iCubGenova09 - added incremental AMO calibration on the shoulders #14
Conversation
|
Only a quick comment. |
|
Ok, sorry for the noise then! |
S-Dafarra
left a comment
There was a problem hiding this comment.
Thanks @davidetome. I have noticed that you probably lost the joint (7) from the calibration order. In addition, the limits you set for the shoulder pitch seem a bit weird. I remember we checked them on the robot, and the shoulder pitch should have +25/-180 range of motion. If you read a different limit it is possible that the encoder was not calibrated
| <!-- 0 1 --> | ||
| <param name="hardwareJntPosMax"> 25 165 </param> | ||
| <param name="hardwareJntPosMin"> -180 0 </param> | ||
| <param name="hardwareJntPosMax"> 36.2 165 </param> |
There was a problem hiding this comment.
Is it this wide? From https://github.com/icub-tech-iit/tickets/issues/1410#issuecomment-953972983 it should be +25 - 180
There was a problem hiding this comment.
36.2 is the measured angle of the hard stop; this is necessary to let the incremental procedure work
cc @ale-git
There was a problem hiding this comment.
Measured how? Is it possible that the encoder was miscalibrated? The mechanical range of motion should be 205deg
There was a problem hiding this comment.
moving by hand to the hard stop and reading the motorgui.
@ale-git I'm right?
| <!-- 0 1 --> | ||
| <param name="hardwareJntPosMax"> 25 165 </param> | ||
| <param name="hardwareJntPosMin"> -180 0 </param> | ||
| <param name="hardwareJntPosMax"> 36.2 165 </param> |
There was a problem hiding this comment.
As above, this limit is a bit weird
There was a problem hiding this comment.
36.2 is the measured angle of the hard stop; this is necessary to let the incremental procedure work
cc @ale-git
There was a problem hiding this comment.
As @davidetome said, we measured the back hard stop position after absolute calibration, in order to make the encoder reading the same when switched to hard stop. We did it more than once, so it is not likely (though still possible) that we ran into a miscalibration. Anyway, a difference of 16 degrees must be evident after hard stop calibration... how does the shoulder seem when you put it in 0 position? We checked the 0 and 90 degrees position and it seemed fine. Maybe the limit was reduced for control reasons? In this case only the sotware limit should have been changed, but maybe this wasn't clear at that time. Another possibility is that something wasn't properly mounted, as happened once in the shoulder yaw. I suggest 1) to visually check the 0 and 90 degrees positions and 2) to check the CAD limits.
There was a problem hiding this comment.
@S-Dafarra , I thought we should check again the limits (both) to see if one increases the other one decreases.. what about @ale-git?
There was a problem hiding this comment.
Yes you can check if the other limit is ok. Good idea.
There was a problem hiding this comment.
Unfortunately the documentation doesn't help
Anyway, if we want the back limit at 25 deg all we need to do is to put 25 instead of 36.2 in the calibration file. I strongly suggest to make the quick test of moving the joint by hand to hard stops and see 1) if the upper limit is still -180 and 2) if the arm really reaches the vertical position, since I doubt about this last point.
There was a problem hiding this comment.
The shoulder pitch has been recently changed by Elio Massa: see https://github.com/icub-tech-iit/tickets/issues/1410#issuecomment-953972983.
Maybe, @ale-git @davidetome you were not aware of this.
And the documentation is sadly not up-to-date, yep.
There was a problem hiding this comment.
Anyway, if we want the back limit at 25 deg all we need to do is to put 25 instead of 36.2 in the calibration file. I strongly suggest to make the quick test of moving the joint by hand to hard stops and see 1) if the upper limit is still -180 and 2) if the arm really reaches the vertical position, since I doubt about this last point.
I did exactly this 😉
I guess also the corresponding parameter in the calibration file (
) should change right?There was a problem hiding this comment.
Yes, the 36.2 value must be changed to your actual desired limit.
There was a problem hiding this comment.
@pattacini I didn't know about the limit upgrade! It seems 170 degrees more than 180 with respect to the vertical axis btw, but the 0 is arbitrary of course :)
added joint 7
added joint 7
|
Today w/ @ale-git we fixed the shoulders pitch limits. |
|
Thanks a lot @davidetome! Merging |
As per the title, this PR contains the mods made w/ @ale-git for the AMO incremental calibration on the shoulders.
cc @S-Dafarra @pattacini