Try this a second time since it didn't appear to post...
Are you really able to get a repeatable and accurite resolution of
0.0005" on your ACME screw? What about motor oscilations and over
shoot
(gotta love steppers)? Then there is the question of your guides
and
play in them. You might be asking a lot to fabricate that. Take a
look a
Dadel / Parker and see thier specs. Then remember they are doing
this
for a living, granted in larger quantities with out the TLC. I am
just
asking the question.
John
Hi John, All questions comments, ect. welcome.
First off, this is my first stepper project, its a learning
experience that I plan on putting to use and basically a fun project.
Its for hobby use, not to compete with $5K+ commercial units that
intend to be sold to NASA, MIT... IOW It has to be Good Enough for me
and my use.
I estimate I'll have about $100-200 invested (less software) in the
unit. Steppers, linear rails and bearings, 8020 (
) and plate aluminum frame, drivers, power
supply... most scrounged from a local scrap dealer and some bought
new or as new old stock. The Camtronics PWM drive was donated by a
friend with a machine shop who wants to use it ocassionally, and will
help with any machining beyond my abilities. The computer to run it,
well, I have a lot of old parts that need put to use. Considering my
investment (time is not a factor since this is a for fun hobby
project) I will have a substantial savings over a commercial unit,
and can upgrade it when/if needed.
Resolution and repeatibility (accuracy): Good points, and no I don't
expect extreme accuracy and repeatability out of this setup,
especially using Acme rod, but i think it will be rather accurate.
What Accuracy (resolution and repeatibility) is Really Needed?
Going by lathes, mills... stiffness of the frame lends to accuracy,
this is why i'm using 8020 and aluminum plate ($0.35/lb scrap price).
It will be rather hefty. Also backlash is a factor so I'm going to
use antibacklash nuts that I will make out of Delrin (for reduced
friction and wear of the Acme rod lead screws).
My Guides: For X and Y axis (right and left), Thompson TWN 10 (5/8")
Super Pillow Block Bearings and Precision Ground Rod. their used but
are in Excellent mechanical condition, smooth and tight.
For Z axis (router verticle), used Robohand Inc. model PS-4-4-SH
(similar to these ) with 12mm
shafts and 4" travel. Very smooth and tight. I also have some 3/4"
supported linear rods and bearings but I think they would be
overkill. :)
The part that I have the least knowledge about are the stepper motors
themselves, so I'm going a bit larger than the ones I've seen on
commercial equipment I've found on the web. I have 2 new (old stock)
Vexta 83 oz/in, 1 used 100 oz/in and 1 used 120 oz/in, all 200 full
steps per rev and NEMA23 frame. This one (
tech.com/products/quickcircuit/ ) appears to be using 60 oz/in
motors, .00025" resolition and .00025" (one halfstep ?)
repeatibility. Check out the PCBs it makes. :)
Speed of the cut is also a big factor in accuracy, so I intend to
keep my cutting inches/sec on the slow side, at least until I see how
fast I can move and not affect accuracy to a great degree.
What Accuracy is Really Needed?
Last but not least the Acme rod. I won't be sure of its linearity
until I build this project and test it with a dial indicator. I can
also get some precision Acme rod if needed, and then theres
ballscrews. As long as its relatively linear I can adjust the overall
accuracy in software. I do have a long ways to go (cost wise) to get
close to the cost of a comercial unit. :)
The design is a Fixed Gantry, the table only moves in the Y axis
under the gantry. The gantry holds the X axis and the Z rides on the
X. I may modify this swapping the X and Y for added rigidity.
The power supply parts are scrounged (free) from old power supplies
from work.
BTW using a RTX for the spindle has been changed to using a small
laminate trim router. I found one (and a spare) at Harbor Freight for
half the cost of a RTX ($20 ea.). Thats cheap enough to replace as
needed. :)
What do you think?
mike
