Re: [CAD_CAM_EDM_DRO] Re: Rutex Z axis problem--more info

tig415 wrote:

>>OHHHH! You may be on to something, there! I've been trying to
>>figure out what was going on, but a loose encoder coupling could
>>provide EXACTLY this kind of error. It would include a backlash-
>>
>>
>like
>
>
>>effect. If you are not "losing steps", ie. a change in absolute
>>position, it
>>is not a classic encoder problem like a dirty wheel or weak LEDs.
>>
>>
>But
>
>
>>a shaft looseness could do things like this. Do you ever see the
>>
>>
>axis
>
>
>>twitch
>>suddenly?
>>
>>
>
>Hi Jon
>
>
>No twitching
>
>I'm using glass scales for my encoders. when sitting at idle The
>axis pulse back and forth a small amount. I suppose this has to do
>with the fact I'm using scales as opposed to shaft mount encoders.
>If I set the incremental jog to .001" and make a move the axis
>moves very smoothly till it gets to the end of the over travel then
>it gently pulsates while it's correcting back.
>
>
>

A number of glass scale designs use a piece of music wire to move the
read head
across the slide. If the read head had got gunk in the track, it might
be causing the
read head to stick. This could also explain the motion you see.

>It never jumps. It always moves very smoothly.
>
>
>

So you only see this when you STOP, and then start motion again, I guess
that's
what you are saying. What is the backlash? You can measure it two
ways. One
is to move to a specific location, read physical position on a dial
indicator, then
move past that point, reverse and come back to the same coordinate on the
computer. The other way is to move slowly in one direction, stop, then
start
making tiny incremental jogs in the other direction until the indicator
moves,
and note how much the computer diaplay has changed. Maybe comparing these
two tests might bring out what is happening. Do you have a digital storage
oscilloscope, or can borrow one? It would be REAL instructive to look at
the velocity command out of the Rutex converter, and see if the Rutex is
commanding
such a move. You'd see the voltage increase in one direction, then pass
through zero
as it overshoots and moves back.

Jon