Re: MaxNC modifications

Jon Elson wrote:

>
> Wow! this really DOES sound like a software problem. I
> really don't believe this is the motor losing steps due to
> drag, electrical noise, etc.

Thanks Jon!! I'm no expert at all this, but it's been my growing feeling
that it's software. With the MAX step/direction version, I was quick to
asssume software bug. Otto certainly has done a fantastic job of keeping
prices down, but there's a cost associated with that.... I recommended
my Dad buy Ah-ha software due to the number of folks using it
commercially, and bought my own copy before realizing we still had the
lost step problem. (I'm finishing up a Sherline for Dad, right now I'm
using his computer/software/drivers on my mill, but I've got all the
same hardware for mine when his Sherline finally goes home with him)
Ah-ha has suggested several items that can result in lost steps and
we've addressed them with no cure. I cannot fault Ah-ha for not bending
over backwards to help diagnose our home brew setup, I'm sure they'd
show more interest had we bought a complete system.

I have taken great care to insure the slides and screws are not binding.
I'd originally used two bronze acme nuts, one on either side of the
saddle for backlash adjustment. One fixed, the other adjustable with the
screws prevented from moving by use of VibraTite.
Going for a fringe problem, part of the reason I went with the ActiveCam
nuts was to insure there was some drag on the screw during axis
reversals. Someone had mentioned to me that ultra-low inertia motors
like the PacSci we are using, can be prone to losing steps during
reversals if the motor suddenly has no load for an instant. We've also
fitting 2" dia brass gears to the rear shaft on the motors trying to use
the mass to damp vibrations. (though since they had no effect, they've
been removed)
I know what lost steps sound like from binding (please, don't ask
how..<G>) and even a few lost steps are audible especially if one is
listening for them.

This then, is the attraction of EMC for us. We are both on a tight
budget and spending big bucks on commercial software isn't much of an
option. Fred has said he can modify EMC to run an encoder in stepper
mode and close the position loop much like a servo.
It's my guess that EMC in stepper mode probably won't lose steps at all,
but if we can get the encoders inplemented, that will give us backup
protection. The eventual solution as far as I'm concerned, is to go
servo. Towards that goal, the MAX and Sherline frames are just test
beds, we will eventually build some something with linear rail bearings
and ball screws.

Jon