Edge Scan Incorrect Vectors

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Edge Scan Incorrect Vectors

    I'm digitizing the die steel in the picture below. I'm running version 3.5 and attempting to do an edge scan. My current scan settings are 1 initial and 1 perm, so what it is doing is sampleling the top surface (one with the black marker) and then dropping down and taking the point on the surface I'm digitizing (one with the red line). It does appear to be compensating (learning) the vectors correctly for the initial hit on the top surface. My concern is it is not compensating (learning) the vectors for the actual surface I'm digitizing. It will not go around the curves, but just attempt to continue in the original vector. I did find out that if you up the initial and perm to 3 that it will take one sample point on the top surface, then take two sample points on the edge I'm digitizing to find the vector or the final edge point, but this method is taking to long. Is there a setting or something that will make the 1 initial and 1 perm method learn the vectors for both the top sampleing surface and the digitized edge? I've got about 20 of these things to do and those extra 2 points per hit add up. Thanks.

    http://i303.photobucket.com/albums/n...ector/Scan.jpg

  • #2
    Originally posted by blind829 View Post
    I'm digitizing the die steel in the picture below. I'm running version 3.5 and attempting to do an edge scan. My current scan settings are 1 initial and 1 perm, so what it is doing is sampleling the top surface (one with the black marker) and then dropping down and taking the point on the surface I'm digitizing (one with the red line). It does appear to be compensating (learning) the vectors correctly for the initial hit on the top surface. My concern is it is not compensating (learning) the vectors for the actual surface I'm digitizing. It will not go around the curves, but just attempt to continue in the original vector. I did find out that if you up the initial and perm to 3 that it will take one sample point on the top surface, then take two sample points on the edge I'm digitizing to find the vector or the final edge point, but this method is taking to long. Is there a setting or something that will make the 1 initial and 1 perm method learn the vectors for both the top sampleing surface and the digitized edge? I've got about 20 of these things to do and those extra 2 points per hit add up. Thanks.

    http://i303.photobucket.com/albums/n...ector/Scan.jpg
    You ain't gonna like this, well, your CAD guys ain't gonna like this.

    SCREW doing edge scans, you won't like it, they don't work for crap.

    Use VECTOR POINTS.

    Turn the probe comp OFF and just do a series of short linear open scans. Take 3 touches that are basically in a straight line, use the last touch of the previous scan for the first touch of the next one, basically, TURN OFF measure in the scan window. Take 3 hits, start, direction, and end, and hit the END button to create the scan (without measuring it). Then touch again where you just ended the previous scan, and take DIR, and END touches and continue until it is ALL done. Then go back and execte them.

    Sorry, but this is pretty much the way you will have do to it, then the cad department will have to comp for the PROBE RADIUS you used.

    Also, it will run for CRAP. When you turn the COMP OFF and try to run, it will AUTOMATICALLY (more or less) add the probe radius to ANY prehit/retract you have set in the program. BEST bet will be to set a VERY LOW prehit/retract (like 0.25mm). BUT, if that is ALL you do, it will never measure a hit, you will need to add
    Code:
      CHECK/ 5,1
    to the program so that it will make contact. Trust me, that will be required. AND, no worries about SIN or COSIN errors since the probe comp is off. I can't TELL you the number of details I've had to do that look EXACTLY like that and this is the only way.
    sigpic
    Originally posted by AndersI
    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

    Comment


    • #3
      Essentially if you need two vectors to be found from one measurement you've got to do the extra points.
      --Brian

      "The best way to predict the future is invent it. This includes your very next action."

      Support: Hexagon Metrology Support Center
      Training: Hexagon Metrology University

      Comment


      • #4
        Matt, You seem to always be the bearer of bad news. I will use your method for things that need scanning in house, however these were contracted to me so I want to make it as easy as I can for the other companies CAD department. I did ask them about your method and having the CAD department just offset for the probe thickness and they said it's just too risky. I guess I've got to make them happy first. I also have these running on a second machine so the more it can run by itself the better. But thanks Matt, I will use that for in house stuff.

        Brian, that's kind of what I gathered but I figured since PC-Dmis can calculate the correct vectors based off the previous points in a vector scan, why can't it just do it now...

        Comment


        • #5
          Originally posted by blind829 View Post
          Matt, You seem to always be the bearer of bad news. I will use your method for things that need scanning in house, however these were contracted to me so I want to make it as easy as I can for the other companies CAD department. I did ask them about your method and having the CAD department just offset for the probe thickness and they said it's just too risky. I guess I've got to make them happy first. I also have these running on a second machine so the more it can run by itself the better. But thanks Matt, I will use that for in house stuff.

          Brian, that's kind of what I gathered but I figured since PC-Dmis can calculate the correct vectors based off the previous points in a vector scan, why can't it just do it now...
          I have TRIED the 5 sample hit edge point scans, they DIDN'T work! You will NOT get correct probe comp! As I said, I have done "those" details. They do trim-after-form. They trim edge is "straight" up and down, but you will ALSO get a Z vector on ALL those points and thus it will NOT be correct. I've been there, done that.
          sigpic
          Originally posted by AndersI
          I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

          Comment

          Related Topics

          Collapse

          Working...
          X