LSP Pressure Settings Different on Other CMM

[acgarcia]

Good Afternoon,

Under F10>Probe Options. I have different pressure settings on 2 machines. Is it a good idea to change them?

I have 2 Tigos with identical probes running the same program using a LSP fixed scanning head. Same calibration routine for both. Both machines are on the shop floor. Measurements are different from 1 CMM to another on the same exact part. Some measurements are right on, some different by a couple of tens (0.0001) and some by differ by 0.005". Some passing measurements are failing on the other CMM.

I checked the probe options and the force settings are very different from 1 cmm to another. I have maintenance guys trying to adjust CNC machines to match what the CMM is reporting. An operator decided to check it on the other CMM and then we found measurement differences.

What do you all think?

[neil.challinor]

In theory, if the machines, probing systems and controllers are identical then they should have the same probe option settings. However, just because one machine gives a "good report" does not necessarily mean it is the one with the right settings. There should be a file on the PC that contains the correct settings - see this post https://www.pcdmisforum.com/forum/pc...033#post464033

[ogcharliebrown]

When I looked the force settings are based on the controller.

[neil.challinor]

Controller and probing system. The settings are also refined by the service engineer when he installs the machine and there should be a pcdparams file somewhere on the PC that contains the ideal settings that were derived for that particular machine.

[ogcharliebrown]

In the PC-DMIS 101 book on page 423-424 there are recommended parameters based on controller.

[acgarcia]

neil.challinor , so you are saying check for the differences in those text files and make changes in the settings editor and not F10 probing options?

I was going to get a QC inspector to check the part by hand to see which machine better match is hand tool measurements.

[neil.challinor]

Yes, check the files and use the settings editor. Not all of the settings are editable under F10

[acgarcia]

I submitted a ticket since I have active HMAs on both machines and also emailed the service team in my region. Waiting for a response and hopefully the outcome is good.

[acgarcia]

I don't have access to the directory. It says windows cannot find C:\autotune\ when I type it in the file explorer

[gibsonridge]

another thought would be to compare calibration results between the cmms

[acgarcia]

So I talked to the Hexagon Applications my region and they stressed to make sure all the probes and magnets where clean. So I got the green putty that comes with the probes and yeah, they were dirty. So I cleaned them and the contacts under the LSP, wiped the rubys with some alcohol, did a lower matrix with a 5x20, then proceeded to calibrate all the probes on both CMMs.

On the F10 probe settings, they asked me to change their first CMM to match the second CMM as I suppose the settings on the second cmm were correct. Did that.

I found a huge chuck missing on the bottom of the cal sphere on my second CMM. So I got the one from my the first CMM, added it to my second CMM probe parameters and re ran the lower matrix and probe calibration. So know they are using the same cal sphere.

My alignment uses a cylinder and a line. I probe a circle in Y- with a L shaped probe and another circle in Y+ with another L shaped probe. Then I create a line. Maybe I should make that a cylinder?

It was also suggested that when doing comparison studies that I add a extra dcc alignment to really lock down the repeatability of the alignment points. Did that.

I was really hoping this was the root cause and it wasn't. I did have better comparison results but a few X axis dims are different by 0.025". 1 CMM reads 0.002 and the other was about 0.023 so I don't know what happened there. Out of 44 dimensions, 14 of them have a difference greater that 0.001. I'm considering difference less than 0.001 acceptable at this point.

Not sure what to do. All eyes are on me on why these machines don't match.

[gibsonridge]

next suggestion would be to write a small program measuring your qual sphere with probe at A0B0, make it x0,y0,z0
then rotate probe to A90B0 measure sphere and continue with A90B90, A90B180, A90B-90
.look for the deviations from zero. do this on both machines . this should give you a better idea of where your errors may be.

[Peter Fuller]

Do both Tigo machines have the same controller? There is a chance that these two machines were bought at a time where we were changing over the controller from a B3 to the DC controller. If they do not have the same controller then they should have different option probe statements. The rub here comes if you're not using rigid stylii and that the different force settings can return variable readings between two machines. For example if you're using long carbon shaft stylii or thin shaft tungsten carbide stylii for small diameter probes the difference in force can report different data between the two controllers.

If the above is the case then it's recommended to run a study on an artifact (IE: ring gage) and adjust the force until your measurements coincide with the gage parameters. A more suitable goal would be to use a stiffer stylus but I understand it's not always possible.

[acgarcia]

Peter Fuller , they were bought a few months apart but both do have DC241 controller. Both machines are using identical styli builds and same calibration parameters. 25 hits on 3 levels, etc..

For the record, diameter and distance measurements match enough to be be considered good correlation. Some of those measurements are exactly the same to the 4th decimal place.

I have been dimensionsing the diameters with their XZ or YZ locations since day 1 of the cell way back. We made a second cell with another Tigo, same probes, same probe calibration, copied over inspection program some dims do not match. My biggest variation in measurements is in my X direction and the value is 0.0053".

I'm going to go through each hit and point to make sure the probes aren't shanking out and maybe I can see where they machines start to differ.

[acgarcia]

Ok, still having a problem with this. I found a huge problem with my dc and final dc alignment. I was asked to rotate my trihedron to make the CNC machine. That way operators can interpret the results easier. So X+ is Z+, Y+ is X-, and Z+ is Y-. I initially just leveled, rotated, and orgined to change direction of the axes to the CNC machinists liking without considering my datums for the CMM alignment. So I fixed that and immediately saw better results on 1 CMM to where a part actually passed inspection but not so much on the other CMM with the same passing part. The CMM still say the axis dims are still way off.

I changed my alignment. I have 2 circles that are different sizes, one at the top end and one at the bottom end. I created a center pnt in each circle and then a line that spans the height of the part (pocekt_cl). Then I have 2 more circles that are the same size, create center pnts in each and make a line (seat_centerline). Then I align to both those lines changing the direction to match the CNC machine. Is using 2 lines for an alignment acceptable? It puts the trihedron exactly where and oriented the way I want it. I'm treating the lines as if they were cylinders.

I did play with the seat centerline for XY orgin with pocket_cl for Z orgin and then X only orgin with seat centerline and YZ orgin for the pocket cl. Made very little difference.

FINAL_DCC_ALIGN=ALIGNMENT/START,RECALL:DC_ALIGNMENT,LIST=YES
ALIGNMENT/LEVEL,ZPLUS,FDC_SEAT_CENTERLINE
ALIGNMENT/ROTATE,XMINUS,TO,FPOCKET_CL,ABOUT,ZPLUS
ALIGNMENT/TRANS,XAXIS,FDC_SEAT_CENTERLINE
ALIGNMENT/TRANS,YAXIS,FPOCKET_CL
ALIGNMENT/TRANS,ZAXIS,FPOCKET_CL
ALIGNMENT/END

[acgarcia]

I'm also using 3 different probes to do this. I have a 5x100 A0B0 and 2 L shaped probes to get in those pockets. 50mm extension, 5 way, and 5x40mm probe. 1 L shaped probe faces the front of the CMM and the other faces the back of the CMM. I know that the L shaped can't be perfectly straight and I think the difference is due to those 2 probes.

[Peter Fuller]

Peter Fuller , they were bought a few months apart but both do have DC241 controller. Both machines are using identical styli builds and same calibration parameters. 25 hits on 3 levels, etc..

For the record, diameter and distance measurements match enough to be be considered good correlation. Some of those measurements are exactly the same to the 4th decimal place.

I have been dimensionsing the diameters with their XZ or YZ locations since day 1 of the cell way back. We made a second cell with another Tigo, same probes, same probe calibration, copied over inspection program some dims do not match. My biggest variation in measurements is in my X direction and the value is 0.0053".

I'm going to go through each hit and point to make sure the probes aren't shanking out and maybe I can see where they machines start to differ.

Which LSPX head to you have? There are several, there should be a sticker on the probe body under where the puck mounts.
Also what are your option probe statements for both machines. I'm looking at mine for DC and X1H, C, S are similar and X1M is different. But it sounds to me you have a lot of non-rigidity going on. I would also inspect every puck and make sure nothing is loose or contacting wrong or missing kinematic balls. I had a customer who had a magnet on one of their pucks that was corroding underneath and allowing for the puck to move about .0015-0025" each time it made a tool change, the issue was his alignments were done with this tool so it looked like the machine was broken when it was just a bad puck.

[acgarcia]

Mystery solved. After weeks of banging my head against the granite and go through each feature theoretical and measured values, it turns out the Y axis scale needed to be cleaned with alcohol.

After making program and alignment changes, cleaning the head and probe magnets, using the same cal and inspection program, the CMMs did correlate better with the same part. My Z axis dims were still out by 0.020" from nominal on 1 CMM. The Y axis scale is upside down under the Y rail so I didn't think it would be dirty but as I cleaned it with alcohol and a lint free cloth, it was dirty. Same part had Z axis measurements down to 0.001" from nominal.

I'm so glad it was simple and easy fix. Definitely learned something here for next time.

[Douglas]

you cleaned it all so well next time might be 2 years from now, plenty long enough to forget at least for me at my age anyway