Base Alignment issues

[[Jo...]]

Hey guys,

So we've come into some issues that were noticing now, especially after a fresh install. We use a load alignment in the fixture program using a ref. sphere, offset the base alignment to first part start location and load that base alignment into the part program. We then do a "part alignment" each time we run a part to locate it and reference features we are measuring. What we are running into now is that sometimes the base alignment becomes just enough "off" after time, that it affects the ability of the part program to effectively measure parts. The base alignments get saved locally, and we run the fixture program before every part run, so in theory the base alignment should never be "off" and on top of that the base alignment from the fixture shouldn't affect the part program or its ability to measure correctly as we do the part alignment each time. Now i know this to be true because when certain features are acting up on a program, my usual is to first check probe calibration and make sure that's kosher, and measure again. if the value is still off then i delete the local base alignment, re run the fixture and re load that base alignment into the part program. now this is what usually fixes it and i don't know why. Does calypso cache base alignment data and never fully wipe it when we running the fixture? is that even the issue?

any and all tips or answers would help guys.

If you need anymore info just let me know.

-Josh

[[Jo...]]

an update to this issue.

There was an issue just now, from CMM to CMM. Over three CMM's, measuring the same part, two of the CMM's had very similar data, within tenth's for one feature. Then the third CMM had an 8 thousandths difference to the other two. I deleted the base alignment locally, re-ran the fixture and re-loaded the base alignment into the part program (for the third CMM). This worked by shifting the data for the feature up but still two thousandths lower than the other two CMM's. I then did the same process to the other two CMM's and the data for those two matched the first CMM, which was initially way off. My question is why does the base alignment affect the part in this way? why does the constant measurement of the fixture not correct the base alignment? is there a solution for this base alignment issue within calypso? Possibly a way to save them differently, or to program a way in which the base alignment cant affect the part program the way it is now.

(running on a rennishaw d12/d8 with tp20 module, 2017 calypso)

[[Ja...]]

Does your fixture program alignment get ran on all three machines, or do you have a separate fixture program alignment per machine?

[[Jo...]]

We run the fixture on all 3 machines. the programs are saved on the network so in that sense they are identical. Then calypso itself saves a base alignment for the fixture program locally on the C: Drive, in that sense they might be slightly different. That base alignment is then used for the part program initially to find the part, then there is a part alignment before any features are measured.

[[Ja...]]

I was concerned that the alignment is being shared between machines, in that case differences in probe calibration could continuously change the alignment. Are you running your alignment[CNC] or just the alignment? Do you have a loop w/conditional on your alignment to hone in on a repeatable alignment?

Maybe you can, just for testing, try taking your programs and making a _0, _1, and _2 copy; one for each machine. Reset the alignment on each and run one program per machine and compare the results. This may eliminate the possibility of problems developing over the network/shared usage.

[[Jo...]]

We typically use the non-cnc alignment, as we seem to run into less issues. We run everything in auto run, running the fixture first then running the part program that has been put into a pallet to run the whole fixture. we seem to get issues where the alignment we use "shifts" slightly, and that somehow affects the part program and can have a decent affect on the features we measure. I am trying to understand why that happens when we re-run the fixture constantly, and the base alignment is referenced to a fixed location (reference sphere).

[[Me...]]

Hi Josh,
This was one of the biggest problems I saw before... In the pallet icon in Autorun, try using the CNC named base alignment for the pallet side only. I found this to work the best for me. Be weary of this method though, on some older versions of the software I had other issues arise. I think they fixed those issues in newer versions though.

Another option is to do a manual alignment on the base alignment program.

Thanks,
JG

[[Ma...]]

Hi Josh, Mark Ensley from Zeiss. The (CNC) alignment goes to where the last part was measured at that location. The non (CNC) alignment is based on the last Manual alignment of the part in that location. When you say that the fixture alignment uses a "ref sphere" does that mean that the fixture alignment only gives you X,Y and Z coordinates without a Spatial and Planer feature ? Does the fixture positively locate the parts accurately every time? I have attached some info on differences of non CNC and the CNC alignments from the User Guide in Calypso under the "?" drop down. Looping the alignment will also improve the accuracy of your measurements.

Base - CNC alignments.PNG

[[Ja...]]

Mark,

When you say looping of the alignment will provide greater accuracy: The current method I use to loop an alignment is by setting a condition baseSystem().valueA < x; is this the best method? I will typically set a value for x depending upon the required accuracy (.0001" - .0005")

[[Jo...]]

In the fixture we use a load alignment as a reference (which is a reference sphere in the front right hole location) to help create the base alignment. Then we typically use a 3-2-1 base alignment set up that is measured on the fixture. (plane, line and point). Offset that alignment to the first location of the part on the fixture. We then use that base alignment created as the base alignment in the part program, then measure features on the part to create a "part" alignment. Then in auto run we do a pallet set up to step around the fixture. Which in auto run we typically don't use the CNC alignment. What i am seeing is that eventually, the alignment becomes misaligned in some way and the only way for me to make it measure correctly again is to delete the base alignment, remeasure the fixture and re-load the base alignment into the part program. I've attached some images for reference.

3.JPG2.JPG1.JPG