. I wish i had an accelerometer to measure the exact speed on my g-force accelerator robot.
accelerometers are only around ~$20 dollars, try robotshop.ca or hvwtech.com
dimension engineering make great ones but i forget their website 