The Rumors are True New HAAS UMC 500 ! - Page 3
Close
Login to Your Account
Page 3 of 17 FirstFirst 1234513 ... LastLast
Results 41 to 60 of 335
  1. #41
    Join Date
    May 2017
    Country
    UNITED STATES
    State/Province
    Minnesota
    Posts
    940
    Post Thanks / Like
    Likes (Given)
    1126
    Likes (Received)
    576

    Default

    You can get to four sides. 45°, 135°, 225°, and 315°.

  2. Likes Mike1974, cameraman liked this post
  3. #42
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by empwoer View Post
    i dont see how...
    think about it, imagine a straight shaft with a ball on top, then rotate your tilt axis 90*, you need to probe the tool from 4 sides, you can get to 3 sides, 4th will have the shaft in the way.
    To be fair, it is necked down quite a bit, so maybe some fancy calculations of centerline of ball based on +.1 / +.2 / +.3 above center...??

  4. #43
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by mhajicek View Post
    You can get to four sides. 45°, 135°, 225°, and 315°.
    Ya, you are right. I think machinist "in general" (or maybe just me) typically think polar- 0/90/180/270 Until we need to "alter" that thinking for a project.

  5. #44
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by empwoer View Post
    i dont see how...
    think about it, imagine a straight shaft with a ball on top, then rotate your tilt axis 90*, you need to probe the tool from 4 sides, you can get to 3 sides, 4th will have the shaft in the way.
    I understand what you are saying, and a lot of the tooling balls used for kinematic calibration use an angled shaft … (like Mazackeck for an i-500 (variaxis) )

    Some are straight and some are angled so I guess 5 points (vertically) are probed rather than 6. and tilted over 90 degrees as you say 4 points can only be probed rather than 5 or 6.

    I'm guessing to calibrate the B axis knuckle (on the UMC layout) then 4 points locates the ball pretty well as the most important measurement there is the height / Z position of the top of the ball to map out / calibrate the angles on the B axis. + find the center line with the other measured positions while tilted.

    HAAS's calibration cycle seems different from other MTBs.


    _________________________________

    I just saw mahjicec's answer , that's a good one / great thinking ! Maybe slightly beyond HAAS's imagination ? lol / maybe not :-)

    __________________________________________________ ______________________________________________




    HAAS MRZP cycle with older gizmo... So maybe they don't tilt the trunnion/ knuckle all the way over ?

    I should look at the end of the other HAAS service video (posted earlier) shows it too.

    OK so it tilts the B axis - 30° , -15°, + 15° and +30°
    (the shaft wont interfere ? right ? Straight or crooked ? ).

    And does the C axis rotary table at 0°, 90°, 180° and 270° ?

  6. #45
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by cameraman View Post
    I understand what you are saying, and a lot of the tooling balls used for kinematic calibration use an angled shaft … (like Mazackeck for an i-500 (variaxis) )

    Some are straight and some are angled so I guess 5 points (vertically) are probed rather than 6. and tilted over 90 degrees as you say 4 points can only be probed rather than 5 or 6.

    I'm guessing to calibrate the B axis knuckle (on the UMC layout) then 4 points locates the ball pretty well as the most important measurement there is the height / Z position of the top of the ball to map out / calibrate the angles on the B axis. + find the center line with the other measured positions while tilted.

    HAAS's calibration cycle seems different from other MTBs.


    _________________________________

    I just saw mahjicec's answer , that's a good one / great thinking.

    __________________________________________________ ______________________________________________



    HAAS MRZP cycle with older gizmo... So maybe they don't tilt the trunnion/ knuckle all the way over ?

    I should look at the end of the other HAAS service video (posted earlier) shows it too.
    That is odd positioning of the cal ball, not like I was "taught", but apparently it works ok? One thing on that cycle it does not update anything but the 3 macro vars. You have to then manually enter those into your parameter page (in case anyone is wondering). Not sure if different than the other models of UMC or if they have updated the control in some way....

  7. #46
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by Mike1974 View Post
    That is odd positioning of the cal ball, not like I was "taught", but apparently it works ok? One thing on that cycle it does not update anything but the 3 macro vars. You have to then manually enter those into your parameter page (in case anyone is wondering). Not sure if different than the other models of UMC or if they have updated the control in some way....


    Yup I agree in that earlier video the tooling ball position seems counter intuitive but good enough ? They seem to set the tooling ball at 45 degrees angles on the 'Clock face" of the C axis table to the cardinal points and positions of the X and Y axes.

    If you scroll to the last 4 minutes or so of the video they show exactly what you are describing with these 3 parameters.

    And to make sure and double and triple check they are entered in correctly ?

    "If you suspect you may have entered in these values incorrectly ??? " I don't suspect anything, I just enter in completely wrong values without knowing it lol At least they do stress that very small errors in those values can really mess everything up (in a big way).

    At least the HAAS videos are pretty clear, even the more service related ones.

    The three parameters you mention point to their idea of true single intersection point.

  8. #47
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by empwoer View Post
    i dont see how...
    think about it, imagine a straight shaft with a ball on top, then rotate your tilt axis 90*, you need to probe the tool from 4 sides, you can get to 3 sides, 4th will have the shaft in the way.


    This is more along the lines you are thinking of perhaps with the likes of the Renishaw Axiset.

    This video is very stylized but towards the end shows the more critical plots and also the strategies used will / can / should vary from machine to machine (type , layout and critical accuracies and available precisions) .

    Dennis of dstryr fame uses Axiset quite a lot for his machines (as I understand it)

    I think HAAS is a law unto itself , HAAS does HAAS (pretty well)

    __________________________________________________ ________________________

    I'm guessing the Fanuc/Matsuura quick calibration is more along the Axiset lines than HAAS (I imagine your linear machine is super accurate so wonder what procedures are used when the errors and residual errors are so low ?

  9. Likes empwoer, PegroProX440 liked this post
  10. #48
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by cameraman View Post


    This is more along the lines you are thinking of perhaps with the likes of the Renishaw Axiset.

    This video is very stylized but towards the end shows the more critical plots and also the strategies used will / can / should vary from machine to machine (type , layout and critical accuracies and available precisions) .

    Dennis of dstryr fame uses Axiset quite a lot for his machines (as I understand it)

    I think HAAS is a law unto itself , HAAS does HAAS (pretty well)

    __________________________________________________ ________________________

    I'm guessing the Fanuc/Matsuura quick calibration is more along the Axiset lines than HAAS (I imagine your linear machine is super accurate so wonder what procedures are used when the errors and residual errors are so low ?
    Cool video. FWIW I did the mzrp (?) cycle multiple times when doing 4/5 axis work and the numbers were always less than (IIRC) .0003". It's actually a number like 123456 (12.3456"), but nonetheless... Maybe that is inaccuracies (ie just not computing enough of a difference to vary more than a few tenths?) in the cycle, maybe the best 'res' Haas could do, dunno...

  11. Likes empwoer liked this post
  12. #49
    Join Date
    Jan 2015
    Country
    UNITED STATES
    State/Province
    California
    Posts
    700
    Post Thanks / Like
    Likes (Given)
    103
    Likes (Received)
    363

    Default

    The angled tooling ball holder is so you can take a measurement at B90. There is a prompted option when running calibration to go to 90° or not. Or maybe they give the option of 90° since the holder is angled and you have clearance? What came first, chicken or egg?

    The built-in calibration cycle gets you close, but I've found on my machine I always needs small adjustments afterward to dial in more closely. I use an indicator(s in multiple directions) and TCPC to follow the ball around 360° in XY, then move in XZ to see how far off I am.

    Movement speed matters too of course, if you go too fast it doesn't follow very closely (and you can see the indicator needle bouncing wildly) and actually I have my post setup to enforce a "speed limit" of sorts on the rotary speed while simultaneous cutting, based on dist from center point. Further out it goes slower. Makes for better finishes in my opinion.


    Haven't looked at the Haas calibration program myself, but I've found you can get just as close with a 3-point circle calculation to find the center points. Probably what they're doing I'd imagine.

    and btw, the X-center of both rotaries are different. What's interesting is the control always accounts for this error in setting 254, whether you are using DWO/TCPC or not. I believe because technically you could still program this machine from centerline and not have to use DWO/TCPC at all, but you would still want this mismatch in rotary x-centers accounted for. The trunnion isn't machined perfectly in the factory and aligned dead nuts, so that's where you make that small correction. (I think mine is at 8 or 9 tenths)

    What's stupid is there's about zero documentation for which way you should adjust it, positive or negative. Just gotta punch in values and see what's what I guess. I don't think the MRZP calibration cycle even tells you? That's why I made my own, but even then I got goofed up results since the control is always accounting for this error. Kinda throws you for a loop like thinking about how time travel works if you kill your father.


    Should be interesting to see how much more accurate the newer and smaller UMC's are with the new rotary gear and scales. I just got a new EC400 and it has their new rotary gears + scales. To me, it seems much more accurate than the C-axis on my UMC. (maybe larger bearings help too? idk) When I setup the rotary center on my EC400, I measured hardly any runout in Y direction when rotating about a 12" circle with the tooling ball from my UMC. Maybe a tenth or so. On my UMC, it's like 5 to 7 tenths in Z, rotating the C axis a full 360° and moving in XY. Or maybe my machine is not as level as I think. Might need to look again.

  13. Likes empwoer, cameraman, Cycle1000 liked this post
  14. #50
    Join Date
    Sep 2018
    Country
    UNITED STATES
    State/Province
    California
    Posts
    265
    Post Thanks / Like
    Likes (Given)
    314
    Likes (Received)
    112

    Default

    Quote Originally Posted by cameraman View Post


    This is more along the lines you are thinking of perhaps with the likes of the Renishaw Axiset.

    This video is very stylized but towards the end shows the more critical plots and also the strategies used will / can / should vary from machine to machine (type , layout and critical accuracies and available precisions) .

    Dennis of dstryr fame uses Axiset quite a lot for his machines (as I understand it)

    I think HAAS is a law unto itself , HAAS does HAAS (pretty well)

    __________________________________________________ ________________________

    I'm guessing the Fanuc/Matsuura quick calibration is more along the Axiset lines than HAAS (I imagine your linear machine is super accurate so wonder what procedures are used when the errors and residual errors are so low ?
    i can see not going to 90/-90, but that essentially means you're extrapolating the errors you get at say 45 degrees out to 90? assuming the error scales linearly...

    either way i think i'd rather use the angled ball.

    as far as what 'thesidetalker' said, measuring 3 points around the ball could work, IF you knew where the shaft is, and could work around it. i'm not sure the software is that smart.

    cameraman: we actually just barely got the machine set up, leveled and ready to machine. a ton of other shit came up, along with setting up our yama seiki LYS 2000 lathe so its been a minute since we bought it. but i'm hoping next week i'll be able to start making some chips on the LX and see how it does! pretty stoked.

  15. Likes cameraman liked this post
  16. #51
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by thesidetalker View Post
    The angled tooling ball holder is so you can take a measurement at B90. There is a prompted option when running calibration to go to 90° or not. Or maybe they give the option of 90° since the holder is angled and you have clearance? What came first, chicken or egg?

    The built-in calibration cycle gets you close, but I've found on my machine I always needs small adjustments afterward to dial in more closely. I use an indicator(s in multiple directions) and TCPC to follow the ball around 360° in XY, then move in XZ to see how far off I am.

    Movement speed matters too of course, if you go too fast it doesn't follow very closely (and you can see the indicator needle bouncing wildly) and actually I have my post setup to enforce a "speed limit" of sorts on the rotary speed while simultaneous cutting, based on dist from center point. Further out it goes slower. Makes for better finishes in my opinion.


    Haven't looked at the Haas calibration program myself, but I've found you can get just as close with a 3-point circle calculation to find the center points. Probably what they're doing I'd imagine.

    and btw, the X-center of both rotaries are different. What's interesting is the control always accounts for this error in setting 254, whether you are using DWO/TCPC or not. I believe because technically you could still program this machine from centerline and not have to use DWO/TCPC at all, but you would still want this mismatch in rotary x-centers accounted for. The trunnion isn't machined perfectly in the factory and aligned dead nuts, so that's where you make that small correction. (I think mine is at 8 or 9 tenths)

    What's stupid is there's about zero documentation for which way you should adjust it, positive or negative. Just gotta punch in values and see what's what I guess. I don't think the MRZP calibration cycle even tells you? That's why I made my own, but even then I got goofed up results since the control is always accounting for this error. Kinda throws you for a loop like thinking about how time travel works if you kill your father.


    Should be interesting to see how much more accurate the newer and smaller UMC's are with the new rotary gear and scales. I just got a new EC400 and it has their new rotary gears + scales. To me, it seems much more accurate than the C-axis on my UMC. (maybe larger bearings help too? idk) When I setup the rotary center on my EC400, I measured hardly any runout in Y direction when rotating about a 12" circle with the tooling ball from my UMC. Maybe a tenth or so. On my UMC, it's like 5 to 7 tenths in Z, rotating the C axis a full 360° and moving in XY. Or maybe my machine is not as level as I think. Might need to look again.
    I'm not even sure... maybe it is too early in the morning...

    There is no "adjustment" You just punch in the numbers recorded in the macro variables (which correspond to xyz centerlines)....? Maybe I am confused on what you are trying to explain...

  17. #52
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by empwoer View Post
    i can see not going to 90/-90, but that essentially means you're extrapolating the errors you get at say 45 degrees out to 90? assuming the error scales linearly...

    either way i think i'd rather use the angled ball.

    as far as what 'thesidetalker' said, measuring 3 points around the ball could work, IF you knew where the shaft is, and could work around it. i'm not sure the software is that smart.

    cameraman: we actually just barely got the machine set up, leveled and ready to machine. a ton of other shit came up, along with setting up our yama seiki LYS 2000 lathe so its been a minute since we bought it. but i'm hoping next week i'll be able to start making some chips on the LX and see how it does! pretty stoked.
    I think what thesidetalker meant was that he could find his center of rotation for the C axis (rotary table) by measuring three different rotary positions for the ball. It's like an old school geometry exercise where you have a drawn circle (made with an old fashioned drawing compass) and you have to find the center of the circle. [You can't cheat by looking at where the compass made a pin prick in the paper.]. Normally by making various arcs you can construct two chords on a circle at moderate angles to each other and then bisect each chord and construct perpendicular lines to each chord by various construction arcs also. Basically it's the intersection of those two perpendicular lines from the two arbitrary chords that "finds" the center from the two chords made with the circle at an angle to each other. TheSidetalker might be computing this method in other ways using trig or cad or 'puter (or even a circular slide rule ;-) .

    So what the "Thesidetalker" is saying (basically) three measured points rotated around the c axis is the minimum solution to find the center axis of rotation. HAAS in their sequence of moves determines XYZ coordinates for the center of the ball that is moved into 4 positions. Theoretically that 4th point is redundant but can be used in various permutations with other constructible triangles to make a slightly more accurate determination of the intersection of the C axis with the XY plane. Personally there are other methods using more points but HAAS does not use those. I think they try to keep things as simple as humanly possible in most cases.

    I agree angled ball is the preferred kit.

    Very cool about all the new machines you got going on there … Nice !

    __________________________________________________ _____________________________________________


    * Might scrounge up a link for finding the center of circle using a drawing compass as I don't believe they teach that kind of stuff in schools anymore. [Just for a lark.].

    YouTube

    ^^^Theory + some math for finding circle centers / centers of rotation.

    YouTube

    ^^^ Using a compass and ruler to find the center of a circle << This is not me >> but I learnt the same method as a 9 or ten year old. It has a parallel with the method used by Rennishaw (in the Renishaw axiset video posetd before) for axiset to find the center of rotation of a an axis (but there are much better mathematical techniques also that can be used that are probably in line with mathematical techniques used by a Zeiss CMM's etc. for finding centers of things from redundant sampled points.). + three measured points minimal solution that thesidetalker mentioned.

  18. #53
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by thesidetalker View Post
    The angled tooling ball holder is so you can take a measurement at B90. There is a prompted option when running calibration to go to 90° or not. Or maybe they give the option of 90° since the holder is angled and you have clearance? What came first, chicken or egg?

    The built-in calibration cycle gets you close, but I've found on my machine I always needs small adjustments afterward to dial in more closely. I use an indicator(s in multiple directions) and TCPC to follow the ball around 360° in XY, then move in XZ to see how far off I am.

    Movement speed matters too of course, if you go too fast it doesn't follow very closely (and you can see the indicator needle bouncing wildly) and actually I have my post setup to enforce a "speed limit" of sorts on the rotary speed while simultaneous cutting, based on dist from center point. Further out it goes slower. Makes for better finishes in my opinion.


    Haven't looked at the Haas calibration program myself, but I've found you can get just as close with a 3-point circle calculation to find the center points. Probably what they're doing I'd imagine.

    and btw, the X-center of both rotaries are different. What's interesting is the control always accounts for this error in setting 254, whether you are using DWO/TCPC or not. I believe because technically you could still program this machine from centerline and not have to use DWO/TCPC at all, but you would still want this mismatch in rotary x-centers accounted for. The trunnion isn't machined perfectly in the factory and aligned dead nuts, so that's where you make that small correction. (I think mine is at 8 or 9 tenths)

    What's stupid is there's about zero documentation for which way you should adjust it, positive or negative. Just gotta punch in values and see what's what I guess. I don't think the MRZP calibration cycle even tells you? That's why I made my own, but even then I got goofed up results since the control is always accounting for this error. Kinda throws you for a loop like thinking about how time travel works if you kill your father.


    Should be interesting to see how much more accurate the newer and smaller UMC's are with the new rotary gear and scales. I just got a new EC400 and it has their new rotary gears + scales. To me, it seems much more accurate than the C-axis on my UMC. (maybe larger bearings help too? idk) When I setup the rotary center on my EC400, I measured hardly any runout in Y direction when rotating about a 12" circle with the tooling ball from my UMC. Maybe a tenth or so. On my UMC, it's like 5 to 7 tenths in Z, rotating the C axis a full 360° and moving in XY. Or maybe my machine is not as level as I think. Might need to look again.

    The New UMC 1000 with the cyclonic bearing and gearing is +/- 40 arc seconds accuracy and +/- 20 arc seconds repeatability.

    At least its nearly bombproof and can sustain substantial "Crashing". Same design intent with the UMC-500 (so they seem to state).

    The UMC 500 vid with the cyclonic gearing states 1 arc minute for backlash. Looking at the other HAAS rotaries and how they compensate for backlash / slop in the mechanism 1 arc minute backlash ---> +/- 30 arc seconds accuracy ---> +/- 15 arc seconds repeatability. (probably somewhere within 20 arc seconds with scales for one rotation to the next.). That can translate to a +/- 10 micron , +/- 0.01 mm , +/- 0.0004" for a 200 mm wide part... (which is not terrible); for a smaller part like 100 mm in diameter +/- 0.0002" for angular errors. Bearing in mind that's only one rotary axis but sometimes it IS possible to get away with murder depending on the function of the part and sequence of rotations and number and types of orthogonal references. never the less not really a true representation of volumetric accuracy and compounding of positional errors from two rotary axes etc. + all the other errors we are usually too scared to talk about (lol).

    Haas is interesting how they make successive and incremental improvements to their product lines. Nothing miraculous (in this case) other than the product intent and problems that it solves (for a lot of people) through it's price point.

    So far in their bumpf they are stressing greater rigidity, so could be better surface finishes, better accuracies etc.

    It will be interesting to see how close the spindle nose can get to the center of the table with the trunnion/ knuckle tilted over at 90 degrees. (not sure if they have changed the spindle head's casting / design a bit ?). Key point being how far off the table do certain parts have to be fixtured ?

    Still not sure how much on control thermal regulation that HAAS implements ? (I.e. temperature sensors in the machine and the control makes corrections ?.) Some sort of thermal regulation is important if one is doing the automation "Thing" / pallet pool and trying to stay within reasonable tolerances running for extended periods.

    @Thesidetalker if the UMC -500 has good sensitivity of movement i.e. can move through very small incremental corrective movements and positions (determined by probing or using dial test indicators .) then with a bit of skill (hands on) perhaps some really pretty good results can be achieved (without any temporal / time travel paradoxes) to fix anything. Hopefully the machine is somewhat less dependent on its foundation and installation that its bigger brothers and sisters.

  19. #54
    Join Date
    Jan 2007
    Location
    Flushing/Flint, Michigan
    Posts
    7,612
    Post Thanks / Like
    Likes (Given)
    379
    Likes (Received)
    6359

    Default

    What are people fixated on the points taken being at 90 degrees?
    Plus here you are trying to define the center of a sphere not a circle which require 4 points.
    Ideally 3 spread around the OD as much as possible and one up top but you can use any 4 with a loss of accuracy.
    Four around the OD at the same height would be bad unless you add a fifth.
    When tracking an axis system like this for alignment or correction values it is the sphere center that counts not a circle..
    You could do it with half a ball or even less hanging out there.
    There point circle is an easy solve, four point sphere not much worse.
    Now a 10 point circle or sphere and the math gets harder as you now assume each point has an error and you have minimize that which is fancy math with a couple of approaches used.
    Ideally you want those ten points spread all over the place in x,y and z and none on the same machine axis plane.

    A cnc machine tool has all the computing power of any cmm so thinking manual machine edge finder is all wrong.
    Bob

  20. Likes sfriedberg liked this post
  21. #55
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by CarbideBob View Post
    What are people fixated on the points taken being at 90 degrees?
    Plus here you are trying to define the center of a sphere not a circle which require 4 points.
    Ideally 3 spread around the OD as much as possible and one up top but you can use any 4 with a loss of accuracy.
    Four around the OD at the same height would be bad unless you add a fifth.
    When tracking an axis system like this for alignment or correction values it is the sphere center that counts not a circle..
    You could do it with half a ball or even less hanging out there.
    There point circle is an easy solve, four point sphere not much worse.
    Now a 10 point circle or sphere and the math gets harder as you now assume each point has an error and you have minimize that which is fancy math with a couple of approaches used.
    Ideally you want those ten points spread all over the place in x,y and z and none on the same machine axis plane.

    A cnc machine tool has all the computing power of any cmm so thinking manual machine edge finder is all wrong.
    Bob
    Not sure what you are carrying on about "points at 90 deg"? The conversation was more about having the ball on an offset stem...

    YouTube

    And if you watch it takes 5 points at every position....

  22. #56
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by Mike1974 View Post
    Not sure what you are carrying on about "points at 90 deg"? The conversation was more about having the ball on an offset stem...

    YouTube

    And if you watch it takes 5 points at every position....


    ^^^ 4 position method technique shown here with a Kunkle style 5 axis mori, they seem not to need to tilt on the B axis at 90 degrees either...

    "The Kinematic module in NC Gage lets users quickly detect and correct errors in 4- or 5-axis motion. Automatically tune the center-of-rotation and angular deviations. For use with Hexagon, Renishaw, Blum or Marposs machine tool probes on Fanuc and MAPPS controllers only. See Machine Tool Measurement | Hexagon Manufacturing Intelligence for more information." *




    TCPC "Party trick" , that you have probably seen at trade shows as well. [I think Okuma does one with a curved probing stem and DTI spinning slowly around the ball while the table moves on two rotations simultaneously. ].


    Then re-read what the Thesidetalker wrote in post #49 ~ (not aimed at anyone but at least one might be able to disambiguate points on the ball versus determined points in 3d space (or on the XY plane) for the c axis rotary using 3 measured positions for the ball after rotating the table into different positions).

    (He made some very good points I thought !).

    Didn't totally expect this to turn into a ball fondling thread lol (too late).

    Jargon busting for the "casual" reader MZRP --->" Machine Rotary Zero Point (MRZP) Offsets"

    I.e. trying to find the center points of rotation for the axes, but in truth there is not one center point of rotation but TWO axes of rotation defined in 3 space. On something like a HAAS UMC 750 the two lines defined in 3d space are not going to intersect with each other. IF proper transformations are carried and the position and orientation of each axis is mapped out it shouldn't matter. Some machine layouts don't have intersecting rotary axes / assemblies that intersect at one point in 3d space. B axis mill turn (5 axis) is just one example of that... The Renishaw axiset routine is therefore necessarily different... As shown in post #47.

    MRZP WIPS Offsets Settings - UMC-750 - NGC

    ^^^ Things you can do with MZRP procedures on a HAAS UMC (if things are not working out ).


    __________________________________________________ __________________________________________________ ____

    * No affiliation

  23. #57
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Florida
    Posts
    3,444
    Post Thanks / Like
    Likes (Given)
    1483
    Likes (Received)
    1627

    Default

    Quote Originally Posted by cameraman View Post


    4 position method technique shown here with a Kunkle style 5 axis mori, they seem to need to tilt on the B axis at 90 degrees either...

    "The Kinematic module in NC Gage lets users quickly detect and correct errors in 4- or 5-axis motion. Automatically tune the center-of-rotation and angular deviations. For use with Hexagon, Renishaw, Blum or Marposs machine tool probes on Fanuc and MAPPS controllers only. See Machine Tool Measurement | Hexagon Manufacturing Intelligence for more information." *




    TCPC "Party trick" , that you have probably seen at trade shows as well. [I think Okuma does one with a curved probing stem and DTI spinning slowly around the ball. ].


    Then re-read what the Thesidetalker wrote in post #49

    didn't totally expect this to turn into a ball fondling thread lol (too late).

    Jargon busting for the "casual" reader MZRP --->" Machine Rotary Zero Point (MRZP) Offsets"


    MRZP WIPS Offsets Settings - UMC-750 - NGC

    ^^^ Things you can do with MZRP procedures on a HAAS UMC (if things are not working out ).








    __________________________________________________ __________________________________________________ ____

    * No affiliation
    OK, but remember this thread was started for the HAAS UMC500... I am sure Renishaw combined with other 'high end' builders can/will do some pretty damn cool stuff with their offset/rotary zero cal cycles...

  24. #58
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default



    ^^^ Like what Thesidetalker said in his post the problem is when the machine runs faster it can't maintain accurate positioning and moving especially in cut (same for a lot of machines). Second half of the video shows the TCPC trick with a sub micron digital gauge (real time)

    So for example this Makino D200Z can make the moves fairly quickly and stay within a couple microns on the ball (or much less). But
    also good basis to derive volumetric accuracy. Notice the use of a straight "Ball 'n stick" lol



    This is why the NIS "Cone test " is good one / good challenge to show how the axes are mover simultaneously in-cut

    The "Goodness" is determined by the circularity of an arbitrary profile taken from the cone. For this machine 4.7 micron circularity was achieved on the cone/ cone test but something of the order of 12 to 17 micron is more typical for a more 'Average" machine, and for really pretty good machine values of 7 micron are considered - really pretty good. Not to be confused with a ball bar test in the XY plane which should have much tighter values for roundness than the NIS "Titled cone test".


    Any large 5 axis machine is hard to be super precise (unless you are paying 20x the $ than a HAAS UMC-750)

    So it should be that the UMC-500 should be somewhat more rigid, tight and accurate than the UMC 750.

    Better IS Better , but then again if a UMC 750 works for your kind of work then that's a win-win.

    Seems the UMC 750 is being discounted pretty heavily these days.

  25. #59
    Join Date
    Nov 2014
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    2,729
    Post Thanks / Like
    Likes (Given)
    3243
    Likes (Received)
    721

    Default

    Quote Originally Posted by Mike1974 View Post
    OK, but remember this thread was started for the HAAS UMC500... I am sure Renishaw combined with other 'high end' builders can/will do some pretty damn cool stuff with their offset/rotary zero cal cycles...

    I agree !


    However... The UMC 500 in respect of the UMC 750... (as the UMC 500 has not been let out into the wild yet... (as of the time of this writing / post)).

    The thrust mainly concerns the UMC -750 --> great for what it does but it is a ponderous machine that relies heavily on the quality of installation and foundation and the operator and machinist to really understand what's going on with the machine and be able to carry out effective Kinematic style calibrations.


    In the right hands and environment I'm sure it has made folks a lot of $ for their businesses/ shops. [Probably made HAAS a boat load of dosh too and the UMC-500 is poised to be a "smash hit", become almost ubiquitous.].

    In the case of the UMC -500 because it is more compact and rigid especially with it's castings and only weighing 12,000 lbs it should be easier to achieve more accurate results than the UMC 750 or to a lesser extent the UMC 1000. (The UMC 500 is proportionately more rigid (smaller force loops and smaller deflection within it's more compact structure.).

    Bear in mind there are 5 axis machines that are set on three points / feet and have a monoblock construction where one does not have to go through quite so much "Bollocks" just to have their machine perform properly + it does not need an engineered foundation. For example Hermle and others have a casting (like modified gantry style and one piece lower casting that's on 3 points where they don't want you to bolt the machine to the floor.). DMG have their 3 point "Monoblocks" like the DMU-60 etc. [Obviously different price points and capabilities. The Hermle's are designed so that easy mechanical alignments and adjustments can be made, in contrast the UMC-750 installation procedure looks really tricky, like what thesidetalker was saying about the seemingly paradoxical nature of making various iterative adjustments where one is not sure of the outcome and what affects what and in what order ??? like he said " Time travel into the past and kill your father" (what happens?)... I.e. Hard to make needed adjustments with confidence and know what one is doing (with predictable outcomes ) without messing the machine up. ] (tricky leveling versus kinematics + squirrelly entry of obscure parameters on the control + need to go "off manual/ off procedure" to effect what is needed / get'er done.). .

    But seems with the HAAS UMC-500 could actually sneak into a more useful accuracy and machine tolerance band. I think a lot of people are seeing the potential opportunity(ies) with the UMC-500 as its going to be a very big seller IMO (and as other have said as well. )…

    So if it is possible to calibrate the snot out of it using advanced methods and write software for that, including things that can monitor foundation loading that might be worth while (something that could be run off a tablet and monitors how stuff is going + kinematic cycles / different more advanced "Hacks"). The problem is IS that worth while or should people just buy a better machine ? So price point and practicality and good interface design (no brainer to operate) could be a very good way to go. [Something I might be interested in taking a risk on... Like what Eric form OrangeVice said, HAAS is low risk as you can sell these machines (on the second hand market (ready buyers )), and still not have lost that much $ on the machine / ROI.].

    Seems there could be a lot of spin off applications and business opportunities created by the HAAS UMC-500 (in general), as well as the educational and institutional "Space"; like Fred Mari chimed in to gauge how many HSK 63 A tool holders he might need to crank out for anticipated demand from the new UMC-500. I'm sure Eric from orangeVice is working on new interesting and imaginative products …

  26. #60
    Join Date
    Mar 2013
    Country
    UNITED STATES
    State/Province
    Oregon
    Posts
    1,734
    Post Thanks / Like
    Likes (Given)
    711
    Likes (Received)
    2004

    Default

    Quote Originally Posted by cameraman View Post
    [B]

    The thrust mainly concerns the UMC -750 --> great for what it does but it is a ponderous machine that relies heavily on the quality of installation and foundation and the operator and machinist to really understand what's going on with the machine and be able to carry out effective Kinematic style calibrations.
    I mean, you're spending a lot of time in this thread bench racing Haas against standards promulgated by 5 axis machine makers who I think everyone understands are playing at a totally different level.

    What Haas obviously understands is that:

    1- 95% of job shop machining in a 5 axis environment is positional.
    2- The vast majority of that work isn't micron accurate fussy business.

    And you know what? Thank god Haas recognizes those facts and builds pretty great quality 5 axis machines around it. The world is swimming in dreadfully expensive German/Japanese 5 axis machines - Hermle, Grob, Yasda, DMG, Makino, Okuma, Matsuura plus a bunch of folks like Kern who are even more exotic. There is a certain point where these machines become exercises in excess that drives the price up to the point where I really scratch my head wondering just how many business plans actually need the price points being pushed on them by the industry. In the real world - there are a shitload of shops out there who recognize that the ability to hit 5 sides of a part is a massive time-saver and that is about all they really want.

    I accidentally walked into a shop doing machining on very thin wall titanium castings. Accidentally because the metrology shop I was meaning to go to had moved, and Google hadn't updated the address, likely because they were owned by the same guy. The manager gave me the nickel tour and low and behold - 7 UMC750s sitting there, knocking out some crazy parts with very tight tolerances. Dude was super happy with the machines, and they had considered Makino, DMG and Hermle. With in-process probing and some creative macro writing to track each casting's distortions, he said the UMC's had absolutely no problem banging out work all day. Given the pricing, it was absurdly more productive to have 7 spindles for the same money that would have bought them 3-4 spindles from the other guys.

    Say what you will of Haas, they don't have their heads up their asses the way a lot of other machine manufacturers do.


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
  •