iCubGrenoble01 S/N:023 – powering the iCub motors fails with blinking red button

[EliseiFrederic]

Robot Name 🤖

iCubGrenoble01 S/N:023

Request/Failure description

Fails to power the motor of the overhauled iCub just back from IIT: when we press the backpack button, it starts blinking green but then blinks red after either 2 seconds or 10-12 seconds

Detailed context

iCub Grenoble is back from a several months overhaul. We're trying to get it powering on and moving.
We changed the cabling of the double PSU setup previously used for PC104, now delivering 40V (+20A max) through one of these PSU (no more remote control DBx link).

Ok for icub_head CPU: button switchs to green after press, and I can connect to it through SSH.

*Not ok for powering the motors: when I press the backpack button, it starts green but it starts blinking green but then blinks red after either 2 seconds or 10-12 seconds

Additional context

Probably not related (but I'll post a video), but a fan is making a lot of noise (touching ?) in a shoulder. Gentle press on the shoulder, or even balancing the robot changes the noise...

How does it affect you?

Cannot validate the bill if I cannot verify that the robot is working after the trip!

[EliseiFrederic]

The emergency button has been checked: one pin was closing the loop and the other one opening it when the button was up. Pressing it closed the opened one.

[EliseiFrederic]

We tried either of the two PSUs: same problem

[EliseiFrederic]

We used a current clamp + oscilloscope, and found that the robot stops -- with either of the two PSUs -- when reaching 5.64A. At every run.

[AntonioConsilvio]

Hi @EliseiFrederic! @Gandoo and I came to Grenoble in person to solve this problem.

We investigated a bit and arrived to the conclusion that if the robot is powered from 39.5V upwards it has the power supply problem described in this issue.

If, on the other hand, it is powered at 38V or 39V, it works correctly.

Since 38V is considered a regular and normal usage regime, we left the robot like this and it works fine.

Then we worked on the neck sheath to be able to move it sideways and allow the robot to go further back with its head and look higher.

In order to do this we had to replace the dragonfly cable as the ones fitted were very stiff, whereas the new ones are very elastic and allow the sheath to be moved sideways.

Then we fixed the neck cover using detachables, so that everything stays tight and in place.

During testing, we worked on the configuration files so that the robot would be able to perform the demos.

Here is the PR for the configuration files:

• [iCubGrenoble01] Change config files after maintenance service robots-configuration#736

The robot has been tested and is working properly. 🚀

MicrosoftTeams-video.12.mp4

If there are any problems in the future feel free to open new issues.

Closing! ✅

[EliseiFrederic]

We've been using the robot during 3 full days now. No problem with the 38V so far, thanks, and the other improvements rock!
For the face though, problems are still very recurrent: calibration fails, jaw does not move to its limits (wether controlled in Position or PositionDirect), lip corners or jaw go in fail mode when driving speech (PositionDirect).
I'll post a new issue on that, as we can't do any face-to-face interaction experiments in the current state.