Question on reliability of position computed from EV3 encoders

[Jonas]

Hi,

I want to ask the users of EV3 system: How much can you rely on position determination made by EV3 unit that computes it from the quadrature encoder inside the EV3 servomotors? Particularly in a situation where the position of a driven mechanical system (e.g. a robot arm) changes quickly and in both directions? Is it precise and robust enough so that limit switches may be omitted? I saw some robots with EV3 and many were without these switches.

I ask because I am experimenting with EV3 motors combined with Arduino and even though my control programs (inspired by this project) seem to be fast and accurate enough, there are always small deviations of the position after several movement repetitions. After tens and hundreds of repetitions the deviations becomes critical.

[doug72]

Have you seen this post 

 

 

[Jonas]

Yes, I have seen it. But the video is rather short to see what happens after hundreds of cycles.

[dr_spock]

Depends on the precision and speed you need.  It seems reliable enough for my slot machine with guests playing with it at a casino night event.  The color dots on the wheels never got out of alignment or sync. My slots program is not particularly taxing the EV3 processor.  

If you want to play it safe and avoid potential liabilities for injury, then limit switches wouldn't hurt to have.  They would be useful for homing and zeroing the positions too.  I had to manually turn the slot wheels to three matching colours (position 0) before starting the program. 

 

[Lok24]

I don't think this is an issue, my experiences are very good.

I think the decviations are more mechanical.

As already mentioned sensors are very useful for homing.

But depending on the software you can use the motor itself as a sensor.

I think LabView doesnt't offer an option to  set the power (not speed!) of the motors, like PU does.

[Coder Shah]

You can use the unregulated motor block in the LabView version of the EV3 software to set power.