Re: [CAD_CAM_EDM_DRO] Newbie - Z-axis slipping

Janice wrote:

>I'm brand new, self taught on my CNC router/engraver. I am
>currently using a single drill point function where I am drilling 30-
>90 holes per part, at a precise depth of 0.043. I am losing depth
>starting at about hole 25, and losing up to .030 worth of depth over
>the course of the remaining holes. Am I expecting an unreasonable
>level of precision, or are there some explanations for why a Z-axis
>would lose depth? Thanks in advance!
>
>

Unfortunately, these things can't be easily diagnosed at a distance.

What software are you using?
What motors, drivers, etc. are you using, or is this a commercial
CNC package?

Are you sure there is no error on the first 24 holes? Unless something is
heating up from the effort of moving the machine, that almost sounds
like a software problem. Can you pause the machine at hole 20 for a few
minutes, then let it run for 20 more, etc.? If there's no error, that
would almost prove it is something being affected by heat. That would
have to be the motor drivers or the step motors themselves. If the error
is identical, then it would point very strongly to software.

If there was a constant drift, that could be step motors losing an
occasional
step during the forward stroke, perhaps, or even an incompatibility between
the timing of the step/diraction signals the computer is putting out and the
requirements of the motor driver.

Is there any possibility the drill bit is retreating into the chuck?
I've certainly
seen that before. Are you sure the machine moves in a parallel fashion, ie.
the spindle remains at the same height above the workpiece surface as it
travels down the part? I can easily see where a flexible workpiece laid
over a lumpy table would flex .030" in a foot or two.

One check would be to see if the spindle returns to the same height over
the first hole. Do you have a dial indicator? Attaching the indicator to
whatever the spindle mounts to, and running the head across the size of
the workpiece would prove that it is not a problem with flatness and
parallelism of the work table.

Jon