Feature Set to Surface Profile

[boydha]

Hi, I am new to the forum so please be considerate.

When you are creating a feature set and say you put in a circle, a line, and a plane, to create (scn1) when you output (scn1) as a profile Form+Location and MaxMin does this give you the profile of each individual point in those features or a profile of the centroids of those features?

If the centroids, Is there an easier way other than probing all the points over again or probing all the points and then creating your features to get what I am looking for?

Thank you for your help!

[JEFMAN]

If you just select features, it will be the centroids.
Then, in the edit window, you just have to write .HIT[1..CIRC.NUMHITS] after the circle, .HIT[1..LIN1.NUMHITS] after the line and .HIT[1..PLN1.NUMHITS] after the plane to use the points.

[AndersI]

Centroids. You need to create the set from the hits of the feature(s). This can only be accomplished by editing the command text in Command mode.

First create the set from your features (I added a way to see how many 'hits' are in the set)

SCN1       =FEAT/SET,CARTESIAN
            THEO/<-32.482,-30.968,-10.997>,<0,0,1>
            ACTL/<-32.482,-30.968,-10.997>,<0,0,1>
            CONSTR/SET,BASIC,PLN1,LIN1,PNT1,,

            ASSIGN/V1=SCN1.NUMHITS

Then change the code to refer to all the individual hits of the features

SCN2       =FEAT/SET,CARTESIAN
            THEO/<-6.539,-21.095,-5.86>,<0,0,1>
            ACTL/<-6.539,-21.095,-5.86>,<0,0,1>
            CONSTR/SET,BASIC,PLN1.HIT[1..PLN1.NUMHITS],LIN1.HIT[1..LIN1.NUMHITS],PNT1.HIT[1],,

            ASSIGN/V2=SCN2.NUMHITS

[boydha]

Thank you! that is just what I needed. learn something new everyday.

[boydha]

So I guess i just dont understand... Shouldnt this give me the same results?

SCN DAT A  =FEAT/SET,CARTESIAN
            THEO/<3.4976,0.6858,0>,<0,0,1>
            ACTL/<3.4853,0.7155,0>,<0,0,1>
            CONSTR/SET,BASIC,DATUM A.HIT[1..1],DATUM A.HIT[2..2],DATUM A.HIT[3..3],DATUM A.HIT[4..4],DATUM A.HIT[5..5],DATUM A.HIT[6..6],DATUM A.HIT[7..7],DATUM A.HIT[8..8],,

DIM PROF11= PROFILE OF SURFACE OF SET [COLOR=#0000CD]DATUM A.HIT[1..DATUM A.NUMHITS[/COLOR]]    FORMANDLOCATION  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  ARROWDENSITY=100  OUTPUT=BOTH
AX    NOMINAL       +TOL       -TOL       MEAS        DEV     OUTTOL        MAX        MIN
M       0.0000     0.0025     0.0025     0.0000     0.0000     0.0000     0.0000     0.0000 ----|----

######
OR
######

SCN DAT A  =FEAT/SET,CARTESIAN
            THEO/<3.4976,0.6858,0>,<0,0,1>
            ACTL/<3.4853,0.7155,0>,<0,0,1>
            CONSTR/SET,BASIC,DATUM A.HIT[1..1],DATUM A.HIT[2..2],DATUM A.HIT[3..3],DATUM A.HIT[4..4],DATUM A.HIT[5..5],DATUM A.HIT[6..6],DATUM A.HIT[7..7],DATUM A.HIT[8..8],,

DIM PROF11= PROFILE OF SURFACE OF SET [COLOR=#0000FF]SCN DAT A[/COLOR][COLOR=#0000FF] [/COLOR]   FORMANDLOCATION  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  ARROWDENSITY=100  OUTPUT=BOTH
AX    NOMINAL       +TOL       -TOL       MEAS        DEV     OUTTOL        MAX        MIN
M       0.0000     0.0025     0.0025     0.0001     0.0001     0.0000     0.0000    -0.0001 ----|----

I understand it is only the matter of .0001 but I am highly doubting our machinists can build a part that can hold that close of flatness/surface profile... and my question is should'nt these be the same?

[vpt.se]

You can't have spaces in the source feature name when referencing hits, that is why it isn't working for you.
Change the name of the source feature "DATUM A" to "DATUM_A" instead and reference that in the feature group.

SCN DAT A  =FEAT/SET,CARTESIAN
            THEO/<3.4976,0.6858,0>,<0,0,1>
            ACTL/<3.4853,0.7155,0>,<0,0,1>
            CONSTR/SET,BASIC,DATUM_A.HIT[1],DATUM_A.HIT[2],DATUM_A.HIT[3],DATUM_A.HIT[4],DATUM_A.HIT[5],DATUM_A.HIT[6],DATUM_A.HIT[7],DATUM_A.HIT[8],,

or...

SCN DAT A  =FEAT/SET,CARTESIAN
            THEO/<3.4976,0.6858,0>,<0,0,1>
            ACTL/<3.4853,0.7155,0>,<0,0,1>
            CONSTR/SET,BASIC,DATUM_A.HIT[1..DATUM_A.NUMHITS],,

[boydha]

Okay, ill give this a shot. Thank you!

[AndersI]

PROFILE compares each hit against its nominal value (or re-pierces the CAD if FINDNOMS is ON), not to the geometric form of the original feature. We don't know what DATUM A is, or how it is measured - it needs to be an autofeature (clicked on CAD or numerically entered) to have real nominals on every hit. If it is a "measured feature", the nominal value for each hit is equal to the measured value when the program was learned, and no form deviation will be detected in the original run. If you run the program again (preferrably on another part) you would probably get higher numbers, but still wrong.

[boydha]

That makes sense how that would be the error if I was manually probing the points and trying to output a profile (I understand how that doesnt work.) However this was to a cad model auto feature plane. I've looked through some of my older progams and concluded that the results of .0001 flatness were probable considering "Datum A" is a machined flat plane. With a span of about 2.0"x3.0" the scan was leveled and origined at the plane.

I was just wondering out of the two methods I have shown above why I am getting a different deviation /and what one would be correct.

*see the edited highlighted text above* My reasoning is that these outputs are the same 8 points therefore their output deviation should be the same. unless there is an error in how I am recalling the points in one of them.

[AndersI]

What value does FLATNESS of DATUM A give?

[lovethepirk]

Wanted to ask this question in a related thread that discussed numhits as well...

Before I knew what numhits was one of my 1st practice programs was on a simple rectangular part with a tight profile of a surface tolerance.

One of the auto features I used for the geotol was a 90 deg circle at a corner of the part and used roughly 5 hits. Most parts from the cnc would pass, but looking back now the centroid should be well within the surface points to cause a fail. Is it possible that the software recognizes you selected a feature and somewhat tries to help you out by comparing it to the CAD feature centroid? Still not the correct approach if so, but I'm just curious about this.... Or should it be just recognizing the centroid as a hit and simply comparing it to where that should be on the surface and most likely showing out of tol?

Thanks.
fillet_centroid.jpg

[JEFMAN]

Comparing the centroid to the cad depends on the number of hits, so you can't get a accurate comparison... (I guess the cad centroid is calculated from a infinity number of "hits")

average	X	Y
181 hits	31,7930475	31,7930475
91 hits	31,7551237	31,7551237
61 hits	31,7172169	31,7172169
31 hits	31,6035962	31,6035962
16 hits	31,3767761	31,3767761
11 hits	31,1504653	31,1504653
6 hits	30,4739646	30,4739646
3 hits	28,4517797	28,4517797

[lovethepirk]

JEFMAN

Thanks, interesting deviations there. That said, if you choose a feature instead of vectors and do a prof of a surface is pcdmis comparing cad centroid to hits centroid or is it comparing where the hits centroid is vs where it should be on the surface?

Thanks