New alignment changed the nominals

[Josh Seiden]

Everything was going good until I added the second alignment near the end of the program. I tried adding the true position for Datum B to the report and it said that it was +250 in Y, so it was using the 2nd alignment for B, even though B is the base alignment. Any suggestions? Remember, I'm new to PC DMIS.

[JacobCheverie]

PC-DMIS will locate features from the active alignment or the DRF, depending on how you have it set up. Did you come from Calypso? Because Calypso uses “Base Alignment” and I feel that you are looking for that behavior.

Are you using Legacy or Xact? What is the DRF that you are dimensioned to? Also, in your second alignment, if the DRF is [A|B|D], then B will still constrain translation in Y due to the order of precedence (can/may/must rule).

[JacobCheverie]

Josh Seiden

DRF is your Datum Reference Frame. This is your datum structure (up to three datum feature references applied RMB, LMB, or MMB) at the end of your positional tolerance.
Legacy is the older PC-DMIS dimensioning style. You can activate it by navigating to INSERT - DIMENSION - USE LEGACY DIMENSIONS.
XactMeasure is the newer (although it is being replaced by the Geometric Tolerance command in 2020R2) dimensioning style. These two attempt to fully adhere to the GD&T standard that your print is drawn to, whether that's ASME Y14.5 or ISO 1101.

There are users that prefer Legacy over Xact and vice versa.

If you are using Xact (USE LEGACY DIMENSIONS is not checked off and you will see XactMeasure GD&T in the title bar of the dialog window), it is quite simple to report your result in a different alignment. Navigate to the Advanced tab at the top and, at the bottom of the window, there will be a drop down menu for your alignment. You can choose Datum Reference Frame (DRF) or Current Alignment.

DRF will report your result in the coordinate system established by your datum structure whilst respecting order of precedence as per your GD&T standard choice in the drop down menu on the Feature Control Frame tab.
Current Alignment will report your result in the coordinate system established by the ACTIVE alignment. You can change your active alignment through the Settings toolbar, typically found at the very top of the PC-DMIS window under the FILE, EDIT, VIEW, etc.. bar. The second drop down menu will let you recall earlier alignments that you have made. This must be done before the Xact dimension command.

[Josh Seiden]

I tried it and it works. Thank you. Now my new problem is that, when I run the program, it changes the nominals on the report. How do I fix this?

[JacobCheverie]

@Josh Seiden

In the Advanced tab of the XactMeasure window (where you went to change the active alignment), at the very bottom there will be a window where Nominal, +Tol, -Tol values can be changed. Make sure that the correct nominals (matching the desired alignment coordinate system) are showing in this window.

[Josh Seiden]

JacobCheverie This all makes sense now. Thank you very much for the help. I will try this out.

[Paperback Rocker]

PC Dmis is linear programming, unlike Calypso. When you load or create an alignment, it is in effect for the rest of the program until a new alignment is created or loaded. Create your Feature Control Frame in Exactmeasure when dimensioning to be equivalent to recalling the alignment into the feature or characteristic in Calypso. Going to PC Dmis from Calypso is going to be somewhat hard. Get as much training as you can. Good luck.

[Josh Seiden]

Paperback Rocker Thank you for the information. It's starting to make sense now.

[Josh Seiden]

JacobCheverie this is my main alignment. Is this what you would do? The part checks just fine, but I think my true position is getting messed up when I choose the datums in XactMeasure. I'm selecting A, B, and C, but I think it's losing the rotation in the process.

[JacobCheverie]

The alignment structure looks fine. I would perform the ROTATE command before any TRANSLATE commands, but that won't cause the problem that you are seeing.

A few questions:
Are you using CAD?
When you punch in A|B|C to XactMeasure, have you looked at the Advanced tab to make sure that the Nominal values are correct per your print?
Are your THEO/TARG values correct on your features? Incorrect THEO values or THEO vectors can cause many issues.
When you say your Position is getting "messed up", are you referring to the nominal values or the measured value?

[Josh Seiden]

• JacobCheverie
• Are you using CAD? No
• When you punch in A|B|C to XactMeasure, have you looked at the Advanced tab to make sure that the Nominal values are correct per your print? Yes, they are correct before I run the program.
• Are your THEO/TARG values correct on your features? Incorrect THEO values or THEO vectors can cause many issues. Not really. I notice that every time I run the part they change.
• When you say your Position is getting "messed up", are you referring to the nominal values or the measured value? Sometimes both. I went to the Advanced tab and corrected everything, but, after running the program, they change. It looks like they're changing to the actuals.

I ran this part on another CMM and I know that it's good. I'm trying to replicate the results, but it seems that the accuracy is lost the further the dimension is away from A, B and C. For instance, Datum D is supposed to be 250mm, which it is according to the other report, but it's checking 251.5. The (2) threads on the Datum D face also have the true position off quite a bit.

I created (4) datums in the DRF (A, B, C, and D). I selected A, B and C for the top diameters. I selected D, A, and B for the (2) threads on Datum D.

Thank you for the help.

[JacobCheverie]

Your THEO values are changing to the ACTL values... This tells me that you are manually measuring the features without CAD.

Are you running in DCC mode and your nominal values are still changing? Perhaps you can post your code on here, separated by [code] & [/_code] (without the "_") tags?

[ctw00]

JacobCheverie this is my main alignment. Is this what you would do? The part checks just fine, but I think my true position is getting messed up when I choose the datums in XactMeasure. I'm selecting A, B, and C, but I think it's losing the rotation in the process.

This isn't relative to your current questions, but I would replace the spaces in the names with underscores (i.e. MAIN_DATUM_B). Try to only use numbers, letters and underscores in the feature names. Special characters like * - + / and others get treated as operators and that can introduce issues.

[Josh Seiden]

JacobCheverie there was too much to just paste, so I made it a text file. I had to delete some unimportant features to get it to size. Thanks again for the help.

[JacobCheverie]

Josh Seiden Just looking quickly at your program code I can say that your THEO values for your features and your THEO vectors are not clean. Your DATUM A B ALIGNMENT (as ctw00 mentioned you shouldn't have spaces in feature/alignment names) uses LEVEL to DATUM A and DATUM A THEO vector is <-.007, .006, .9999602>. All your THEO vectors need to be nice and clean (<0, 0, 1>). I would suggest you add some underscores to all of your feature/alignment names, clean up all of your THEO data (DATUM B CYLINDER shows THEO of -.004 in Y, is this true?), and see how things run for you after that.

Once it is all cleaned up you should try re-running it. If you're still having issues you can re-post the code and we can dig into it a little bit deeper.

P.S. When you are updating your THEO values, PC-DMIS asks if you'd like to update nominals. Respond with YES. This will send your correct nominal values down to your XactMeasure commands.

[Josh Seiden]

Thank you. I'll do all that and get back to you.

[JacobCheverie]

@Josh Seiden

Also, make sure that the THEO values on your hit vectors update when you update the feature THEO values. It should prompt you "Change the hit values to match the theoreticals?" and to that you should also say YES.

Your datum feature B (DATDEF) should be your "DATUM B CYLINDER" instead of "MAIN DATUM B", the reason being that the latter feature is a circle with four hit points. A datum feature should be measured much more densely and as a cylinder, in this case. Same goes for your datum feature C.

[ctw00]

JacobCheverie there was too much to just paste, so I made it a text file. I had to delete some unimportant features to get it to size. Thanks again for the help.

I had a quick look and in your program where it picks up the threads there is room for improvement. In your program you have it picking up 4 points per level at 2 levels with 0 pitch. If you pick up a point on one side of the thread and the point opposite of that the points will be 'off' relative to the thread by 1/2 of the pitch, meaning one could hit the high point of the thread and the other will hit the low point (it won't be exactly that in practice). I usually add the thread pitch in that way if it hits the high spot of the thread on one side, it will hit the same part of the thread on the other side. Likewise, the distance between the two levels should be equally divisible by the thread pitch. As an example, if you have a M10x1.5 thread, put 1.5 in the pitch section, chose CCW direction (if it is a left hand thread choose CW), space the levels 4.5mm apart (or 6/7.5/etc) and the hits should all end up on the same section of the thread. In the graphic display window the movement path of the probe will look helical and match the thread.

[JacobCheverie]

This, or if you have scanning capabilities you can scan very densely and evaluate using a maximum inscribed algorithm.

[Josh Seiden]

Thank you for the suggestion. I'm only a few weeks into PC DMIS, so I'll work my way up to that.

[Matthew D. Hoedeman]

I had a quick look and in your program where it picks up the threads there is room for improvement. In your program you have it picking up 4 points per level at 2 levels with 0 pitch. If you pick up a point on one side of the thread and the point opposite of that the points will be 'off' relative to the thread by 1/2 of the pitch, meaning one could hit the high point of the thread and the other will hit the low point (it won't be exactly that in practice). I usually add the thread pitch in that way if it hits the high spot of the thread on one side, it will hit the same part of the thread on the other side. Likewise, the distance between the two levels should be equally divisible by the thread pitch. As an example, if you have a M10x1.5 thread, put 1.5 in the pitch section, chose CCW direction (if it is a left hand thread choose CW), space the levels 4.5mm apart (or 6/7.5/etc) and the hits should all end up on the same section of the thread. In the graphic display window the movement path of the probe will look helical and match the thread.

OR, you can use a plane old circle and do 15 hits (with the proper pitch) and use a 0 to 1080 start/end angles. Pcdmis will then go around the circle 3 times, giving 5 hits per go-round, and 3 pitches of thread.

[Josh Seiden]

JacobCheverie it worked! Thank you, my friend. That's an easy fix.

[JacobCheverie]

Glad it worked, have fun with PC-DMIS.