From Maslab 2007
|Team 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 17 · 18|
|Team Three’s Journal|
We made a crude assembly of our bot in SolidWorks. In the past week, we've fashioned the baseplate of the chassis and some mounts from 3/4" birch plywood in Edgerton Shop. In terms of sensors, short-range IR sensors on servos will be used to detect walls obstacles (i.e., walls). A gyroscope and a USB optical mouse will be coupled for reasonably accurate odometry. The roller for our bot will be driven a nifty right angle motor that Darthur suggested. Star Simpson suggested that we make the roller by mounting the bristles from paint brushes. They seem to have the right consistency, so we'll try it out.
Alex Hornstein helped us design a drive assembly using pulleys and fitting tooth belts that we crufted. We also purchased mounted bearings for attaching the wheels to the chassis. Hornstein also showed us how to make our own axles (not unlike those for our pegbot) by lathing aluminum. The only (not rusted) blade for the band saw at MITERS had been previously used for cutting fiberglass(!), but was still useful for cutting aluminum brackets for mounting on our chassis.
- Assemble the drive assembly from the pulleys and mount them to the chassis
- Try making a roller from the tips of paint brushes and couple it to the right angle motor
- Fashion an aluminum cage and attach it to the bottom of the chassis
- Figure out how to mount the mouse and the back roller.
- Mount the battery and the computer using aluminum brackets
Monday, 15 January 2022
We rest, in observance of Martin Luther King Day.
Tuesday and Wednesday, 16-17 January 2022
Our parts from McMaster-Carr and All Electronics came in today, so we should have a functioning robot by the end of the day. On a sidenote, don't order from All Electronics unless you really need to: their overnight shipping tacks on ('dollar (+ 33 (* 3 weight))) to invoice. Note that 'weight' is the weight of your part.
It turned out that having our robot done by Tuesday was a bit presumptuous.
On Tuesday we successfully mounted our drive system only to discover on Wednesday that the bolts we used (more out of desperation, as better bolts were not available) were too long and extended into the ball cage beneath the robot. We ultimately resorted to wood screws, that were shorter and better secured the motor block, but made the possibility of easily adjusting the motor block in the future more difficult. The ball-cage and a nifty mount for the optical mouse were built and are waiting to be attached on Thursday.
Note: The process of mounting and remounting the engine block took a very large portion of these two days. MITERS has lots of wonderful tools to work with, but a rather large and disorganized collection of fasteners that makes the process of building much longer than it should.
In any case, the chassis of our bot is very, very well built. The frame was pretty resilient when we accidentally pressed down on the chassis with a mill while drilling holes. The battery and the computer fit very snuggly into their respective places on the chassis, and a second-level hemisphere has been made to fit on top, leaving access to the ports on the computer and a place to mount the OrcPad/OrcBoard.
More on building tomorrow...
Thursday, January 18, 2022
There a couple differences between our SolidWorks design and the ultimate implementation.
A right-hand motor replaces the original motor provided by the MASLab staff. Unspecified parts of the design (such as the drive assembly) have been replaced by pulleys and gear belts. The size of the roller and the ball-cage has been reduced. The camera hasn't been mounted yet. The computer and the battery have switched places so the center of mass is closer to the robots axis of motion...
Friday - Sunday, January 19 - 21
We decided to replace our bearings with lower-friction mounted bearings of our own construction. The robot now appears to be be moving around the playing field with minimal mechanical difficulty. More updates tomorrow as we start testing more software features...
It appears that the new bearings introduced a number of new mechanical problems that eventually simplified to issues of torque:
- The weight of the bot required more torque on the behalf of the motors to actually move. This became evident when the bot failed to clear a carpet hump during the mock contest on Thursday. On a side note, LDVL's main problems during the mock contest were entirely mechanical (as will soon be evident).
- Also related to weight, the weakness of the motors was only made worse by other imperfections in the design that prevented the left motor from powering the pulley assembly at all. Trivial torques (i.e., rubbing against wood) would prevent the left wheel from turning on, and subsequently, the bot would start turning in left-handed circles.
- The pulley assembly pulled down the axle connected to the motor, while at the same time a counterforce was applied to the motor so that the tooth belt would have sufficient tension. As Big Slocvm pointed out, supporting the end of the axle with a simple bearing (like greased wood, for example) would largely prevent those two forces from severely damaging the motor/axle.
A big part of this week was spent on repairing the issues above. Our ultimate hope was that Slocvm's suggestion of adding a block of wood with a joint for a greased axle would work. For a lot of identifiable reasons (that were then fixed), this solution didn't work out so well. Our robot continued to move to the left, and due to those unidentifiable reasons ("Why is this STILL turning left?"), we remounted the motors to couple directory to the wheels.
At this point, you might wonder "Why go back?" It turns out that the lateral space occupied by a motor and that occupied by a bearing assembly is roughly equal. Through are different mechanical drafts, it seemed to us (although less evident in the CAD file), that the remaining space after motors or bearings were added was still sufficient for a winning ball collection strategy.
In any case, we rebuilt the motor assembly and the robot seems to be moving in very, very straight lines. It can proficiently collect balls: