LinkBack URL
About LinkBacks
Dunno your software, but most CAD software has a spline function and you can define the smoothness.
| | |
Does anyone know of a software that can take a csv or dwg file consisting of a set of x and y coordinates that describes an arc and identify the best fitting curve? I am working on a CNC grinder that is making a curved edge but the encoder that I am using leaves some jitter so that the programmed moves are not one continuous arc like move but instead many short segments that do not leave a smooth shape. I am using BobCad and Mach3.
Thanks
Dunno your software, but most CAD software has a spline function and you can define the smoothness.
http://www.kirkcon.com/
Moi3dOriginally Posted by thespindoc
rhino3d
You are looking for software that reads point data.
post a sample file.
here are a few points from the file
they are only significant to .001 I really hope to use the
points to identify the location of the part, then program
an ideal arc move rather that a point to point move so I need
to identify one large arc in the 8-9 foot radius range and another
toward one end in the 4 foot range.
-2.88209,-0.69001968891367
-2.90209,-0.68601968825802
-2.92209,-0.68301968962893
-2.94209,-0.67901968897328
-2.96209,-0.6760196828936
-2.98209,-0.67301968426451
-3.00209,-0.67101968766197
-3.02209,-0.67001968563542
-3.04209,-0.66801968903288
-3.06209,-0.66601968497977
-3.08209,-0.66401968837723
-3.10209,-0.66101968974814
-3.12209,-0.65901968569502
-3.14209,-0.65701968909249
-3.16209,-0.65401968301281
-3.18209,-0.6520196789597
-3.20209,-0.65001968980774
-3.22209,-0.64801968575463
-3.24209,-0.64501967967495
-3.26209,-0.643019690523
-3.28209,-0.64101968646988
-3.30209,-0.63801968039021
-3.32209,-0.63601969123825
-3.34209,-0.63401968718514
-3.36209,-0.63201968313202
-3.38209,-0.62901969195351
-3.40209,-0.62701968790039
-3.42209,-0.62501968384728
-3.44209,-0.62301967979416
-3.46209,-0.62001968861565
-3.48209,-0.61801968456253
-3.50209,-0.61601968050942
-3.52209,-0.61401969135746
-3.54209,-0.61101968527779
-3.56209,-0.60901968122467
-3.58209,-0.60701969207272
-3.60209,-0.6050196880196
-3.62209,-0.60301968396649
-3.64209,-0.60101967991337
-3.66209,-0.59901969076142
-3.68209,-0.5970196867083
-3.70209,-0.59501968265518
-3.72209,-0.59201969147667
-3.74209,-0.59001968742356
-3.76209,-0.58801968337044
-3.78209,-0.58601967931732
-3.80209,-0.58401969016537
-3.82209,-0.58201968611225
-3.84209,-0.58001968205914
Last edited by thespindoc; 09-24-2011 at 12:37 AM.
First I took you points list and created X, Y coordinates with it in G-code format.
Then I opened that file with PowerStation G-code Analyzer and output a DXF file.
Then I opened the DXF file with MasterCAM X3 to see the line segments.
Then I added the points to the end points of the line segments for clarity.
Then I constructed 2 arcs that seemed to best match the points.
See attachments.
http://www.kirkcon.com/
I was a bit confused about how these points would define the 9 and 4 foot arcs you described, but I took a page from txcncman's previous post and defined them as he did.
This file will contain the imported points, the curve that defines them and a 9 foot arc and a 4 foot arc going through the 2 endpoints of the segment that was imported...
If I ran a Tangent circle command on the 2 endpoints and ran along the curve anywhere, the best I could get was around a 40 inch arc (txcncman posted a 50 inch one I believe) So I just defined the 2 arcs at the dims you specified, then I rotated them to run through the points????
Anyway, maybe you could expound more about what you were looking for from these points?
Thanks for all the input!
What I am doing is developing a CNC ice skate sharpening machine, The blades
I am working with now are about 12 inches long. They are built with an initial shape
using an 8 or 9 foot radius circular arc over most of the blade but toward the front
the shape varies with 1 or 2 smaller arcs. I believe splines will not be as good as
one arc so I am trying to use the array of X/Y coordinates to locate the blade in
space and then I can sharpen using point to point moves for blades that are old
and have had many sharpenings but being able hopefully to gradually correct the
shape with subsequent sharpenings. I hope with new blades to just use the location
of the blade and describe the arcs for a smoother more consistent shape or
idealized shape that I can hold over time with subsequent sharpenings. I believe,
but have not tried any analysis, that programming a G03 arc will give a better shape
than following the point to point move.
I have also done the dxf import from G code and it works but finding the arcs and getting
them to fit is time consuming and trying to get the arcs to line up with the points accurately
is pretty difficult with BobCad/Cam. Hopefully I can automate the process mathematically
from CB.
The file I have uploaded is from a blade that is
old and sharpened incorrectly so it is not representative but presents a worst
case scenario. The front of the blade is to the right and it is lying horizontally on its side with the setup like a desktop mill and the grinding wheel is horizontal. It has been a labor of love over several years - love of CNC, love of fabrication, love of machining and love of skating!
Thanks again!
Keith
Hi Spin,
So if your contouring runners, I would think you would setup a template off of the blade type and ignore the existing cut:
Then you want to apply "your method" to that jig to reshape the blades:
So you already have "your method" layed out and ready to go, and are just looking to "place the jig" to apply and remove the least amout needed to reshape the blade???
Here is a file with the points and curve from your last post. I included a bounding box of the curve, one inch up in Y, from the curve.
The definition of the optimal placement of your "method" would need to come from a sampling of "your perfect shape" along the "sloppy existing cut" to determine the furthest point need to be cut, to place the new contour?
Yes that is the ticket!
Just need to program alignment of "perfect shape" along the "sloppy existing cut" to determine the furthest point need to be cut, to place the new contour. Otherwise, a good blade would just need a standard contour dropped onto the old edge.
Very Helpful Thanks!
Keith
Ok, on the same page...
MoI has an existing command to iterate out a line at a .001 interval for the 2 curves. Here's a dramatic representation of what that looks like.
We would need to script it to see if could could determine the "furthest point" between the 2 and just output that location to give your result distance needed...
I could ask the developer to see if he could cook it up, or you could join the MoI forum and ask yourself...
If MoI cant perform this, you may need a higher end software that can perform the analysis between the 2 curves to locate the point. (Dont want to have to manually measure 200 lines!! >)
Let me know if you want me to check, or I'll see it go by in the MoI forum...
Burr,
MoI looks very interesting. If I understand, I would then be able to look at the line lengths as a statistic to report variance from the ideal standard and know were the problem areas were? Should be able though to do it with algebraic formulas in a point by point analysis in basic?
Keith
Last edited by thespindoc; 09-25-2011 at 01:45 AM.
Yeah, you could write your own program to do that too.. I just asked in the MoI forum if it could be scripted .. If it does, you could just import your point file against the optimal curve and hit a button and have your location... BTW: MoI's scripting interface is JavaScript.
I'll post back here with the answer.. If it will, I'll also post an example...
You could look at the MoI forum for the response too.. It will probably come in a few minutes!!! lol
Posting Permissions
| | |
