Finding a radius with smartprobe.ngc

Results 1 to 10 of 10

Thread: Finding a radius with smartprobe.ngc

  1. #1
    Member
    Join Date
    Mar 2006
    Location
    USA
    Posts
    51
    Downloads
    0
    Uploads
    0

    Default Finding a radius with smartprobe.ngc

    I just purchased a Tormach passive probe and have been wondering about using smartprobe.ngc to find the radius of a shaft or in my case a rifle receiver to make a scope base. How would I confugure the grid? It needs to only have one plotted line for the CAD program to measure radius. No X move and a Y grid that starts -Y and finishes +Y? The Z move would have to be? Here is the configuration portion of smartprobe.ngc:
    (Configuration section)
    G20 (Inches)
    F20 (probe speed)

    #1=-1 (X start) would it be X0?
    #2=.04 (X increment) would it be 0?
    #3=51 (X count) 1?

    #4=-1 (Y start) Would it be off center by some amount?
    #5=.04 (Y increment)
    #6=51 (Y count) Distance/.04?

    #7=1.1 (Z safety)
    #8=-.1 (Z probe) would it be a distance to ensure contact at the -Y first probe?
    (End configuration section)

    Similar Threads:


  2. #2
    Member
    Join Date
    Mar 2006
    Location
    USA
    Posts
    51
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    I have had success with the smartprobe file. It probes like I thought but will not put usable data into the directory I created. It really has potential as a simple way for Pathpilot to probe 3D. The programers at tormach are not too keen on folks getting into the file system but I was able to at least send an empty txt to my directory PROBE DATA.



  3. #3
    Registered
    Join Date
    Sep 2008
    Location
    usa
    Posts
    325
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    Unless you can manage to take measurements at (3) 90 degree positions around the radius I don’t believe your measurement will be accurate. Otherwise you will be measuring at a tangency point between the probe tip radius and receiver radius and you will have to try to calculate the receiver radius based on that. Also, unless you are deflecting the probe tip either directly vertically or horizontally I’m not sure if the trip point of the probe will be the same. The smaller the area of the receiver radius you are measuring the greater the error due to measuring innacuracy!



  4. #4
    Member
    Join Date
    Mar 2006
    Location
    USA
    Posts
    51
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    How then does a point cloud work to digitize a 3d object? I used to use a Renishaw probe on a Fadal VMC. I could probe with the same dia tip as my ball mill and make very good copies of 3d objects. Is it that the Renishaw program would measure tip deflection?



  5. #5
    Registered
    Join Date
    Sep 2008
    Location
    usa
    Posts
    325
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    Quote Originally Posted by choppero View Post
    How then does a point cloud work to digitize a 3d object? I used to use a Renishaw probe on a Fadal VMC. I could probe with the same dia tip as my ball mill and make very good copies of 3d objects. Is it that the Renishaw program would measure tip deflection?
    Unless a sharp point is being used for a probe tip, as the probe starts to measure a vertically curved surface, the probe will not contact along the vertical axis of the probe. Therefore the location that is being measured is in a different xy location than the vertical centerline of the probe. Because of that something has to be looking at the slope of the curve and calculating where that tangency point is. If the VMC is using the same probe diameter as the ball endmill the calculation for the curved surface was already been done in the CAM software and one is just looking at the deviation from the toolpath location and the actual. One way of compensating is to ignore the probe tip diameter in the measurements, bring the point cloud into software that will create a surface and then offset that surface for the probe tip radius. I haven't used such point cloud software but I would imagine that it can do that in one operation. I am familiar with all of this because we had a CMM where I worked that was purchased years before the capability of having an automated method of measuring a 3 dimensional surface existed. The machine had a programming language built in so my task was to create a program that would take measurements on a grid automatically and then perform those calculations.



  6. #6
    Registered
    Join Date
    Sep 2008
    Location
    usa
    Posts
    325
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    In addition to the shift from the vertical axis of the probe when measuring a 3D surface there is another factor that plays into this. That is that there is usually a small amount of deflection of the probe tip before the switch contact in the probe actually generates the electrical signal The accuracy and repeatability of that deflection distance in all directions is what makes good quality measuring probes so expensive. Low quality or homemade probes with rudimentary contacts will not have that accuracy and repeatability.



  7. #7
    Member
    Join Date
    Mar 2006
    Location
    USA
    Posts
    51
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    Check my theory out please. Tormach Pathpilot can only probe X and Y from a known Z so finding the center of a boss or rectangle is a task that uses the tip dia offset in calculating the point of trip. I'll try probing Y at a known Z distance from top of a partial radius manually and record that; then again .005 up and so on. maybe that can give me some points to enter into my CAD program. I'll use a shaft with a known diameter, so if the CAD points equal the known radius at least I can have a way to measure.



  8. #8
    Registered
    Join Date
    Sep 2008
    Location
    usa
    Posts
    325
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    That sounds like a great plan choppero. I think the thing that complicates things is that in Pathpilot the cutter diameter (in this case the probe tip diameter) is used for XY offsets along each axis for the probe tip. But the probe tip diameter in the Z direction is built into the tool length value for the probe. So if you were to bring the recorded values into the CAD and offset the curve for the probe tip diameter you should have to reduce the tool length by the probe tip radius while making the probe tip diameter zero. Then the measurements would appear to be originating from the exact center of the probe tip sphere and the offset of the curve in CAD should work out correctly.

    That is my view of things based on the assumption that the probing program is only automating the movement of the probe through the grid. In addition, I have no idea if any calculations are being done in Pathpilot to account for the true Z value at the point of tangency between the probe tip and the curve that is being measured. I doubt it very much but you never know. It will certainly be interesting to hear what you come up with.



  9. #9
    Member
    Join Date
    Aug 2010
    Location
    US
    Posts
    130
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    If you are probing in the z direction you would want to measure across the x or y direction (depending on how you orient it on the machine) and can do that without the double loop in the code by just deleting it. You can just position the probe along the x axis where you want to measure and let it move in the chosen direction (the Y direction for instance). You would want to probe closest to the highest point where the z change in the surface is the smallest for a given y change. That would give the least error due to the probe tip diameter and would be measuring the actual location where the scope base would sit.

    I did some calculations of that error as an example. If the probe tip is .25" diameter and the receiver top or tube you are measuring is 2" diameter, the error in the z direction is about .005" when you measure about .3" from the centerline. That goes to .017" at .5" from the centerline, but is .052" at .75" from the centerline. For a scope base that is typically less than an inch in width, the error is not huge for a surface with a relatively large radius. You could probably probe the profile, then estimate the radius from the probe data, then apply a z correction for that size receiver.

    If the top radius (diameter) of the receiver or tube being measured is much larger, like 10" diameter, the errors will be much less as you will be measuring a section only about an inch wide, and that will be relatively "flat". For that 10" diameter tube, at .3" from the centerline, the z error is only .0002", at .5" from the CL it is .0006", and at .75" from the CL the z error is just then .001".

    I'm glad you brought this up as it made me think about it again, and get an idea of how much error there is, and how the probe tip diameter can be accounted for in measurements. Testing would help verify this plus any of the other errors that were mentioned. When I get back to the shop I will test my theory and correction method.

    Did you have a particular receiver in mind? I would digitize one on my end if I have one and we could compare notes.



  10. #10
    Member
    Join Date
    Aug 2010
    Location
    US
    Posts
    130
    Downloads
    0
    Uploads
    0

    Default Re: Finding a radius with smartprobe.ngc

    I used gridprobe and adjusted it to only do one iteration in the x direction, and to measure in the positive y direction to check the curvature of a 2.257" diameter piece of aluminum hollow bar. I did not turn the section of the bar to be as round as possible as this was only to see if a pure calculation could be added to a prove value to get a more accurate position of the actual surface based on the curvature of the surface being measured and the probe ball diameter.

    I probed in the y direction in .050" increments from 0 to .95 inches. When I added my calculated error for the probe diameter to the actual probe measurements, I ended up with a maximum error (at the position Y=.95 inches where the curvature is the "steepest") of .0011". The point on the round tube should have been at z= -.285" but the probe tripped at z=-.2501. My calculated error to add to that point was .0344" which when added to the probe trip point estimated the z position at -.284". The difference with the added digit was the .0011".

    At y=.5" the error with the calculated correction added was .0003". The z point should have been at -.1168. The probe tripped at -.1043, and the calculated adder was .0121 which when added estimated the point at z=.1164 which is only off by .0003 (with all digits in the calculations, but here it rounds to .0004").

    At y=.3" the error was .0000". The z point should have been at -.0406. The probe tripped at -.0367. The calculated adder was .0039 for an estimated point at -.0406".

    So, it looks like this method will work well, especially when the surfaces are not approaching vertical. It would definitely work for a scope base.

    If anyone is interested I can post more detailed results and further explain what I did.



Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

Finding a radius with smartprobe.ngc

Finding a radius with smartprobe.ngc