Pulleys, motor boards, motors, etc.

[Tim Vanderhoek <vanderh@ecf.utoronto.ca>]

The following notes are copied from my lab notes for today...

	- if the metal pulley were smoothed more, I believe the short belts
	  would not snap.  I think the short belts will work fine, but,
	  it will be necessary to either smooth the metal (it's currently
	  grooved/threaded) or to plastic-coat it (Joyce suggested heat
	  shrink).  The metal grooves quickly eat into the rubber belt,
	  right now.  I did hacksaw several bolts in order to lower the
	  motors more, thus reducing the force on the pulley from the
	  belt.  This did not prevent the belt from being chewed-up.
	  I know the belt is getting chewed-up by the fact that little
	  bits of rubber come flying off of the belt...  (This was in
	  question at an earlier time.)  I think it will work fine from
	  both the higher height and the lower height provided that
	  something is used to cover the threaded metal on the pulley.
	- o I am really concerned that the touch sensors are too sensitive.
	    I can add some code to the computer to try and ignore random
	    bounces, if necessary, but that means that when we do hit a _real_
	    wall, Zeta will be slower to react.  
	  o The touch sensors are fragile.  The metal prongs get bent easily.
	    That is indeed what happened to the three that I had to take off
	    of Zeta on Monday.  While working with Zeta, that happened to
	    another two of them.  It's possible to bend them back into place,
	    but this can only be done so many times before the prongs fall
	    off of the sensors.
	  o For these reasons, I think we should investigate replacement
	    touch sensors.
	- Joyce put new bigger capacitors on the motors to try and prevent
	  them from resetting the comp board.  I think they work, but I can't
	  be sure since the capacitors shorted after about 5s of running,
	  which is not long enough to verify that they don't reset the comp
	  board.  Once the caps shorted, of course, the comp board promptly
	  reset itself.  My understanding is that these were 500V capacitors.
	  We have some 1KV capacitors that we can try, but if these actually
	  were 500V, then we should probably go all the way to 3KV capacitors.
	  PWM control of the motor obviously creates much more backvoltage
	  than does normal use of the motor (the capacitors were fine under
	  normal use).
	- The following components need replacement (next shopping trip):
	  o The metal prong no longer stays in one (two?) of the metal touch
	    sensors.  We still have two (three?) spare touch sensors, and it
	    can be glued back in, if necessary, but, see earlier comment.
	  o Should buy some nice large 3KV caps.  If the caps that were
	    supposed to be 500V blew so quickly, I can only assume that the
	    1KV caps I bought will also blow sooner or later.
	  o When we were taking the batteries ot of our of the locker, the
	    battery case that fell cracked.  It was an 8x1.5V = 12V case.  We
	    have one 8x1.5 left.
	  o Many small rubber belts have been destroyed.  I now have a very
	    set procedure for putting them onto the robot, however:
	    - Remove wheel.  Belt may snap if you try to pull belt over wheel.
	    - Remove screws behind wheel.
	    - Put belt onto the small motor pulley first, then pull it onto
	      the large pulley.  Belt may snap if you try to put it onto the
	      large pulley first (snaps when you try to grab it afterwards).
	    - Rescrew everything (or glue bonds break).
	    This is a significant deviation from the procedure we
	    followed during lunch.  If we have any thoughts of using the
	    small rubber belts, we must purchase many more of them.
	  o Neither the motor control circuit nor the computer appear to have
	    been permanently damaged by the shorted capacitors.