hexagon logo
Register Sign in
  • Replies 13 replies
  • Subscribers 223 subscribers
  • Views 227 views
  • Users 0 members are here
Options
Related

Understanding Calibration Results

TheBradish
I don't usually look at the probe calibration results because I don't really understand what I'm looking at. I'd really like to gain a firm understanding of the data and what it should mean to me.

I've been digging through old threads, and I'm getting closer. What I think I get is that StdDev is the standard deviation of single-point error of the measured sphere for each tip angle. (I was told by another inspector here that that number should be .0002 or less.) I also understand that the MEAS D determines the effective probe diameter for each tip, and is used for probe size compensation.

I don't fully understand the THEO/MEAS XYZ (where the THEO comes from, what the MEAS means), or what is a maximum permissible error. What do you guys look for when you read your calibration results? What do you hope to learn from the numbers? What do you consider to be a red flag?
  • I don't know for sure what, specifically, is being calculated for StdDev, but AFAIK it is spherical form deviation.

    THEO is generated when you build a probe in the probe utility. It comes from the base of the Z rail, centered on the probe head. MEAS is what it actually measured and is what is used to compensate, along with the probes measured diameter.

    I run in MM and have it alert at .005 StdDev. The THEO/MEAS values I don't pay as much attention to. Diameter I have alert at .025 deviation, but you can double check the compensation by measuring a ring gage, or using probe qualification in the probe utilities measure options.
  • Vinni is right, theo comes from the build (probe.dat). The meas is calculated from real offsets of theprobe from the theo and the difference from the "master probe" (which one you use the last time when you said "yes, the ball have changed")
  • So the XYZ is a coordinate for the center of the ruby?

    I drew a quick sketch to see if I understand where the XYZ origin is, according to Vinni. Do I have it right?

  • So the XYZ is a coordinate for the center of the ruby?

    I drew a quick sketch to see if I understand where the XYZ origin is, according to Vinni. Do I have it right?



    As far as I know. I asked the Hexatech some questions last time he was here. He said Z zero is set at the bottom of the z-rail during the initial machine mapping. So, when your machine homes, the probe tip isn't at Z0, but at Z0 - probe head - probe build. For me this tends to be 0,0,-230 at "home".

    You can verify this at A0B0. Modify your build to include or remove a 50mm_CF extension. The Z will shorten by 50mm. The pivot point of the head makes calculating the XYZ awkward when at any other angle. But you can see the PH10 dimensions in its data sheet. I think there is some compensation for your master probe. For mine, A0B0 is 0,0,230 (which makes it looks like it calculates backwards from the probe center). It forces A0B0 then to be 0, and where it calculates everything else from.

    This isn't really "Official" information, but from what I have gathered out of curiosity. As such, it might be mis-informed, but it probably is correct, in general.
  • Looks like you're right about that. I just compared the difference in THEO Z for two of my probes (A0B0) and the difference in Z matches the difference in stylus length. I also compared the THEO Z of my master probe to the total length of my Renishaw hardware (using nominals from the Renishaw site) and they match within a couple mm.

    From looking at my results, it looks like my probe head might be slightly out of square. But if the qualification routine compensates for that, will it really make a difference? How much is too much? Here are some of my results for a 1x30 tip: 

    AUTOCALIBRATE/PROBE, PARAMETER_SET=ALL-TIPS-WITH-DEFAULTS, QUALTOOL_MOVED=YES,                           SHOW_SUMMARY=YES, OVERWRITE_RESULTSFILE=YES
BEGIN AUTOCALIBRATE RESULTS FOR PROBE MASTER USING SET ALL-TIPS-WITH-DEFAULTS
Probe file=MASTER    Date=11/17/2014    Time=4:37:04 PM
BALL IN STAND  CENT X35.152186 Y48.827637 Z-37.963942 D0.983550
T1A0B0         THEO X  0.00000 Y  0.00000 Z  5.96535 D0.15748
T1A0B0         MEAS X  0.00000 Y  0.00000 Z  5.96535 D0.15733 StdDev0.00003
END AUTOCALIBRATE RESULTS FOR PROBE MASTER USING SET ALL-TIPS-WITH-DEFAULTS


AUTOCALIBRATE/PROBE, PARAMETER_SET=ALL-TIPS-WITH-DEFAULTS, QUALTOOL_MOVED=NO, 
                          SHOW_SUMMARY=YES, OVERWRITE_RESULTSFILE=NO
BEGIN AUTOCALIBRATE RESULTS FOR PROBE 1MM X 30MM USING SET ALL-TIPS-WITH-DEFAULTS
Probe file=1MM X 30MM    Date=11/17/2014    Time=8:47:47 PM
T1A30B-180     THEO X  0.00000 Y -2.68740 Z  6.03267 D0.03937
T1A30B-180     MEAS X -0.04038 Y -2.68127 Z  6.03372 D0.03874 StdDev0.00006
T1A37.5B-180   THEO X  0.00000 Y -3.27197 Z  5.64207 D0.03937
T1A37.5B-180   MEAS X -0.04254 Y -3.27025 Z  5.64188 D0.03873 StdDev0.00008
T1A37.5B-172.5 THEO X  0.42708 Y -3.24398 Z  5.64207 D0.03937
T1A37.5B-172.5 MEAS X  0.38588 Y -3.24671 Z  5.64099 D0.03874 StdDev0.00008
T1A45B-180     THEO X  0.00000 Y -3.80056 Z  5.17851 D0.03937
T1A45B-180     MEAS X -0.04481 Y -3.80282 Z  5.17663 D0.03874 StdDev0.00007
T1A45B-172.5   THEO X  0.49607 Y -3.76805 Z  5.17851 D0.03937
T1A45B-172.5   MEAS X  0.45315 Y -3.77486 Z  5.17559 D0.03878 StdDev0.00007
T1A37.5B-165   THEO X  0.84685 Y -3.16048 Z  5.64207 D0.03937
T1A37.5B-165   MEAS X  0.80732 Y -3.16749 Z  5.64026 D0.03872 StdDev0.00009
T1A37.5B165    THEO X -0.84685 Y -3.16048 Z  5.64207 D0.03937
T1A37.5B165    MEAS X -0.89041 Y -3.15117 Z  5.64251 D0.03877 StdDev0.00005
T1A37.5B172.5  THEO X -0.42708 Y -3.24398 Z  5.64207 D0.03937
T1A37.5B172.5  MEAS X -0.46934 Y -3.23901 Z  5.64190 D0.03875 StdDev0.00006
T1A22.5B180    THEO X -0.00000 Y -2.05685 Z  6.34362 D0.03937
T1A22.5B180    MEAS X -0.03730 Y -2.04704 Z  6.34485 D0.03880 StdDev0.00007
T1A30B180      THEO X -0.00000 Y -2.68740 Z  6.03267 D0.03937
T1A30B180      MEAS X -0.04040 Y -2.68133 Z  6.03368 D0.03874 StdDev0.00009
  • Looks like you're right about that. I just compared the difference in THEO Z for two of my probes (A0B0) and the difference in Z matches the difference in stylus length. I also compared the THEO Z of my master probe to the total length of my Renishaw hardware (using nominals from the Renishaw site) and they match within a couple mm.

    From looking at my results, it looks like my probe head might be slightly out of square. But if the qualification routine compensates for that, will it really make a difference? How much is too much? Here are some of my results for a 1x30 tip:

    .....


    Bolded above are questions that I wish I knew the answer to also. Slight smile

    One of my machines was retrofitted to Renishaw hardware and then upgraded again sometime later (several years before me). The PH10 shaft fits in a hole that is several thousandths larger than it, with a single set screw in the back to hold it in place. The part that holds it is a modified bracket attached to the Z-Rail. Between the two pieces, there is a crapton of slop with how the probe head actually fits. It is definitely not quite square, and I had to add a custom spacer in the probe build before the PH10 to get the height approximately right...

    I used the THEO/MEAS values to try to fine tune it. A0B0, and A90B(0, 90,180, -90) are enough to try to fine tune any skewing. But again, how close is close enough? If I only knew.
  • Surely someone here must be able to enlighten us. Anyone else care to chime in?
  • The THEO is the number calculated to where it SHOULD find the probe offsets to be. If the head was perfectly square, the lengths were all perfect ect.....if that was the case, there would be no need to calibrate in a perfect world. As you will always see, there is some deviation of location, this is just fine...it is the actual amount it is going to use to offset. as long as you can get the probe to calibrate without crashing, it doesn't matter if the head is all that square...but once you add extensions, that out of squareness makes it next to impossible to calibrate without squaring it up...there is my chime...
  • From looking at my results, it looks like my probe head might be slightly out of square. But if the qualification routine compensates for that, will it really make a difference? How much is too much?

    The qualification mathematics take care of this, and if you calibrate manually there shouldn't be a problem (except that it has only been done manually).

    But if you calibrate in DCC, PC-DMIS assumes that the tips are as specified theoretically. For the probe angle A90B0, applying simple geometry gives (assuming I have understood the problem):

    The difference between where the tip actually is, and where PC-DMIS thinks it is (when running in DCC with THEO definitions) is roughly "Length of tip from rotation point" times SIN(angle between probe tip A90B0 and the closest machine axis [X or Y]). If this value gets as large as the prehit distance, you should expect problems.

    To this should be added that you get a small cosine error in the measuring points as soon as the hits are not exactly normal to the surface. I don't know (or have the time to calculate) if that will be significant.
  • hello, i want to add one more question with this thread...


    i'm using Hexagon Probe Head HH-AS8-T2.5 with Leitz_LSPX1H_T Probe and my calibration results are... 

    Probe file=2BY50 Date=06-Apr-18 Time=8:49:11 AM

MASTER CENT X 260.735 Y 384.713 Z -456.232 D 25.000
T1A0B0 THEO X 0.000 Y 0.000 Z 274.000 D 2.000
T1A0B0 FAST X -0.510 Y -0.081 Z 274.068 D 2.000 PrbRdv -0.018
T1A0B0 MEAS X -0.509 Y -0.084 Z 274.072 D 2.000 PrbRdv -0.004 StdDev 0.002

T1A90B0 THEO X 0.000 Y 208.800 Z 65.200 D 2.000
T1A90B0 FAST X 1.122 Y 209.040 Z 64.931 D 2.000 PrbRdv -0.012
T1A90B0 MEAS X 1.121 Y 209.039 Z 64.932 D 2.000 PrbRdv -0.002 StdDev 0.002

T1A90B90 THEO X -208.800 Y 0.000 Z 65.200 D 2.000
T1A90B90 FAST X -209.326 Y 1.207 Z 64.939 D 2.000 PrbRdv -0.012
T1A90B90 MEAS X -209.323 Y 1.207 Z 64.939 D 2.000 PrbRdv -0.002 StdDev 0.002

T1A90B180 THEO X -0.000 Y -208.800 Z 65.200 D 2.000
T1A90B180 FAST X -1.490 Y -209.235 Z 64.644 D 2.000 PrbRdv -0.010
T1A90B180 MEAS X -1.489 Y -209.233 Z 64.644 D 2.000 PrbRdv 0.000 StdDev 0.000



    In the above result what does the Fast and PrbRdv denotes ?
    The PrbRdv was repeating 2 times one in fast and another in measure but both the results are completely different.....
    can anyone explain this ?
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement