Jim Struts sort of Walked...

 

Third week of July

Have spent most of the week continuing writing the leg management software to see if I can get Jim walking.

Phoned Mentorn to see if they will be using Jim Struts in the program. The answer I got was some what unclear, so I await the promised letter. The letter did not turn up on Saturday, will now have to wait for Monday (did not receive letter until the following weekend). I am assuming that because Jim Struts is overweight he will not be used in the competition. But because he is a walker they want to use him in their program! For good measure I registered for a grudge match against the Robot Wars Production Team this is probable a very stupid thing to do in light of the problems I have had this week end. " C'est La Vie "

Have spent the weekend installing, debugging and testing the leg management software, it has not gone well.

EyeEye - Jim Struts - Eddie the on Board Computer
Eddie the onboard computer is mounted on hinges so that he folds out the back of Jim Struts to allow easy access for removing and fitting the micro-controllers between programming. The two outer tag strips are for the mounting of the Mosfets and terminating the Hydraulic solenoid cables. The centre tag strip is for terminating the sensor cabling.

First problem the cpu kept resetting itself by the watchdog. This meant the walk cycle would keep restarting and because of the 'where to start the cycle' detect routines, was continually being recalled. The legs where behaving very un-predictably. But at least they went backwards and forward, just not very linearly, and no up and down.

Fixed resetting problem. Now the legs are working in the correct cycles, but cannot control the walking parameter from the remote control. Adjusting the parameters results in unpredictable actions. The height adjustments seems to be just about okay, but there is no control over the step size etc.

Could not find the leg parameter problem so by passed it and coded up some fixed parameters. This has resulted in a reasonable leg action. So removed Jim from test stand and tried him out walking for real.

Well Jim Struts sort of walked... But Oh dear its awful. He is ever so slow, AND I MEAN SLOW. The legs stutter, I think that is caused by the feedback loop from the sensors via the cpu back to the cylinders oscillating!

Have spent the rest of the weekend trying to fix the oscillating problem but to no avail, and feeling very depressed.

Thinking of dropping 'Jim Struts' out, as my timetable has me down for completing the walking software this weekend. There will not be time to grip the software and upgrade the engine to improve the speed and walking. He is very slow. Even the slowest wheeled Robot would run rings around Jim.

Last Week of July

Took a day off work to have another go at the software, with some success. The oscillating problem was caused by the feed back loop between the Leg sensor and the Ram/Solenoid. The walking parameter problem was caused by not resetting the carry flag before rotating a register.

We have managed to have several very successful runs with Jim Struts; he walks, but not well, but well enough, and very slowly. He fell on his back into a flower bed crushing a rose bush. My wife was not entirely happy, I thought it was very funny. My son William has drawn a cartoon of the event. Jim Struts and the Rose...

We also pushed a car tyre up the garden about 10m. The legs got jammed once as the garden slopes and Jim Struts walked into a hole which aggravated this, this was quickly solved by a few moves of the right legs via the manual control on the remote control, and tyre and robot continued up the garden.

Jim Struts Pushes a Car Tyre
 EyeEye - Jim Struts - Pushes a tyre up the garden   The picture is poor quality as it is a grab off the video. It is at an angle as the garden slopes, not because of a drunken cameraman.

We are still unclear as to what Mentorn will be using Jim for if at all....

We are in...

 

 [ Jim Struts ]
Copyright © Ian Inglis, 1998-2014. All rights reserved.