"Bob La Londe" wrote in message
...
"Bob La Londe" wrote in message
...
"Bob La Londe" wrote in message
...
http://www.yumabassman.com/pictures/...20Axis%20b.JPG
Well, there it is. It's a combination of design to fit and machine to
fit techniques, but it screams up and down at 40 IPM. (Fastest I can
run without changing the kernel speed.) I expected to be able to stall
it very easily with the stepper running that fast, but I was unable to
stall it by bracing my hand against the table and bushing upward with
the axis rapiding downwards. (I think the G450 must switch from micro
steps to full steps over as certain step rate. I'll have to ask Marcus
about that.) Now I probably need to map the screw. Repeatability is
very good, but accuracy is pretty bad at different distances along the
screw. I never had to map a screw before. Funny that the first ball
screw I install has the worst accuracy of any screw I've used. LOL.
Just for the heck of it I bumped the kernel speed up to 65khz and I got
53 IPM on rapids without stalling on a 25.4 TPI (1mm) screw. That's
nuts. I'm tempted to bump it up again just to see where it does stall.
Oops. 45khz. Not 65.
Cut my first mold with the new setup. Wow! It looks really good. I admit
I took several steps to stiffen up the Z axis and spindle mount all at the
same time, but this little ball screw is awesome. Oh, and when I measured
travel with my dial indicator (to start mapping the screw) it was only off
about .0004 per inch at the worst spot. Actually less near the end were it
will do most of its work. Nowhere near as bad as I thought when I quick
checked it with my caliper. Its still not any better than C7, but its not
as bad as I first feared. I haven't checked backlash on it yet, but linear
movement is pretty good. I now have a machine I can smaller prototype molds
on while the Taig is grinding its way through the "bigger" stuff. I never
trusted the little machine before.
Still kills me I am running 50800 steps per inch and over 50 IPM with good
holding/moving force and no stalling.