I’m having very inconsistent result because of this even though we have quad encoders.
If you have the time, slowing it down will almost ALWAYS make it more accurate. Also make sure you don’t have excess friction in the drive.
Edit: you could also use a gyroscope. If you do, try and put it in the center of the robot.
We use the gyroscope with PID, it is about the most precise you can get for autonomous turns.
1 Like
@DanielIW26 Do you guys mount it straight to metal or do you do some kind of vibration reduction?
We have it mounted straight to the metal. I am not sure wether you could get a good vibration reduction piece.
Rubber links 
@DanielIW26 I was contemplating some form of rubber band housing? However I probably won’t bother with it since you guys are fine with mounting it straight to metal.
It works just fine this year to mount it to metal. Last year with flywheels it would have been better.
Rubber links seem that they would cause more problems then they would solve.
Can confirm, rubber links are op op.
How necessary are rubber links? @Nehalem because I see you have 3x Programmings skills world championships…which means you must know what you’re talking about when it comes to sensors.
From my experience, the rubber links help to reduce a lot of the drift that gyros tend to have. If they’re screwed directly into the metal, vibrations from the robot’s movement will most likely generate noise that will affect the accuracy of the sensor. In that case, it’ll be necessary for you to program a way to deal with the extra drift/noise.
For the record, @Henry and I were on the same team when we won the programming skills titles.
Did you have a problem with the gyro not being completely parallel to the ground while it was screwed to rubber links?
My team uses a combination of the gyro and quadrature encoders mounted on unpowered wheels for our autonomous functions. The gyro works well over the short team and quads works well over the long term, so the two can combine to form a more accurate estimate.
Edit: I should also say we use two PID controllers in tandem for driving straight and turning. One controller controls the overall movement, and the other controls deviance from the optimal path to the target (i.e., one controller drives us to a point and another keeps us straight).
The gyro wasn’t completely parallel to the ground but only negligibly. It still returned an accurate reading.
Not really, if it is a major concern, just make sure you screw it in so that the rubber links aren’t twisted.