Odd results carving 3D STL file in Aspire

[blackhawk]

I have a car that I designed in a separate 3D design software. I exported it to Aspire as an STL. Everything in the preview looks good, but I get these unexpected cuts on the front and back. My machine cut the back with an extra groove along the X direction. At the front, it didn't clean up completely and had a few gouges. As you can see, the Aspire preview looks clean. I ran both roughing and finishing with a 1/4" ball nose at 2.8 inch/sec. After I saw this problem, I recalculated the finish toolpath a second time and ran the file again, but the same results. I have a PRT Alpha and I checked all the pinions for tightness and the collet and dustboot had no interference. This is the first time that I have ever cut anything that didn't match the Vectric preview perfectly. Please, any help is appreciated.

8420
8421
8422

[blackhawk]

OK, I just noticed something. If you look at my Aspire preview you can see that I left 1/4" on all 4 sides of the model. Now if you look at the 2nd (front) picture of the actual carving, you can notice that it carved the nose at exactly the zero point. The back side is carved at 1/2" from the edge of the blank. So, for some reason this thing cut exactly 1/4" off in the Y direction. I went back and checked the Shopbot screen preview and it looks fine, just like Aspire. I checked my zero and it is correct. I bumped the Y to +.002 and then rezeroed here. I ran the file again but it still cut the same. I also looked at the file in the editor and there is no command that sends the Y negative past 0.0000

I went back to Aspire and changed my finish toolpath from an offset strategy to a raster strategy along the Y. Without changing any zero points, I ran the raster file. It ran perfect and I had my 1/4" spacing front and back.

This is crazy. Why would the offset toolpath look good in Shopbot editor and preview, but then actually cut exactly 1/4" off in the Y?

[blackhawk]

I shut down the machine and computer and went to dinner. Turned everything back on tonight and ran my roughing toolpath, no problems and everything had 1/4" all around. I then calculated a brand new offset strategy finish toolpath just like before. Ran a shopbot preview and looked at the numbers in the editor, everything was good. Ran the file and got the exact same bad cut as earlier in the day. This time I watched the position window and I could see the Y axis going negative by as much as -1.67 inches. I double checked in the editor and the minimum Y is exactly 0.000000 in the .sbp file.

It can't be Aspire. Anyone have any ideas? I am still baffled. If anyone would volunteer to look at my .sbp file, I will email it to them.

[khalid]

Can I suggest you one thing.. Export the file in another format instead of STL and then bring it into Aspire... Then check the results.

P.S: Some time STL files having minor problems that not can be seen by the CAM software.... One of my friend was having problem with GRIP stl file..He gave me to make a toolpath for it...I did toolpath in VisualMill and he was getting the same results some notches in corners, that were visible in machined part but hidden in CAM software... then I enlarge that portion in visual mill ans inspected the toolpath..their was some tool collision problemm...
I did the same part in Artcam and that problem was resolved...
Download NCPLOT and inspect the tool retract i think the tool not retract completly and keep travelling in X/Y direction making these notches...

[blackhawk]

I first suspected the STL and/or Aspire was the source of the problem. But I inspected the actual cut file line by line and there is no position command to send the Y axis to a negative value. Since I could actually see the position move negative while cutting, it seems to be a Shopbot machine or control problem. I used the same STL file to generate a raster strategy and that worked fine.

[blackhawk]

I ran some more experiments today. I ran my problem .sbp file again and watched the actual code values and compared them to the red position window. At line 1500, the Y was off by only .003, but then continually got worse:
line 2200 was .010" off
line 2700 was .022" off
line 3000 was .049" off
line 4000 was .163" off
line 5000 was .270" off
line 5450 was .169" off (starting to get better)
line 5500 was the last line of code sending the machine back to the 0,0 start point. It truly returned to the exact zero where I started. I know because I measured the position of the Y carriage to a relative point using calipers before starting the file.

I ran the file a second time and the values were getting off identically as before. Around line 5000, I hit the space bar and paused the machine. While paused, I measured my position with calipers again and it was off exactly the amount that the red position window was saying (about .300"). So, the red position box was accurately giving the position of the Y car.

Now, I resumed the cut and now the commanded positions started to match the red position window values! So, I let the file run to the end and when it went back to its 0,0 home position, the Y car went to rest past the Y zero point by .300 inches. This is about exactly what I had measured during the pause. So it seems like the pause reset the red position box to match up with the code, but the physical machine position did not change.

Finally, I re-zeroed Y, since the previous file knocked it off. I put in a new piece of wood and ran the raster finish strategy that I mentioned in my original post. The raster strategy worked fine and the car came out perfect. I watched the code run on this file and it was never off by more than .001 of an inch.

Has anyone ever seen such a thing as this? All of these tests can be repeated with identical results so it does not seem like anything sporadic is happening?

[blackhawk]

Anyone seen this before, help?