hexagon logo
Register Sign in
  • Replies 56 replies
  • Subscribers 222 subscribers
  • Views 575 views
  • Users 0 members are here
Options
Related

LSP X1h + 2015.0 / 2015.1 = no go?

vpt.se
For anyone out there with an LSP X1h and are using 2015.0 or 2015.1:
Experienced any problems with calibrating your probes?

Example:

For the first two probes below, I have answered YES to the question if the tool has moved (only way in order to get values that matches the THEOs): 
T1A-5B-15      TEO X  -1.5046 Y  -5.6152 Z 284.6301 D 5.0000
T1A-5B-15      SNABB X  -1.4635 Y  -5.5247 Z 284.6052 D 5.0000 PrbRdv 0.0035
T1A-5B-15      MÄTT X  -1.4635 Y  -5.5248 Z 284.6044 D 5.0000 PrbRdv-0.0012 StdDev 0.0004

T1A0B0         TEO X   0.0000 Y  12.0000 Z 284.3600 D 5.0000
T1A0B0         SNABB X   0.0016 Y  11.9994 Z 284.3585 D 5.0000 PrbRdv 0.0031
T1A0B0         MÄTT X   0.0014 Y  11.9994 Z 284.3579 D 5.0000 PrbRdv-0.0017 StdDev 0.0004

T1A-90B-90     TEO X-203.8600 Y  -0.0000 Z  92.5000 D 5.0000
T1A-90B-90     SNABB X-203.0024 Y   0.0033 Z  93.3571 D 5.0000 PrbRdv 0.0027
T1A-90B-90     MÄTT X-203.0017 Y   0.0032 Z  93.3571 D 5.0000 PrbRdv-0.0015 StdDev 0.0005

T1A-90B90      TEO X 203.8600 Y  -0.0000 Z  92.5000 D 5.0000
T1A-90B90      SNABB X 202.5958 Y   0.4775 Z  93.2956 D 5.0000 PrbRdv 0.0034
T1A-90B90      MÄTT X 202.5957 Y   0.4776 Z  93.2956 D 5.0000 PrbRdv-0.0009 StdDev 0.0004

T1A90B-180     TEO X   0.0000 Y-203.8600 Z  68.5000 D 5.0000
T1A90B-180     SNABB X  -0.3890 Y-202.5555 Z  69.5875 D 5.0000 PrbRdv 0.0026
T1A90B-180     MÄTT X  -0.3891 Y-202.5557 Z  69.5873 D 5.0000 PrbRdv-0.0015 StdDev 0.0004

T1A90B0        TEO X   0.0000 Y 203.8600 Z  68.5000 D 5.0000
T1A90B0        SNABB X  -0.0159 Y 203.0495 Z  69.6721 D 5.0000 PrbRdv 0.0033
T1A90B0        MÄTT X  -0.0159 Y 203.0492 Z  69.6725 D 5.0000 PrbRdv-0.0015 StdDev 0.0005


The machine can't complete a DCC calibration of the angles, only A0B0... It either touches the calibration sphere or errors out giving me the "Invalid reading of measured point".

Hexagon tech tested the calibration using 2013 MR1 which passed, ie. PC-DMIS could calibrate the angles as expected so it seems that the issue is related to 2015.0 AND 2015.1. I have tried with deleting the probe files, creating new ones, creating new tool, hit/retract values, analog probe values etc., to no avail. Needless to say, machine is not fit for use in it's current state - if used together with 2015.*

Why not downgrade? Because with our CMM programs collection, backsaving is out of the question - that's why.

Anyone else seen this?
  • Vpt, a technician is planned to visit you on Monday. Also please check the things I outlined in mail.

    Got that confirmed by the higher-ups today, Anders - thanks.

    In my experience many techs do not yet have a good understanding of the FDC controllers. Could very well be that the first technician missed something. Sounds like Anders has someone heading your way so hopefully theyll get you sorted out.

    Well, I have zero knowledge about the innards of those, so I'll just assume the tech's know their stuff. AndersI's service record with me (us) is flawless. Very seldom he can't deliver - when he can't, he'll make sure someone else in his address book do.
  • An odd thing is that using a retract distance of 0.5 mm during calibration yields the "invalid reading of the measured point". Using a 2 mm retract distance and the machine might pass through calibration. I just tested a star probe calibration (2 mm retract) and the A0B0 passes just fine, but when moving to the angled positions (A90B-90), I get the "Probe deflection during CNC movement".

    0.5mm seems very little for a calibration... I always use between 3 and 5... and it's long, ok !
    The failure of star probe looks like a alignment problem.
    Could you check the alignment by yourself ? A90B90 and A-90B90 along a block aligned on inserts in the granit.
    Did you try to calibrate one angle once in manual mode, then dcc?
  • 0.5mm seems very little for a calibration... I always use between 3 and 5... and it's long, ok !
    The failure of star probe looks like a alignment problem.
    Could you check the alignment by yourself ? A90B90 and A-90B90 along a block aligned on inserts in the granit.
    Did you try to calibrate one angle once in manual mode, then dcc?

    I will leave this problem as it is for now, the tech will arrive Monday but I will update the thread with if/when the machine is up again and what the problem was.

    Enjoy your weekend, people!
  • Ok, it seems that the LSP X1h definition in the 2015.x versions of PROBE.DAT is wrong. It is about ~1.8 mm short...
    AFAIK, it only affects new angles/probes in 2015.x.

    I believe this calls for a big, red WARNING[/SIZE]
  • Ok, it seems that the LSP X1h definition in the 2015.x versions of PROBE.DAT is wrong. It is about ~1.8 mm short...
    You are only affected if you create new angles/probes in 2015.x.

    I believe this calls for a big, red WARNING



    You've got to wonder how that happens don't you?

    It's not like it's a brand new component definition, or the format has changed in some way.
  • Ok, it seems that the LSP X1h definition in the 2015.x versions of PROBE.DAT is wrong. It is about ~1.8 mm short...
    AFAIK, it only affects new angles/probes in 2015.x.


    It wasn't the first place to look for problems, and the fact that old probe files worked, and newly created ones didn't, both muddied the waters for a while and led to the finding - when I did a diff of probe.dat between 2015 and 2014 it stood out quite visibly...

    I might add that LSPX1S, LSPX1M and LSPX1C are unaffected by this bug. Increasing the prehit value also works around it, but creating the probe file in 2014.1 is the better temporary solution (as the length value stored in the probe file will be correct).

    Any probe file created with the wrong value needs to be re-created from scratch, on a system with a correct(ed) PROBE.DAT.
  • Copying values from an old probe.dat in a usrprobe.dat, and deleting values in the actual probe.dat could it be a solution ?
  • Ok, it seems that the LSP X1h definition in the 2015.x versions of PROBE.DAT is wrong. It is about ~1.8 mm short...
    AFAIK, it only affects new angles/probes in 2015.x.

    I believe this calls for a big, red WARNING[/SIZE]

    This is quite odd, because we are having no length problems with 2015.0, 2015.1 and LSPX1h over here.

    However - in the past we have had problems with the incorrect Z-ram-to-Wrist adapter plate being put into the virtual probe build, which throws the lengths off.
  • All credit for this finding (which was true Sherlock Holmes-style) goes to AndersI (and any additional crew he needed for it).

    I know the EULA states that you are not responsible (yada, yada...) but for the love of God...
    Looks like we have to take the fall with this one - if this error has affected our measuring results...

    Maybe I should start authoring our official "We're sorry, but..." letter to our suppliers?
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement