Alignments and dimensions

[bjacobson]

To simplify what I'm doing, let's say I have a measured point with a +Z vector, a feature set containing that point, and a formandlocation profile of that feature set reporting a deviation of 0.43mm. Now, I want to evaluate that profile in another alignment. So I add an alignment with a 100mm Z translation, create a new profile and use the previous feature set again reporting a deviation of 0.43mm.

The 2 profiles report exactly the same deviation. I would expect that the deviation of the second profile would report an additional 100mm of deviation in Z, but it does not.

However, if I do the same trial with a location dimension, the deviation shows the additional 100mm.

What am I missing? It appears that the location dimension does not update the nominal, but the profile dimension does. Why would the profile dimension do that? Is there a way to stop it, and make it behave like the location dimension? I need to see the effect the second alignment has on the profile, but it's not reflected in the deviation.

Thanks!!

[RIDER]

Are you using CAD? Are you using legacy or Exact? Are you selecting form and location or form only? Is your alignment locking down all 6 degrees of freedom? Need a little more info please.

[Roberto]

Your 2nd alignment is just "offseting" Z, meaning profile is evaluating a basic distance that is 100 mm larger but the deviation should be the same (i.e. 0.43mm)

[bjacobson]

Yes, then why isn't the location reflecting that also? Either the location changed, and would be reflected in the profile, or it didn't.

[Ego Murphy]

To simplify what I'm doing, let's say I have a measured point with a +Z vector, a feature set containing that point, and a formandlocation profile of that feature set reporting a deviation of 0.43mm. Now, I want to evaluate that profile in another alignment. So I add an alignment with a 100mm Z translation, create a new profile and use the previous feature set again reporting a deviation of 0.43mm.

The 2 profiles report exactly the same deviation. I would expect that the deviation of the second profile would report an additional 100mm of deviation in Z, but it does not.

However, if I do the same trial with a location dimension, the deviation shows the additional 100mm.

What am I missing? It appears that the location dimension does not update the nominal, but the profile dimension does. Why would the profile dimension do that? Is there a way to stop it, and make it behave like the location dimension? I need to see the effect the second alignment has on the profile, but it's not reflected in the deviation.

Thanks!!

Part 1: Profile
Both Profile and Location compare actual to nominal and compute the difference, however the values displayed by reporting are different.

Profile is a GD&T dimension. All GD&T dimensions are this: a tolerance zone of allowed deviation from nominal.
If measured data lands outside of the tolerance zone, it's bad. For all GD&T dimensions, the reported result (whether MEAS or MIN & MAX) is expressed as deviation from nominal, not deviation along an axis from current coordinate system origin. The deviation from nominal should not be greater than the allowed zone.

Location is a Linear dimension. It too has a tolerance zone that should not be gone past, however the MEAS result is expressed as deviation along an axis from current coordinate system origin. Note that deviation from nominal - the DEV header item in F10 Dimension Options - is also reported by default.

But wait, there's more:
Profile tolerance zone has two boundaries - the high limit and low limit, usually equally spaced around the nominal.
If you are dimensioning Profile to datums, it is imperative to use F10 Dimension Options and turn on MAX & MIN as the default MEAS value can't give you the whole story - is the dimension failing because the actual shape's high point is past the high boundary, or is the low point past the low boundary?

Part 2: Alignments
A new alignment is a new coordinate system. PC-DMIS's frame of reference is cursor-dependent. If your cursor is past the new alignment, all PC-DMIS's thinking is in the new frame of reference.
As you noted, Location reports an updated version of the actual-and-nominal status of the data - transformed to the new coordinate system.
The profile results remain unchanged, because they report deviation from nominal. Note that in the Location dimension the DEV results are also unchanged.

Part 3: Correlation
Here's an excellent, fun, and easy way to get perfect correlation between Profile and Location: T-Value, which is deviation along feature vector. This is basically a "profile of a single point" dimension.
Take a bunch of points and make sure they all have perfect nominals. The easiest thing is a flat plane, like the top of a block, where they all have the same Z coordinate.
First, dimension them all for T-value. Read all the results and write down the maximum MEAS T value and the minimum MEAS T value. Note that in each one, the MEAS and DEV values are exactly the same.
Next, create a Constructed Set containing all the points. Hit F10 Dimension Options and turn on MAX & MIN, then create a Profile dimension of this set - and make sure it has the same + & - values that you used in the T value dimensions. Vioila! The profile's reported Max and Min will exactly match the max & min T values you found.

I hope this helps!

[bjacobson]

Thanks for the input everyone.

When making changes to an alignment, PCDMIS asks "Alignment XXX has changed! Update dependent commands?" Clicking no means the profile deviation will include the offset.

Here's the thing...if you put a translation into an alignment ABOVE the measured points and click no to the question above, PCDMIS adjusts the nominals and their deviations reflect the change. If you put it in BELOW the measured points, and still click no, PCDMIS will offset both the nominal and the measured, and the deviation will not reflect the changes...but only in profiles.

So if I put a Z translation of 100mm in to my initial best fit BEFORE the measured points, the report shows the deviations to have the additional 100mm. But if I add an alignment AFTER the points (but before the dimensions, or even the feature sets) with a 100mm offset, and click no when asked to update, the profile deviations are unchanged. Yet all my locations show the additional 100mm in their deviations.

If the point's LOCATION changed 100mm from nominal, then the PROFILE (with the same nominal location) should also reflect the 100mm deviation. It either moved or it didn't.

Further...a formandlocation profile uses a point's nominal from datum. So if that datum is moved then the nominal location moves with it and the deviation from the nominal to the measured location changes. Exactly like the location dimension shows is happening. Why not in the profile?

[Schrocknroll]

Let's say your Z nominal is .500 and measures .930, that's a deviation of .430.
Then you offset the alignment by 100mm. Your new nominal is 100.500 and the measured value is 100.930. That's a deviation of .430, the same thing.

As has been stated, just offsetting an alignment will not change a profile, because the nominal will change by the offset amount.

[bjacobson]

If that's true then why does it change a location? If you do the same exact thing, the locations will show 100mm additional deviation.

[Schrocknroll]

If that's true then why does it change a location? If you do the same exact thing, the locations will show 100mm additional deviation.

It should show 100mm difference in the nominal as well as the measured value giving you the same deviation, unless you override it by manually changing the nominal in the dimension. PC-MDIS is a dynamic software that is constantly recalculating things. It makes it powerful and very dangerous at the same time. See below.

PNT1       =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<0,0,0.5>,<-1,0,0>
            ACTL/<0,0,0.5>,<-1,0,0>
            TARG/<0,0,0.5>,<-1,0,0>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=NO
A1         =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
              ALIGNMENT/TRANS_OFFSET,ZAXIS,-100
            ALIGNMENT/END
DIM LOC1= LOCATION OF POINT PNT1  UNITS=MM ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH  HALF ANGLE=NO
AX    NOMINAL       +TOL       -TOL       MEAS        DEV     OUTTOL
Z      100.500      0.100      0.100    100.500      0.000      0.000 ----#----
END OF DIMENSION LOC1

[bjacobson]

It should show 100mm difference in the nominal as well as the measured value

But it doesn't, and I didn't override anything...

DIM PRF_1= PROFILE OF SURFACE OF POINT PNT_1    FORMANDLOCATION  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=25.00  ARROWDENSITY=31  OUTPUT=BOTH
AX    NOMINAL       +TOL       -TOL       MEAS        MAX        MIN        DEV     OUTTOL
M        0.000      0.000      0.000      0.029      0.029      0.029      0.029      0.029 |->
DIM LOC_1= LOCATION OF POINT PNT_1  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=25.00  OUTPUT=BOTH  HALF ANGLE=NO
AX    NOMINAL       +TOL       -TOL       MEAS        MAX        MIN        DEV     OUTTOL
X        4.684      0.002     -0.002      4.685      4.685      4.685      0.002      0.000 --#
Y       -0.112      0.002     -0.002     -0.057     -0.057     -0.057      0.055      0.053 -->
Z       -0.598      0.002     -0.002     -0.633     -0.633     -0.633     -0.035      0.033 <--
T        0.000      0.002     -0.002      0.029      0.029      0.029      0.029      0.027 -->
END OF DIMENSION LOC_1
ALGN       =ALIGNMENT/START,RECALL:PART_LFT_ALGN,LIST=YES
              ALIGNMENT/TRANS_OFFSET,ZAXIS,100
            ALIGNMENT/END
DIM PRF_2= PROFILE OF SURFACE OF POINT PNT_1    FORMANDLOCATION  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=25.00  ARROWDENSITY=31  OUTPUT=BOTH
AX    NOMINAL       +TOL       -TOL       MEAS        MAX        MIN        DEV     OUTTOL
M        0.000      0.000      0.000      0.029      0.029      0.029      0.029      0.029 |->
DIM LOC_2= LOCATION OF POINT PNT_1  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=25.00  OUTPUT=BOTH  HALF ANGLE=NO
AX    NOMINAL       +TOL       -TOL       MEAS        MAX        MIN        DEV     OUTTOL
X        4.684      0.002     -0.002      4.685      4.685      4.685      0.002      0.000 --#
Y       -0.112      0.002     -0.002     -0.057     -0.057     -0.057      0.055      0.053 -->
Z       -0.598      0.002     -0.002   -100.633   -100.633   -100.633   -100.035    100.033 <--
T        0.000      0.002     -0.002      0.029      0.029      0.029      0.029      0.027 -->
END OF DIMENSION LOC_2

[Mike Ruff]

Your measured value is 100mm lower after you offset your alignment by 100mm. Recreate the DIM_LOC_2 and your nominal should change 100mm. If not, then you can change it manually to -100.598.

[Schrocknroll]

But it doesn't, and I didn't override anything...

DIM PRF_1= PROFILE OF SURFACE OF POINT PNT_1 FORMANDLOCATION UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 ARROWDENSITY=31 OUTPUT=BOTH
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
M 0.000 0.000 0.000 0.029 0.029 0.029 0.029 0.029 |->
DIM LOC_1= LOCATION OF POINT PNT_1 UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 OUTPUT=BOTH HALF ANGLE=NO
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
X 4.684 0.002 -0.002 4.685 4.685 4.685 0.002 0.000 --#
Y -0.112 0.002 -0.002 -0.057 -0.057 -0.057 0.055 0.053 -->
Z -0.598 0.002 -0.002 -0.633 -0.633 -0.633 -0.035 0.033 <--
T 0.000 0.002 -0.002 0.029 0.029 0.029 0.029 0.027 -->
END OF DIMENSION LOC_1
ALGN =ALIGNMENT/START,RECALL:PART_LFT_ALGN,LIST=YES
ALIGNMENT/TRANS_OFFSET,ZAXIS,100
ALIGNMENT/END
DIM PRF_2= PROFILE OF SURFACE OF POINT PNT_1 FORMANDLOCATION UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 ARROWDENSITY=31 OUTPUT=BOTH
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
M 0.000 0.000 0.000 0.029 0.029 0.029 0.029 0.029 |->
DIM LOC_2= LOCATION OF POINT PNT_1 UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 OUTPUT=BOTH HALF ANGLE=NO
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
X 4.684 0.002 -0.002 4.685 4.685 4.685 0.002 0.000 --#
Y -0.112 0.002 -0.002 -0.057 -0.057 -0.057 0.055 0.053 -->
Z -0.598 0.002 -0.002 -100.633 -100.633 -100.633 -100.035 100.033 <--
T 0.000 0.002 -0.002 0.029 0.029 0.029 0.029 0.027 -->
END OF DIMENSION LOC_2

That's odd. I can't reproduce those results. I'm guessing it's a setting, in F5 maybe, but I tried a few things but it always shifts the nominal along with the measured for an offset alignment. I'll keep thinking about it, but that's all I have for now.

[Mike Ruff]

That's odd. I can't reproduce those results. I'm guessing it's a setting, in F5 maybe, but I tried a few things but it always shifts the nominal along with the measured for an offset alignment. I'll keep thinking about it, but that's all I have for now.

This usually happens when alignments are inserted/adjusted in front of dimensions/ features in the dimensions. My workaround is to change then unchange the affected dimension/ features associated with the dimensions. This will force PC-DMIS to reevaluate the dimensions without actually changing anything. This should trigger the popup "Do you want to update nominals of related dimensions?" or something along those lines... If you answer yes, then nominal should update to the expected value. If this popup isn't triggered then that question is likely set to "No" in the F5 warnings as mentioned above.

[NinjaBadger]

But it doesn't, and I didn't override anything...

DIM PRF_1= PROFILE OF SURFACE OF POINT PNT_1 FORMANDLOCATION UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 ARROWDENSITY=31 OUTPUT=BOTH
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
M 0.000 0.000 0.000 0.029 0.029 0.029 0.029 0.029 |->
DIM LOC_1= LOCATION OF POINT PNT_1 UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 OUTPUT=BOTH HALF ANGLE=NO
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
X 4.684 0.002 -0.002 4.685 4.685 4.685 0.002 0.000 --#
Y -0.112 0.002 -0.002 -0.057 -0.057 -0.057 0.055 0.053 -->
Z -0.598 0.002 -0.002 -0.633 -0.633 -0.633 -0.035 0.033 <--
T 0.000 0.002 -0.002 0.029 0.029 0.029 0.029 0.027 -->
END OF DIMENSION LOC_1
ALGN =ALIGNMENT/START,RECALL:PART_LFT_ALGN,LIST=YES
ALIGNMENT/TRANS_OFFSET,ZAXIS,100
ALIGNMENT/END
DIM PRF_2= PROFILE OF SURFACE OF POINT PNT_1 FORMANDLOCATION UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 ARROWDENSITY=31 OUTPUT=BOTH
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
M 0.000 0.000 0.000 0.029 0.029 0.029 0.029 0.029 |->
DIM LOC_2= LOCATION OF POINT PNT_1 UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=25.00 OUTPUT=BOTH HALF ANGLE=NO
AX NOMINAL +TOL -TOL MEAS MAX MIN DEV OUTTOL
X 4.684 0.002 -0.002 4.685 4.685 4.685 0.002 0.000 --#
Y -0.112 0.002 -0.002 -0.057 -0.057 -0.057 0.055 0.053 -->
Z -0.598 0.002 -0.002 -100.633 -100.633 -100.633 -100.035 100.033 <--
T 0.000 0.002 -0.002 0.029 0.029 0.029 0.029 0.027 -->
END OF DIMENSION LOC_2

How did you get to this point?

Was this code created in the order it's shown? Or did you do the dimensions and then insert the offset alignment after?

[bjacobson]

This usually happens when alignments are inserted/adjusted in front of dimensions/ features in the dimensions.

The thing is, I WANT this to happen. But I want it to happen to all dimensions. Only profiles don't respond the same way. We target the parts (cut the datums) and I want to adjust the datums prior to targeting, and see the result in order to manually best fit the parts. Currently, we take a series of points scattered around the part and use a best fit alignment to locate the part. Then, we measure the entire part. We can make manual adjustments in the best fit alignment to alter the datums, and see the result.

Currently we do this:

measure 25-30 best fit points
best fit alignment (where we make adjustments AFTER the part is ran)
measure entire part (up to 3000 points)
Dimensions

What I want to do:

measure 25-30 best fit points
best fit alignment (where we make adjustments AFTER the part is ran)
measure entire part (up to 3000 points)
2nd best fit alignment using all measure points (where we can still make manual adjustments)
Dimensions (these should ALL be affected by the changes in the second alignment)

I want the dimensions to show the additional deviation that results in altering the datum. I don't have the setting saved in F5 warnings, because sometimes I click yes, sometimes I click no.

How did you get to this point?

Was this code created in the order it's shown? Or did you do the dimensions and then insert the offset alignment after?

Created in that order, then the 100 was added. The dimension did exactly what I wanted it to do, I just want profiles to do it too...they don't.

[Mike Ruff]

So you want to tell PC-DMIS that the datums are somewhere they actually aren't and measure it to those fake datums based on manual offsets to the actual measured datums?

[bjacobson]

Fake datums? No. If I move the datum, it's not fake, it's simply in a different location.

[NinjaBadger]

Okay, so this is what's going on.


Profile is looking at deviation from nominal. It doesn't matter where you move your origin to, it's comparing the measured value of that point with the theoretical value of that point, you move your origin by 100 and in that new alignment, the theo is 100 different, but so is the measured. The deviation from measured to theo remains the same.

Now for the location, that's reporting the distance from the origin, which of course has changed (because you moved the origin.)

When you inserted the 100 offset, you answered 'No' - to the 'update the commands below question' (or, I suspect someone may have ticked the 'Don't ask me again' tickbox. Check Warnings on the general tab of Setup(F5))

A LOT of people don't fully understand quite how this works...

When you say 'Yes' to this question, everything gets updated; so if you substantially move your origin, the theo's and measured and nominal and actuals all get updated. Typically you'd answer Yes if you're actually moving your origin.

When you say 'No' to this question, the Theo's of subsequent features and nominals remain the same, and the actuals DO NOT move relative to the physical part, but they DO in relation to the origin. (Hence why your Measured values have shifted by 100 - but the nominal hasn't changed). As someone said, if you recreated the dimension the nominal would be different by 100 from how it shows in your second location dimension.


Edit: If you stick the 100 offset before the feature and answer No to update commands below you will see the shift you want.

Make sure 'Allow fine tuning of Alignments' is unticked (General Tab of Setup).

[bjacobson]

"Edit: If you stick the 100 offset before the feature and answer No to update commands below you will see the shift you want."

No, I don't. That's the whole point. I understand the difference between answering "no" and answering "yes". That's not the issue. The issue, is exactly what I stated above. The location shows the 100mm offset, but the profile does not. Even you state that it should, but it doesn't.

[NinjaBadger]

@bjacobson


No. I said if you put the offset before the features then you will see the difference in the profile dimension. (which I believe you stated above in post #5).

With the offset after the features and before the dimensions this is behaving exactly as it should.

Location is reporting deviation from the origin.
Profile is reporting the deviation from nominal - in your new/offset alignment the deviation from measured to nominal is the same as before the offset alignment.

In post #5 (last paragraph) you state:

"Further...a formandlocation profile uses a point's nominal from datum. So if that datum is moved then the nominal location moves with it and the deviation from the nominal to the measured location changes. Exactly like the location dimension shows is happening. Why not in the profile?"

I think this might be your stumbling block in understanding what's going on, although with form and location the profile is fixed with respect to the datum(s), it's a comparison between nominal and measured. You move the alignment and they both move by the same amount, the difference doesn't change.

I hope this helps

https://www.dropbox.com/s/hw1xoluo8e7cgce/profile%20vs%20location.docx?dl=0

[Roberto]

Okay, so this is what's going on.


Profile is looking at deviation from nominal. It doesn't matter where you move your origin to, it's comparing the measured value of that point with the theoretical value of that point, you move your origin by 100 and in that new alignment, the theo is 100 different, but so is the measured. The deviation from measured to theo remains the same.

+1
I tried to explain this on post# 3, my English is obviously a "work in progress"