|
Dear hackers,
we have the following problem with our iCub: when the eyes are at their home position, they look about 10 degrees too far up. We have the feeling that the band that controls the tilt movement has shifted. What would be the solution for this, adjust the band, or somehow adjust the 0 position in icub_head_torso_safe.ini? --Leo |
|
Hi Leo,
you can follow instructions for fine calibration in the manual, direct link to instructions for the head: http://eris.liralab.it/wiki/HeadFineCalibration Lorenzo > -----Original Message----- > From: leo_pape [mailto:[hidden email]] > Sent: lunedì 18 giugno 2012 15:45 > To: [hidden email] > Subject: [rc-hackers] wrong icub eyes tilt > > Dear hackers, > > we have the following problem with our iCub: when the eyes are at their > home position, they look about 10 degrees too far up. We have the feeling > that the band that controls the tilt movement has shifted. What would be the > solution for this, adjust the band, or somehow adjust the 0 position in > icub_head_torso_safe.ini? > > --Leo > > -- > View this message in context: http://robotcub- > hackers.2198711.n2.nabble.com/wrong-icub-eyes-tilt-tp7577397.html > Sent from the RobotCub Hackers mailing list archive at Nabble.com. > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and threat > landscape has changed and how IT managers can respond. Discussions will > include endpoint security, mobile security and the latest in malware threats. > http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Robotcub-hackers mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
Ciao Lorenzo,
we will try this. Is this a common phenomena that it changes by this much over just a weekend? I am just wondering if there is maybe a mechanical fault and if we can make sure that there is not :) //Juxi On Wed, Jun 20, 2012 at 5:02 PM, Lorenzo Natale <[hidden email]> wrote: > Hi Leo, > you can follow instructions for fine calibration in the manual, direct link to instructions for the head: > > http://eris.liralab.it/wiki/HeadFineCalibration > > Lorenzo > > >> -----Original Message----- >> From: leo_pape [mailto:[hidden email]] >> Sent: lunedì 18 giugno 2012 15:45 >> To: [hidden email] >> Subject: [rc-hackers] wrong icub eyes tilt >> >> Dear hackers, >> >> we have the following problem with our iCub: when the eyes are at their >> home position, they look about 10 degrees too far up. We have the feeling >> that the band that controls the tilt movement has shifted. What would be the >> solution for this, adjust the band, or somehow adjust the 0 position in >> icub_head_torso_safe.ini? >> >> --Leo >> >> -- >> View this message in context: http://robotcub- >> hackers.2198711.n2.nabble.com/wrong-icub-eyes-tilt-tp7577397.html >> Sent from the RobotCub Hackers mailing list archive at Nabble.com. >> >> ------------------------------------------------------------------------------ >> Live Security Virtual Conference >> Exclusive live event will cover all the ways today's security and threat >> landscape has changed and how IT managers can respond. Discussions will >> include endpoint security, mobile security and the latest in malware threats. >> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >> _______________________________________________ >> Robotcub-hackers mailing list >> [hidden email] >> https://lists.sourceforge.net/lists/listinfo/robotcub-hackers > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Robotcub-hackers mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers -- "We need dreamers 'cuz in our dreams we see not what is but what can be!" http://about.me/Juxi ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
I would say it should not happen... but it could just be that the belt slipped for some reasons during operation. Maybe Stefano can confirm or suggest things to check (when he is back from vacation, next week).
Cheers, Lorenzo > -----Original Message----- > From: Juxi Leitner [mailto:[hidden email]] > Sent: mercoledì 20 giugno 2012 17:20 > To: Lorenzo Natale > Cc: leo_pape; [hidden email] > Subject: Re: [rc-hackers] wrong icub eyes tilt > > Ciao Lorenzo, > we will try this. Is this a common phenomena that it changes by this much > over just a weekend? > I am just wondering if there is maybe a mechanical fault and if we can make > sure that there is not :) > > //Juxi > > On Wed, Jun 20, 2012 at 5:02 PM, Lorenzo Natale <[hidden email]> > wrote: > > Hi Leo, > > you can follow instructions for fine calibration in the manual, direct link to > instructions for the head: > > > > http://eris.liralab.it/wiki/HeadFineCalibration > > > > Lorenzo > > > > > >> -----Original Message----- > >> From: leo_pape [mailto:[hidden email]] > >> Sent: lunedì 18 giugno 2012 15:45 > >> To: [hidden email] > >> Subject: [rc-hackers] wrong icub eyes tilt > >> > >> Dear hackers, > >> > >> we have the following problem with our iCub: when the eyes are at > >> their home position, they look about 10 degrees too far up. We have > >> the feeling that the band that controls the tilt movement has > >> shifted. What would be the solution for this, adjust the band, or > >> somehow adjust the 0 position in icub_head_torso_safe.ini? > >> > >> --Leo > >> > >> -- > >> View this message in context: http://robotcub- > >> hackers.2198711.n2.nabble.com/wrong-icub-eyes-tilt-tp7577397.html > >> Sent from the RobotCub Hackers mailing list archive at Nabble.com. > >> > >> --------------------------------------------------------------------- > >> --------- > >> Live Security Virtual Conference > >> Exclusive live event will cover all the ways today's security and > >> threat landscape has changed and how IT managers can respond. > >> Discussions will include endpoint security, mobile security and the latest in > malware threats. > >> http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > >> _______________________________________________ > >> Robotcub-hackers mailing list > >> [hidden email] > >> https://lists.sourceforge.net/lists/listinfo/robotcub-hackers > > > > ---------------------------------------------------------------------- > > -------- > > Live Security Virtual Conference > > Exclusive live event will cover all the ways today's security and > > threat landscape has changed and how IT managers can respond. > > Discussions will include endpoint security, mobile security and the > > latest in malware threats. > > http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > > _______________________________________________ > > Robotcub-hackers mailing list > > [hidden email] > > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers > > > > -- > "We need dreamers 'cuz in our dreams we see not what is but what can be!" > > http://about.me/Juxi ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
In reply to this post by Juxi Leitner
Hi Hackers,
as Juxi mentionned it, it could be a mechanical problem related to transmission belt (see below) which slipped on the wheel attached to the output of the motor actuating the tilt of the eyes. ![]() To check if it is true, unmount the face cover, check if the belt is loose or consummed by ages: these are the only possibilities which could lead to slippering belt. The simplest one to solve is the loosing belt, for that just straighten the screw on the opposite side of the nose (see below picture) ![]() until having a good and tensed belt. Afterwards just make a little head fine calibration http://eris.liralab.it/wiki/HeadFineCalibration the other one is less easy to treat because means that it needs to be changed and to make sure is used and replacing the part has to be done just run the tilt movement and verify the slippery of the belt during the movement. If it is the case just ask to your icub support @ icub.support@icub.iit.it for a replacement part |
|
Thank you I will try this next week. We were already thinking that the slack on the belt could be a bit too much.
I hope this might fix that. I'll let you know. //Juxi Sent from my iPhone On 22-giu-2012, at 16:23, julijenv <[hidden email]> wrote: > Hi Hackers, > as Juxi mentionned it, it could be a mechanical problem related to > transmission belt (see below) which slipped on the wheel attached to the > output of the motor actuating the tilt of the eyes. > > http://robotcub-hackers.2198711.n2.nabble.com/file/n7577408/belt.jpg > > To check if it is true, unmount the face cover, check if the belt is loose > or consummed by ages: these are the only possibilities which could lead to > slippering belt. > The simplest one to solve is the loosing belt, for that just straighten the > screw on the opposite side of the nose (see below picture) > > http://robotcub-hackers.2198711.n2.nabble.com/file/n7577408/screw_tenser.jpg > > until having a good and tensed belt. Afterwards just make a little head fine > calibration http://eris.liralab.it/wiki/HeadFineCalibration > http://eris.liralab.it/wiki/HeadFineCalibration > the other one is less easy to treat because means that it needs to be > changed and to make sure is used and replacing the part has to be done just > run the tilt movement and verify the slippery of the belt during the > movement. If it is the case just ask to your icub support @ > [hidden email] for a replacement part > > -- > View this message in context: http://robotcub-hackers.2198711.n2.nabble.com/wrong-icub-eyes-tilt-tp7577397p7577408.html > Sent from the RobotCub Hackers mailing list archive at Nabble.com. > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Robotcub-hackers mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
In reply to this post by Lorenzo Natale-2
Thanks, Lorenzo.
I followed the instructions in the fine head calibration manual, using a -14 degree change to axis 4 in Zeros. When the iCub starts up, the eyes are now at the right position. After this change, however, setting the eyes to the maximum tilt value in robotMotorGui doesn't result in the eyes looking all the way up, as it could do before. Similarly, the eyes go way further down than is supposed to be possible when setting the tilt to the minimum value in robotMotorGui. My guess is that the Zeros entry is not the problem. Is there something else we can calibrate to fix this? I couldn't find any calibration instructions for the head other than the fine calibration manual. Also, I checked the belt. It seems quite tight, and doesn't slip during iCub startup or subsequent operation. --Leo -----Original Message----- From: Lorenzo Natale Sent: Wednesday, 20 June, 2012 17:02 To: leo_pape ; [hidden email] Subject: RE: [rc-hackers] wrong icub eyes tilt Hi Leo, you can follow instructions for fine calibration in the manual, direct link to instructions for the head: http://eris.liralab.it/wiki/HeadFineCalibration Lorenzo > -----Original Message----- > From: leo_pape [mailto:[hidden email]] > Sent: lunedì 18 giugno 2012 15:45 > To: [hidden email] > Subject: [rc-hackers] wrong icub eyes tilt > > Dear hackers, > > we have the following problem with our iCub: when the eyes are at their > home position, they look about 10 degrees too far up. We have the feeling > that the band that controls the tilt movement has shifted. What would be > the > solution for this, adjust the band, or somehow adjust the 0 position in > icub_head_torso_safe.ini? > > --Leo > ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
Hi Leo,
In the hypothesis that the problem is "software" (i.e. the head configuration file changed for some reason) you could try to fix the behavior you describe by changing the limits for the eye tilt joint (reduce the lower limit, increase the upper one). It may be worth trying... Lorenzo > -----Original Message----- > From: Leo Pape [mailto:[hidden email]] > Sent: mercoledì 27 giugno 2012 17:23 > To: [hidden email] > Subject: Re: [rc-hackers] wrong icub eyes tilt > > Thanks, Lorenzo. > I followed the instructions in the fine head calibration manual, using a -14 > degree change to axis 4 in Zeros. When the iCub starts up, the eyes are now > at the right position. After this change, however, setting the eyes to the > maximum tilt value in robotMotorGui doesn't result in the eyes looking all > the way up, as it could do before. Similarly, the eyes go way further down > than is supposed to be possible when setting the tilt to the minimum value in > robotMotorGui. > My guess is that the Zeros entry is not the problem. Is there something else > we can calibrate to fix this? I couldn't find any calibration instructions for the > head other than the fine calibration manual. > Also, I checked the belt. It seems quite tight, and doesn't slip during iCub > startup or subsequent operation. > > --Leo > > > > > -----Original Message----- > From: Lorenzo Natale > Sent: Wednesday, 20 June, 2012 17:02 > To: leo_pape ; [hidden email] > Subject: RE: [rc-hackers] wrong icub eyes tilt > > Hi Leo, > you can follow instructions for fine calibration in the manual, direct link to > instructions for the head: > > http://eris.liralab.it/wiki/HeadFineCalibration > > Lorenzo > > > > -----Original Message----- > > From: leo_pape [mailto:[hidden email]] > > Sent: lunedì 18 giugno 2012 15:45 > > To: [hidden email] > > Subject: [rc-hackers] wrong icub eyes tilt > > > > Dear hackers, > > > > we have the following problem with our iCub: when the eyes are at > > their home position, they look about 10 degrees too far up. We have > > the feeling that the band that controls the tilt movement has shifted. > > What would be the solution for this, adjust the band, or somehow > > adjust the 0 position in icub_head_torso_safe.ini? > > > > --Leo > > > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and threat > landscape has changed and how IT managers can respond. Discussions will > include endpoint security, mobile security and the latest in malware threats. > http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Robotcub-hackers mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
Hi Lorenzo,
I haven't tried changing the limits yet, but it doesn't seem like a permanent fix. The encoder values before and after the eyes started acting weird would be different for the same eye position. Much of our software was trained/calibrated using the old settings, I foresee further problems with iKinGazeCtrl, etc. Isn't there something else we can try? The last time we changed the head configuration files was several months ago (from last modified date), so I am pretty sure the cause is not a sudden change in configuration file. --Leo -----Original Message----- From: Lorenzo Natale Sent: Wednesday, June 27, 2012 18:00 To: Leo Pape ; [hidden email] Subject: RE: [rc-hackers] wrong icub eyes tilt Hi Leo, In the hypothesis that the problem is "software" (i.e. the head configuration file changed for some reason) you could try to fix the behavior you describe by changing the limits for the eye tilt joint (reduce the lower limit, increase the upper one). It may be worth trying... Lorenzo > -----Original Message----- > From: Leo Pape [mailto:[hidden email]] > Sent: mercoledì 27 giugno 2012 17:23 > To: [hidden email] > Subject: Re: [rc-hackers] wrong icub eyes tilt > > Thanks, Lorenzo. > I followed the instructions in the fine head calibration manual, using > a -14 > degree change to axis 4 in Zeros. When the iCub starts up, the eyes are > now > at the right position. After this change, however, setting the eyes to the > maximum tilt value in robotMotorGui doesn't result in the eyes looking all > the way up, as it could do before. Similarly, the eyes go way further down > than is supposed to be possible when setting the tilt to the minimum value > in > robotMotorGui. > My guess is that the Zeros entry is not the problem. Is there something > else > we can calibrate to fix this? I couldn't find any calibration instructions > for the > head other than the fine calibration manual. > Also, I checked the belt. It seems quite tight, and doesn't slip during > iCub > startup or subsequent operation. > > --Leo > > > > > -----Original Message----- > From: Lorenzo Natale > Sent: Wednesday, 20 June, 2012 17:02 > To: leo_pape ; [hidden email] > Subject: RE: [rc-hackers] wrong icub eyes tilt > > Hi Leo, > you can follow instructions for fine calibration in the manual, direct > link to > instructions for the head: > > http://eris.liralab.it/wiki/HeadFineCalibration > > Lorenzo > > > > -----Original Message----- > > From: leo_pape [mailto:[hidden email]] > > Sent: lunedì 18 giugno 2012 15:45 > > To: [hidden email] > > Subject: [rc-hackers] wrong icub eyes tilt > > > > Dear hackers, > > > > we have the following problem with our iCub: when the eyes are at > > their home position, they look about 10 degrees too far up. We have > > the feeling that the band that controls the tilt movement has shifted. > > What would be the solution for this, adjust the band, or somehow > > adjust the 0 position in icub_head_torso_safe.ini? > > > > --Leo > > > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and threat > landscape has changed and how IT managers can respond. Discussions will > include endpoint security, mobile security and the latest in malware > threats. > http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Robotcub-hackers mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/robotcub-hackers ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
|
Hi RobotCub hackers,
I see that the tilt movement of the eyes is still an issue for you, so as you said, Leo said :"The last time we changed the head configuration files was several months ago (from last modified date), so I am pretty sure the cause is not a sudden change in configuration file" so i agree that if you dont change any parameters the robot should still behave the same. So taking in account the fact that nothing in the software part changed (ini file and other configuration file coming form update) it sounds like it is a electrical problem. I made an attempt to resolve the issue in my last post with the hypothesis that it was a mechanical problem because you experienced 1 single -10 degree slippery of the eye position, but now I will focus more specifically my answer by pointing out the fact that the cord of the tilt motor could be eaten or used at the bottom of the motor, along the cord or inside the header socket connecting it to the MC4. So as to enlighten this, I will ask you to remove the face cover and check visually the electrical cord of the motor, it's a 6 grey stripped wire flat cable (see example of cord from picture below) ![]() and take a look at how it is (dont hesitate to disconnect the header connector from the MC4 board to have a entire view of the cable, of course do first this trial with the PC104 closed). If it isnt obvious for you to detect the used part, it means it may come from the header connector crimps which may got looses or uncrimped or worse coming form the inside of the motor ![]() On this above picture, try, with a plier, to detect a failure in the crimps by pulling out the wires one after the other (the first 4 ones, by your own issue) and see if one of thoses wires get off its crimp. If yes just recrimp it and that's it. Let me know if it helps. |
|
Dear Julien, Lorenzo,
after checking all your suggestions, our conclusion is that the belt is probably worn out. We tightened it to the maximum, which got rid of about 11 degrees of the original 14 degrees tilt shift. The rest we fixed with the icub_head_torso.ini files. I guess we'll need a new belt soon, though.. Thanks for all your help! best, Leo -----Original Message----- From: julijenv Sent: Thursday, 28 June, 2012 13:13 To: [hidden email] Subject: Re: [rc-hackers] wrong icub eyes tilt Hi RobotCub hackers, I see that the tilt movement of the eyes is still an issue for you, so as you said, Leo said :"The last time we changed the head configuration files was several months ago (from last modified date), so I am pretty sure the cause is not a sudden change in configuration file" so i agree that if you dont change any parameters the robot should still behave the same. So taking in account the fact that nothing in the software part changed (ini file and other configuration file coming form update) it sounds like it is a electrical problem. I made an attempt to resolve the issue in my last post with the hypothesis that it was a mechanical problem because you experienced 1 single -10 degree slippery of the eye position, but now I will focus more specifically my answer by pointing out the fact that the cord of the tilt motor could be eaten or used at the bottom of the motor, along the cord or inside the header socket connecting it to the MC4. So as to enlighten this, I will ask you to remove the face cover and check visually the electrical cord of the motor, it's a 6 grey stripped wire flat cable (see example of cord from picture below) http://robotcub-hackers.2198711.n2.nabble.com/file/n7577428/Flat_cable.jpg and take a look at how it is (dont hesitate to disconnect the header connector from the MC4 board to have a entire view of the cable, of course do first this trial with the PC104 closed). If it isnt obvious for you to detect the used part, it means it may come from the header connector crimps which may got looses or uncrimped or worse coming form the inside of the motor http://robotcub-hackers.2198711.n2.nabble.com/file/n7577428/connettore.jpg On this above picture, try, with a plier, to detect a failure in the crimps by pulling out the wires one after the other (the first 4 ones, by your own issue) and see if one of thoses wires get off its crimp. If yes just recrimp it and that's it. Let me know if it helps. -- View this message in context: http://robotcub-hackers.2198711.n2.nabble.com/wrong-icub-eyes-tilt-tp7577397p7577428.html Sent from the RobotCub Hackers mailing list archive at Nabble.com. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Robotcub-hackers mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/robotcub-hackers |
| Powered by Nabble | Edit this page |
