Milestone I Demo Feedback

Logistics

• The score for motors are higher -> The score for control is scaled
• Self assessments need to be more concise, specific, and measurable

Demo

• Prioritize show first, then explain
• Organization is vital (use PowerPoint if necessary)
• Show the best features clearly (design freeze)

Control

• Keep in mind about the spacing of the encoders
• Obtain ALL of the motor parameters (especially mechanical parameters)
  • Extremely important for later on to find paramters on the custom made motor
  • Electrical
    • R, L, C
  • Mechanical:
    • J, B, K (don’t need to unless we’re actually adding spring)
  • Electromechanical:
    • Torque constant
    • Back EMF constant

Motor

• Windings are done nicely and cleanly
• In the future when using laminated cores, make sure there is minimum airgap
• (Potential gotcha) When using metal windings, the cross section of the core is rectangular, but the winding is curved; thus the insulation near the corners of the core are worn. This can cause the current to all go through the core and render the rotor useless.
• Cylindrical commutators are common. Commutators in general needs more work
• Most motors are in a test bench (good), Make sure the parts of the test bench are contrained (fix them together)
• Most brushes are not mechanically mounted. Making sure the brushes are making just enough contact with the commutator without introducing too much friction
• Know if the motor will self-start (for fixed commutator brushes)
• Make sure the base of the test bench is not flexible and flimsy

Couple of More Things

Control

• Derivative term (using finite difference derivative): (Pf-Pi)/(change in time)
  • If the (discrete) signal is slower than the sampling time, the velocity over time will appear like an impulse train
  • Some filter (such as averaging) is required
  • Faster control frequency is better, but will cause problem with derivative term
• Make sure ISR is short and as fast as possible
  • Bloated ISR will impose lower control frequency
  • Never want ISR to be interrupted
• A nice milestone is to have a real open-loop (no feedback) step response similar to the simulated step response
• Then do a close loop step response with PID, and compare it with the simulated step response

Motor

• Problems with cylindrical commutator
  • If the commutator is not perfectly round, the brushes are bouncing around
  • Short gap cause sparking and bouncing brushes
• Try disk commutator
  • Put brushes near the center of the disk (low speed, low friction)
  • Make disk out of PCB (put spot of solderable spot for winding connection)
• Quickfield.com shows magnetic field when you put windings around magnetic material; shows saturation.
• Glue won’t work for laminating metal; paint works well
  • Use paint to laminate
  • Use nuts, bolts, and washers to clamp them overnight
• Generally bad idea to glue the shaft with rotor
  • Drill a hole and use a set screw

MileSTONE II

• Build a closed motor for use