ISO 1101 / ASME and Iterative alignment

[Robert Hulman]

Hello Guys,

Can anyone tell me how exactly iterative alignment calculates aligmnent? Is it a kind of BestFit?

Is Iterative alignment according to standard ISO 1101 or ASME??

Thanks

[NinjaBadger]

It's not really a best fit.

What it's doing is making certain features right in certain axis.

Take the following example...

3 points on three separate nominally parallel (but not co-planer) faces

2 points on two separate nominally parallel (but not co-planer) faces (which are nominally perpendicular to the previous three faces

1 point on a face which is nominally perpendicular to the previous faces.

In effect it creates an offset plane from the first three points and levels to that

It then creates an offset line from the two points and rotates to that

It then origins on the final point in the remaining axis (but offsets it by its theoretical value)

See below for code

A1         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<79.2669,70.5993,[COLOR=#FF0000][B]6.35[/B][/COLOR]>,<0,0,1>
            ACTL/<79.2669,70.5993,6.35>,<0,0,1>
            TARG/<79.2669,70.5993,6.35>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
A2         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<108.3136,7.7473,[COLOR=#FF0000][B]0[/B][/COLOR]>,<0,0,1>
            ACTL/<108.3136,7.7473,0>,<0,0,1>
            TARG/<108.3136,7.7473,0>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
A3         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<174.4408,90.0379,[COLOR=#FF0000][B]-7[/B][/COLOR]>,<0,0,1>
            ACTL/<174.4408,90.0379,-7>,<0,0,1>
            TARG/<174.4408,90.0379,-7>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
B1         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<76.0992,[COLOR=#0000FF][B]6[/B][/COLOR],-4.9737>,<0,-1,0>
            ACTL/<76.0992,6,-4.9737>,<0,-1,0>
            TARG/<76.0992,6,-4.9737>,<0,-1,0>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
B2         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<201.4813,[COLOR=#0000FF][B]20[/B][/COLOR],-6.0981>,<0,-1,0>
            ACTL/<201.4813,20,-6.0981>,<0,-1,0>
            TARG/<201.4813,20,-6.0981>,<0,-1,0>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
C1         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<[COLOR=#FF8C00][B]239[/B][/COLOR],3.5235,-27.6121>,<1,0,0>
            ACTL/<239,3.5235,-27.6121>,<1,0,0>
            TARG/<239,3.5235,-27.6121>,<1,0,0>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=BOTH,DISTANCE=10
            SHOW HITS=NO
PLN_A      =FEAT/PLANE,CARTESIAN,TRIANGLE,NO
            THEO/<120.6737,56.1281,0>,<0,0,1>
            ACTL/<120.6737,56.1281,0>,<0,0,1>
            CONSTR/PLANE,OFFSET
            ID = A1,A2,A3,,
            OFFSET = [B][COLOR=#FF0000]-6.35,0,7[/COLOR][/B]
LIN_B      =FEAT/LINE,CARTESIAN,UNBOUNDED,NO
            THEO/<76.0992,0,-5.5359>,<1,0,0>
            ACTL/<76.0992,0,-5.5359>,<1,0,0>
            CONSTR/LINE,OFFSET
            SURFACE NORMAL = <0,0,1>,MULTI POINT
            ID = B1,B2,,
            OFFSET = [B]-[COLOR=#0000FF]6[/COLOR][/B][COLOR=#0000FF],-[B]20[/B][/COLOR]
A4         =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
              ALIGNMENT/LEVEL,ZPLUS,PLN_A
              ALIGNMENT/TRANS,ZAXIS,PLN_A
              ALIGNMENT/ROTATE,XPLUS,TO,LIN_B,ABOUT,ZPLUS
              ALIGNMENT/TRANS,YAXIS,LIN_B
              ALIGNMENT/TRANS,XAXIS,C1
              ALIGNMENT/TRANS_OFFSET,XAXIS,-[COLOR=#FF8C00][B]239[/B][/COLOR]
            ALIGNMENT/END

[Robert Hulman]

Thanks for explanation. But is this alignment according with standards ISO 1101 OR ASME?

[NinjaBadger]

I'm not 100% but I believe they both cover iterative alignments - I know I've seen examples in both where the datum target symbols are used.

[Liambo734]

Thanks for explanation. But is this alignment according with standards ISO 1101 OR ASME?

Can you share the text in ISO or ASME that explains how an RPS or Iterative alignment should be performed, according to each standard?

[VinniUSMC]

I'm not aware of what it means to be "according to ASME" either.

As NinjaBadger wrote up, iterative is not exactly best fit. But, when you use more than the minimum number of features, it does best fit to those features.

And, it gets more complicated when the features aren't normal to X, Y, or Z axis. It still fits the feature in only 1 axis, even if it's 45 degrees between, for example, X and Y. You have to pick 1 axis, and it puts that axis to 0, and allows the other axis to have deviation.

[vpt.se]

ISO1101 covers GD&T and nothing about alignments.

[Robert Hulman]

Yes you,re right. I FCF, maybe GPS?? My problem is that our supplier uses iterative alignment and i prove them they are not measuring according to our drawings which are according to Iso1101. We have bad results and they have good ones. These parts cause problems in our production but they think their measurements are ok

[Matthew D. Hoedeman]

IF they are not using the correct datums per the print, then yeah, they are wrong.

[Liambo734]

Generally speaking the print should have set datum target points when using iterative alignments. That way you can input the XYZ IJK nominal values from the print into your points used in the iterative alignment. If your supplier is just making these up or using an iterative alignment instead of the correct alignment for each print then they are wrong. It's not based of a standard, it's what is required as an alignment on each of the drawings. HTH

[Robert Hulman]

This is special case where datum A is plane, B is hole and C is plane in angle of 6,5 but perpendicular to A. And theres tp of C to dátum A,B. But whole pat is rotated to C. I found this case where it wants tp of plane which part is rotated to in gdt book, so this case is bot a mistake in drawing

[NinjaBadger]

Could you post the relevant part of the drawing, or something done by hand which conveys the same information?

[Liambo734]

This is special case where datum A is plane, B is hole and C is plane in angle of 6,5 but perpendicular to A. And theres tp of C to dátum A,B. But whole pat is rotated to C. I found this case where it wants tp of plane which part is rotated to in gdt book, so this case is bot a mistake in drawing

So why would you use an iterative alignment when you have a Plane, Plane & Cylinder called out as Datum's?

[Robert Hulman]

That's what i m trying to find out. Supplier uses iterative alignment from minimum points which are lying on these datums. I use fcf alignment. I want to prove that they reasults are not according to drawing which is iso1101.

[vpt.se]

Alignment can be one thing, the evaluation another. What I am trying to say is you can align in one way (not according to the datums) for measurement purposes but the evaluations needs to be done according to the datum structure specified in the FCF.

Have you asked your supplier to do it the way you do it, to see if they see the same error?

A specific use for iterative is explained by Liambo734 above - datum targets, but that doesn't mean you cannot use iterative when not having datum targets.

[KIRBSTER269]

you're so smart

[vpt.se]

Thank you, I'll be here all week

[Ncogneto]

You both should use XACT measure and see what the results are.

[Robert Hulman]

draw.jpg

[NinjaBadger]

Yep - Level A (and origin one axis)
Origin x2 Axis in B
Rotate to C

When you report the TP the Basic/Boxed/TED in the dimension should be in one axis only with a Nominal of 16.6

[Robert Hulman]

yes, I know it should be in one axis only. It is a plane, it cant be dimensioned in two axis. Problem is that with AB only, there are no all 6 degrees of freedom locked. With no rotation locked i have nothing to rotate from by 6.5. And also xactmeasure takes rotation from alignment when used datums dont lock part in all 6 deg of freedom. and part is rotated to datum C, so result wil have no deviation. It wants to rotate by 6.5 from itself.

[NinjaBadger]

Nope.

Don't rotate from B to C, just rotate to the vector of C.

For examples sake, if the part was set on CMM as in top-left view (looking down on part), Level Z on Dat A and Origin Z.
Origin X and Y on Dat B
Rotate Y Minus to Dat C

Report Position of Dat C (Y axis only) - Unless part is perfect it will have some deviation on the [16.6] boxed dimension.

https://www.dropbox.com/s/ocy4ysqiwu8batd/Alignment%20Question.png?dl=0

[Robert Hulman]

Yes it has some deviation but only 0.001-0.003 but it is not real dev because that part is bad. But when i choose option snandard iso1101 to custom in FCF and i uncheck fit to datums, when it shows deviation closer to real condition. But i don't think this is the best way.