Are there CMM macros available for VMC's? - Page 3
Close
Login to Your Account
Page 3 of 3 FirstFirst 123
Results 41 to 54 of 54
  1. #41
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    1,501
    Post Thanks / Like
    Likes (Given)
    2347
    Likes (Received)
    823

    Default

    Quote Originally Posted by snowshooze View Post
    It is not.
    you're an idiot.

  2. #42
    Join Date
    Nov 2015
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    608
    Post Thanks / Like
    Likes (Given)
    144
    Likes (Received)
    370

    Default

    Quote Originally Posted by empwoer View Post
    then go do something about it...

    you're surprised that nobody has come up with a way to do something (idiotic i might add) when much better ways exist to do this?

    and go get fucked with your snarky shit. i was trying to help you, fucking douchebag
    Ya this guy seems like the engineer type who I usually say to "yes you can, but why the hell would you? Here are 10 established and repeatable solutions to the problem you have created for the sole purpose of solving it". Then the engineer tells me to get fucked and does it their own way and screws it up. Only then do they consider that we might have some pretty good ideas!

    or

    "You can, but why would you? Here are 10 alternatives to the problem you have created just for the sole purpose of saying you solved it"

    In both cases usually the problem they are trying to solve is not a problem. "it is not a bug, it is a feature" usually applies. Trying to do this is a major time sink unless there is some ulterior motive that we don't know about, for example you are doing research and development for a machine tool builder to try and improve the probing cycles. This is not likely the case. Hence, waste of time (imo)

  3. Likes empwoer liked this post
  4. #43
    Join Date
    Aug 2006
    Location
    Wisconsin
    Posts
    1,790
    Post Thanks / Like
    Likes (Given)
    812
    Likes (Received)
    846

    Default

    I'll start with this, I use Powermill and probe my finished parts before they come off the machine, then I import that data into Fusion 360 and read the results, and that's as far as we go with on machine probing. When we need what you are asking, we just digitize it with a scanner to make a point cloud, then you get a triangle file that needs allot of post model smoothing to say the least. Doing that on the actual cnc machine I think would be a nightmare unless your doing very simple stuff. After all the cnc machine will only go where you ask it to, and if you already know where that is, it sort of defeats the purpose of what your trying to do.

  5. Likes empwoer, mhajicek liked this post
  6. #44
    Join Date
    Jan 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    888
    Post Thanks / Like
    Likes (Given)
    130
    Likes (Received)
    462

    Default

    So tracing like this, but 3d?




    It isn't too complicated. This one just uses a circular logic. Move Y-, if skip signal found, move X+

    If no skip signal, move Y- again.

    etc.

    If no Y- signal, move X-... etc, etc. Just moving CW around a part.

    3D wouldn't be much different, but you'd need to figure in some kind of safety so you don't drop into a hole and keep going until the probe body is squashed.

    And you'd have to figure a grid size to run, where this one just moves in a circle and is stopped manually.

  7. #45
    Join Date
    May 2017
    Country
    UNITED STATES
    State/Province
    Minnesota
    Posts
    2,311
    Post Thanks / Like
    Likes (Given)
    3197
    Likes (Received)
    1680

    Default

    Quote Originally Posted by thesidetalker View Post
    3D wouldn't be much different, but you'd need to figure in some kind of safety so you don't drop into a hole and keep going until the probe body is squashed.

    And you'd have to figure a grid size to run, where this one just moves in a circle and is stopped manually.
    That's the grid thing we were talking about, that's been around 30+ years. It's a very limited approach. It works for scanning a model car body or boat hull, but takes a very long time, and has geometry limitations. No undercuts, Z step limits, points only on your grid, etc.

  8. Likes gregormarwick liked this post
  9. #46
    Join Date
    Feb 2007
    Location
    Aberdeen, UK
    Posts
    4,811
    Post Thanks / Like
    Likes (Given)
    1785
    Likes (Received)
    2371

    Default

    Quote Originally Posted by mhajicek View Post
    That's the grid thing we were talking about, that's been around 30+ years. It's a very limited approach. It works for scanning a model car body or boat hull, but takes a very long time, and has geometry limitations. No undercuts, Z step limits, points only on your grid, etc.
    AKA raster pattern probing.

  10. Likes mhajicek liked this post
  11. #47
    Join Date
    Jan 2007
    Location
    Flushing/Flint, Michigan
    Posts
    11,124
    Post Thanks / Like
    Likes (Given)
    668
    Likes (Received)
    8893

    Default

    Quote Originally Posted by thesidetalker View Post
    So tracing like this, but 3d?
    It isn't too complicated. This one just uses a circular logic. Move Y-, if skip signal found, move X+
    This neat but..... What to do with all that data?
    Here is where cmms computers are very different from a cnc machine tool control.
    A 3-5 axis mill can act like a cmm for somethings.
    Yes it will not be as square or true but as here able to do it at a lower accuracy. Both "know" where they are at all times.
    Most built in measuring of features demand that the orient be known. Cmms don't care and the part can and often is sitting at bit a angle to the world.
    Squeezing even the basics of a early 80's 3d DNC cmm into a OM control would be quite a challenge. Transform matrices are not an exposed function to macros even though done inside with ease.

    There are many ways to trace a contour with a point contact probe. This one sort of brute force but if working it works.
    Knowing the previous points and new directions is nice but putting that into a cnc macro sort of a lot of work and if a 3d thing like a intake port... Oh my.
    Laser or optical 3D scanners absolutely do not work in these holes either.
    Clamped once however in space I see no way any 3 axis cmm or other can do a head port no matter the probe tip used.
    5 axis is needed to check. But 5 axis is also needed to do the work.

    Easier.. pour rubber into the hole/feature of good. Measure that optically on a compator or microscope with defined rotation. Cross sections.
    I do get what the op is after and good lord it should be easier nowadays.
    Bob

  12. #48
    Join Date
    May 2002
    Location
    South Central PA
    Posts
    14,669
    Post Thanks / Like
    Likes (Given)
    2874
    Likes (Received)
    4401

    Default

    Quote Originally Posted by CarbideBob View Post
    I do get what the op is after and good lord it should be easier nowadays.
    Bob
    Centroid advertises an all-in-one package, digitize and cut on the same machine.

    A560 5-Axis CNC Articulating-Head Porting Machine - high-speed cylinder head porting, digitize and port intake and exhaust ports and combustion chambers...

  13. #49
    Join Date
    Feb 2007
    Location
    Aberdeen, UK
    Posts
    4,811
    Post Thanks / Like
    Likes (Given)
    1785
    Likes (Received)
    2371

    Default

    Quote Originally Posted by CarbideBob View Post
    This neat but..... What to do with all that data?
    Here is where cmms computers are very different from a cnc machine tool control.
    A 3-5 axis mill can act like a cmm for somethings.
    Yes it will not be as square or true but as here able to do it at a lower accuracy. Both "know" where they are at all times.
    Most built in measuring of features demand that the orient be known. Cmms don't care and the part can and often is sitting at bit a angle to the world.
    Squeezing even the basics of a early 80's 3d DNC cmm into a OM control would be quite a challenge. Transform matrices are not an exposed function to macros even though done inside with ease.

    There are many ways to trace a contour with a point contact probe. This one sort of brute force but if working it works.
    Knowing the previous points and new directions is nice but putting that into a cnc macro sort of a lot of work and if a 3d thing like a intake port... Oh my.
    Laser or optical 3D scanners absolutely do not work in these holes either.
    Clamped once however in space I see no way any 3 axis cmm or other can do a head port no matter the probe tip used.
    5 axis is needed to check. But 5 axis is also needed to do the work.

    Easier.. pour rubber into the hole/feature of good. Measure that optically on a compator or microscope with defined rotation. Cross sections.
    I do get what the op is after and good lord it should be easier nowadays.
    Bob
    Which is why I recommended PowerInspect in the very first reply to this thread.

    It is CMM software that also supports posting gcode for probing cycles on CNC machines. It is the only CMM software package that can do so AFAIK.

  14. #50
    Join Date
    Jan 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    888
    Post Thanks / Like
    Likes (Given)
    130
    Likes (Received)
    462

    Default

    Quote Originally Posted by gregormarwick View Post
    Which is why I recommended PowerInspect in the very first reply to this thread.

    It is CMM software that also supports posting gcode for probing cycles on CNC machines. It is the only CMM software package that can do so AFAIK.
    But don't you need a model that's fairly close to be able program the data collection points? Or just set an area and it will figure it out for you? It seems the OP is after 'scanning' with his probe to establish a model or point cloud of some kind.


    Seems pointless on a CNC and a waste of time to scan anything like this. FWIW I never used the program I made. I suppose it was more of an exercise for fun.


    OP should just pay someone to scan whatever needs scanning. Time and money ahead when a 3D laser scanner can take millions of points in a relatively short amount of time. We have that on our Faro arm for comparing, mostly used on sheetmetal stampings. We didn't bother to go all out with the software though, so not really for manipulating things.

    Although Bob has a point about scanning holes or ports...

  15. #51
    Join Date
    Oct 2009
    Location
    Oregon
    Posts
    6,625
    Post Thanks / Like
    Likes (Given)
    7061
    Likes (Received)
    3937

    Default

    OP-

    IMO, your whole idea is pretty dumb. You just want the port geometry right? You can model it up and generate the toolpaths right?

    So you take a head that's ported the way you want. You set it up in the same fixture you'd use to machine it and gather all the point data you need to draw the mfer up.

    It doesn't take that long. I've done thousands of automotive parts, many with complex geometry. I do it with a VMC, radius gauges and mostly use a dowel pin and a piece of paper to establish contact. Sounds pretty archaic when I write it, but with the MPG in one hand, paper in the other and a notebook and pencil to plot points I go to town. On more complex stuff I walk the 15 feet to the cam computer and plot the points as I take them so I don't lose track.

    I'd bet an hour per port per fixture position. I'm sure you'd need 3 positions per port to reach at all on a 4 axis VMC. That really isn't that much time.

    I did a project working with an engineer at a big OEM awhile back. I used my method and he used his $200k scanning bullshit machine thing to come up with a point cloud. I was done in a couple hours and all my numbers made sense. His point cloud was pretty worthless to me and the model he made from it was not much better. His shit averaged a bunch of points that it shouldn't have so most locations were wrong.

    I'm not really concerned about how long it takes to manually reverse engineer geometry because I'm also thinking about machining strategy and getting a thorough dimensional understanding of what I'm working with.

  16. Likes empwoer liked this post
  17. #52
    Join Date
    Feb 2007
    Location
    Aberdeen, UK
    Posts
    4,811
    Post Thanks / Like
    Likes (Given)
    1785
    Likes (Received)
    2371

    Default

    Quote Originally Posted by thesidetalker View Post
    But don't you need a model that's fairly close to be able program the data collection points? Or just set an area and it will figure it out for you? It seems the OP is after 'scanning' with his probe to establish a model or point cloud of some kind.


    Seems pointless on a CNC and a waste of time to scan anything like this. FWIW I never used the program I made. I suppose it was more of an exercise for fun.


    OP should just pay someone to scan whatever needs scanning. Time and money ahead when a 3D laser scanner can take millions of points in a relatively short amount of time. We have that on our Faro arm for comparing, mostly used on sheetmetal stampings. We didn't bother to go all out with the software though, so not really for manipulating things.

    Although Bob has a point about scanning holes or ports...
    Whether you need a model depends on what you are doing. PI works just like any other CMM software. For OMV (On Machine Verification) you require a model.

    I think you can set a touch probe raster grid arbitrarily without a model, but I have never done that.

    FWIW PI has always been one of the foremost names in laser and white light scanning and it has comprehensive support for point clouds. Perhaps that is mainly in this country though, being British developed (originally Delcam before the Autodesk buyout) - I think that it is not widely used in the US. Faro used to offer PI as the high end software option when you bought a new arm, not sure if that is still the case, and again may have only been in this country.

    Wholeheartedly agree that it's not a good use of machine time, but if OP wants to use his machine like that there is no reason why he shouldn't do so.

    As long as he understands why he shouldn't be using his machine to qualify parts, as opposed to just reverse engineering them...

  18. #53
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    1,501
    Post Thanks / Like
    Likes (Given)
    2347
    Likes (Received)
    823

    Default

    Quote Originally Posted by gregormarwick View Post
    Whether you need a model depends on what you are doing. PI works just like any other CMM software. For OMV (On Machine Verification) you require a model.

    I think you can set a touch probe raster grid arbitrarily without a model, but I have never done that.

    FWIW PI has always been one of the foremost names in laser and white light scanning and it has comprehensive support for point clouds. Perhaps that is mainly in this country though, being British developed (originally Delcam before the Autodesk buyout) - I think that it is not widely used in the US. Faro used to offer PI as the high end software option when you bought a new arm, not sure if that is still the case, and again may have only been in this country.

    Wholeheartedly agree that it's not a good use of machine time, but if OP wants to use his machine like that there is no reason why he shouldn't do so.

    As long as he understands why he shouldn't be using his machine to qualify parts, as opposed to just reverse engineering them...
    why the distinction between qualifying and reverse engineering?
    if the fear is that there are inaccuracies in the machine, that would affect the model one reverse engineers, which then wont qualify correct, no?
    the way i see it, you're either comfortable with your machine's accuracy to trust it with probing/reverse engineering, or not.
    what am i missing?

  19. #54
    Join Date
    Feb 2007
    Location
    Aberdeen, UK
    Posts
    4,811
    Post Thanks / Like
    Likes (Given)
    1785
    Likes (Received)
    2371

    Default

    Quote Originally Posted by empwoer View Post
    why the distinction between qualifying and reverse engineering?
    if the fear is that there are inaccuracies in the machine, that would affect the model one reverse engineers, which then wont qualify correct, no?
    the way i see it, you're either comfortable with your machine's accuracy to trust it with probing/reverse engineering, or not.
    what am i missing?
    You're obviously not wrong technically, but there is a a critical distinction.

    To qualify a part is to prove that it is correct to a print or a model, and you are rarely if ever proving this to yourself.

    When you reverse engineer a part there is rarely a requirement (or obligation) to prove it is correct to anything, beyond that it works properly, or looks good.

  20. Likes mhajicek liked this post

Tags for this Thread

Bookmarks

Posting Permissions

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