Progressive Alignment

[angerji]

Good afternoon,

I took the PC DMIS 101 class last september and in the alignments section there is a sub section that briefly covers "Progressive Alignments" and how they are required to be used if you are going to use 2D features in an alignment. Now i know its not possible to look through every single post on this forum but i have not once seen anybody reference this "progressive" style alignment. Also when i do a google search of "Progressive alignment pc dmis forum" i get literally NO results.

So whats the deal with this topic, is it just hooplah?

[acgarcia]

Its referred to here at 3-2-1 alignment or plane, line, point alignment.

Alignments in general are required to measure your part.

The 3 2 1 alignment is just a quick align before you probe your datums or more critical alignment features.

If you want to use a manual 3 2 1 and a DCC 3 2 1 you can as well also.

[angerji]

I understand alignments are required to measure parts and all that, i more specifically was just wondering what the deal was with being called a "progressive alignment" in hexagons official training material.

[acgarcia]

maybe because its faster to say 321 than to say progressive?

[Sora5]

Level, rotate, translate. Always lock all 6 DOF or else the demon will play its tricks on you.

[Mike Ruff]

Hexagon's training material is very helpful, but there is more than one way to skin a cat. The main purpose of an alignment is to lock down all 6 DOF on the part in a way that is accurate and repeatable. I don't have a copy of the 101 training material and I'm sure a lot of other people on here also do not. We can't help and try to explain something to you if we aren't sure what you are talking about. Can you explain what the subsection is talking about? The more info you provide, the easier it will be to provide an explanation

[dph51]

Just remember this helpful little phrase..........



”Liberals promote progressive alignments, conservatives don’t.”

[WM321]

Possibly referring to doing your 3-2-1 alignments as you probe them, so level straight after you probe the plane, rotate straight after you probe the line etc., rather that probing all elements and then doing 3-2-1 in one alignment.

[kingsld1]

Could be talking about an iteritive alignment. (Spelling sucks) That didn't use to be covered in Level 1. I learned about it in a Level 2 class but thats been a while. I could see it being refered to as a 'progressive alignment.

[ChairsWithWheels]

The progressive alignment is aligning each feature after you hit it. After taking 3-4 hits on PLN1, leveling to that. Then probing your line, which would be level with your plane (If you make it that way), rotating to the line. Then taking a point to translate to. With that all 6 ROF are constrained. I've never used this style. Honestly I don't know the benefits of using it. Besides using 2-D features for an alignment.

[louisd]

Yeah what ChairsWithWheels said.
Measure plane
A1 Align/level plane
-----Align/origin plane
Measure line
A2 Align/Recall A1
----Align/Rotate to line
----Align/origin to line
Measure point
A3 Align/Recall A2
-----Align/origin to point

A3 now controls all 6 DOF.

The benefit is that the data points you collect help correct any vector error in your probed initial alignment... making for a more accurate and more robust initial align.
If you set a very prudent prehit/retract, (ex: 0.250" for large parts, or for smaller parts maybe 0.150" will do) this can essentially find the part on the machine within a solid 1/2" from where it is supposed to be. If you have lazy peers or operators, it should still run the part without error the first go, even if they put part on one-handed while surfing FB with their phone in the other hand...

[A-machine-insp]

I have lazy peers. They hate my read point alignments cause they have to take an extra 10 seconds to align the part on the CMM.

[AndersI]

If you use lines and/or circles in your alignment, you need to do the level before measuring the lines and/or circles, as they are 2D features and take the info for the 3:rd dimension from the current alignment. This is mostly not a large problem on a CMM where you put the part on a table with a known [good] relation to the machine axes, but with a measuring arm (for example) there is no known correlation between the part support and the measuring unit, and the resulting errors in circles and lines measured before levelling can be substantial.

[Matthew D. Hoedeman]

Yeah what ChairsWithWheels said.
Measure plane
A1 Align/level plane
-----Align/origin plane
Measure line
A2 Align/Recall A1
----Align/Rotate to line
----Align/origin to line
Measure point
A3 Align/Recall A2
-----Align/origin to point

A3 now controls all 6 DOF.

The benefit is that the data points you collect help correct any vector error in your probed initial alignment... making for a more accurate and more robust initial align.
If you set a very prudent prehit/retract, (ex: 0.250" for large parts, or for smaller parts maybe 0.150" will do) this can essentially find the part on the machine within a solid 1/2" from where it is supposed to be. If you have lazy peers or operators, it should still run the part without error the first go, even if they put part on one-handed while surfing FB with their phone in the other hand...

The only issue with what you show is that tech support did confirm (a number of years ago) that if you didn't have ONE alignment that locked all 6 DOF in a single alignment, there could be a nominal-change issue with your program. So, instead of 3 alignments that only lock a partial, keep locking the previous DOF when you make a new alignment

measure a plane
(1) Level & origin
measure a line
(2) Level & rotate, origin, origin
measure a point
(3) level, rotate, origin, origin, origin

I pointed out that their 'quick start' alignments were making alignments that could cause that very issue, but so far, I've not seen them fix this, and this would be a pretty simple fix for them to make.

[InspectorJester]

Interesting. I didn't know not locking all 6 DOF at least once could change nominals (I've had that happen in the past, more than likely due to that very reason). I know to do that, of course, but that would help with troubleshooting..
+1

[louisd]

Yeah typically I recall startup once all 6DOF are captured, then lock down all 6DOF within the ONE alignment definition. I didn't post that, as it wouldn't exemplify the concept of "progressive alignment" as the OP was inquiring about. Good to know that this could lead to nominals dancing however. I would think recalling previous aligns until all 6dof are controlled would be equivalent, but the demon never ceases to impress/shock me.

[InspectorJester]

Agreed; finicky machines!
I do the same; recall startup and lock all 6DOF in 1 alignment. I typically perform iterative or "progressive" alignments, and have a final "fine alignment", which recalls startup. At least this way I can see that they are all, in fact, locked down

[AndersI]

I pointed out that their 'quick start' alignments were making alignments that could cause that very issue, but so far, I've not seen them fix this, and this would be a pretty simple fix for them to make.

This is fixed in the new "Quick Align" - i just tested (offline) by clicking four points on a plane, two on a line and finally a single point, then clicked Quick Align. This is what it generated:

MAN_ALIGN  =GROUP/SHOWALLPARAMS=YES,EXECUTION CONTROL=AS MARKED

PLN1         =FEAT/PLANE,CARTESIAN,TRIANGLE
              THEO/<122.116,57.782,0>,<0,0,1>
              ACTL/<122.116,57.782,0>,<0,0,1>
              MEAS/PLANE,4
                HIT/BASIC,NORMAL,<93.037,96.548,0>,<0,0,1>,<93.037,96.548,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<150.774,92.431,0>,<0,0,1>,<150.774,92.431,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<163.567,26.269,0>,<0,0,1>,<163.567,26.269,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<81.088,15.88,0>,<0,0,1>,<81.088,15.88,0>,USE THEO=YES
              ENDMEAS/

A1           =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1
                ALIGNMENT/TRANS,ZAXIS,PLN1
              ALIGNMENT/END

LIN1         =FEAT/LINE,CARTESIAN,UNBOUNDED
              THEO/<95.851,0,-7.822>,<1,0,0>
              ACTL/<95.851,0,-7.822>,<1,0,0>
              MEAS/LINE,2,ZPLUS
                HIT/BASIC,NORMAL,<95.851,0,-6.955>,<0,-1,0>,<95.851,0,-6.955>,USE THEO=YES
                HIT/BASIC,NORMAL,<173.359,0,-8.689>,<0,-1,0>,<173.359,0,-8.689>,USE THEO=YES
              ENDMEAS/

A2           =ALIGNMENT/START,RECALL:A1,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1
                ALIGNMENT/TRANS,ZAXIS,PLN1
                ALIGNMENT/ROTATE,XPLUS,TO,LIN1,ABOUT,ZPLUS
                ALIGNMENT/TRANS,YAXIS,LIN1
              ALIGNMENT/END

PNT1         =FEAT/POINT,CARTESIAN
              THEO/<0,10.933,-27.229>,<-1,0,0>
              ACTL/<0,10.933,-27.229>,<-1,0,0>
              MEAS/POINT,1,WORKPLANE
                HIT/BASIC,NORMAL,<0,10.933,-27.229>,<-1,0,0>,<0,10.933,-27.229>,USE THEO=YES
              ENDMEAS/

A3           =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1
                ALIGNMENT/TRANS,ZAXIS,PLN1
                ALIGNMENT/ROTATE,XPLUS,TO,LIN1,ABOUT,ZPLUS
                ALIGNMENT/TRANS,YAXIS,LIN1
                ALIGNMENT/TRANS,XAXIS,PNT1
              ALIGNMENT/END

            ENDGROUP/ID=MAN_ALIGN

DCC_ALIGN  =GROUP/SHOWALLPARAMS=YES,EXECUTION CONTROL=AS MARKED
              COMMENT/OPER,NO,FULL SCREEN=NO,AUTO-CONTINUE=NO,
              Changing to DCC mode! Do not continue unless avoidance moves have been added.

              MODE/DCC
              WORKPLANE/ZPLUS

PLN1_DCC     =FEAT/PLANE,CARTESIAN,TRIANGLE
              THEO/<122.116,57.782,0>,<0,0,1>
              ACTL/<122.116,57.782,0>,<0,0,1>
              MEAS/PLANE,4
                HIT/BASIC,NORMAL,<93.037,96.548,0>,<0,0,1>,<93.037,96.548,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<150.774,92.431,0>,<0,0,1>,<150.774,92.431,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<163.567,26.269,0>,<0,0,1>,<163.567,26.269,0>,USE THEO=YES
                HIT/BASIC,NORMAL,<81.088,15.88,0>,<0,0,1>,<81.088,15.88,0>,USE THEO=YES
              ENDMEAS/

A3_DCC1      =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1_DCC
                ALIGNMENT/TRANS,ZAXIS,PLN1_DCC
                ALIGNMENT/ROTATE,XPLUS,TO,LIN1,ABOUT,ZPLUS
                ALIGNMENT/TRANS,YAXIS,LIN1
                ALIGNMENT/TRANS,XAXIS,PNT1
              ALIGNMENT/END

LIN1_DCC     =FEAT/LINE,CARTESIAN,UNBOUNDED
              THEO/<95.851,0,-7.822>,<1,0,0>
              ACTL/<95.851,0,-7.822>,<1,0,0>
              MEAS/LINE,2,ZPLUS
                HIT/BASIC,NORMAL,<95.851,0,-6.955>,<0,-1,0>,<95.851,0,-6.955>,USE THEO=YES
                HIT/BASIC,NORMAL,<173.359,0,-8.689>,<0,-1,0>,<173.359,0,-8.689>,USE THEO=YES
              ENDMEAS/

A3_DCC2      =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1_DCC
                ALIGNMENT/TRANS,ZAXIS,PLN1_DCC
                ALIGNMENT/ROTATE,XPLUS,TO,LIN1_DCC,ABOUT,ZPLUS
                ALIGNMENT/TRANS,YAXIS,LIN1_DCC
                ALIGNMENT/TRANS,XAXIS,PNT1
              ALIGNMENT/END

PNT1_DCC     =FEAT/POINT,CARTESIAN
              THEO/<0,10.933,-27.229>,<-1,0,0>
              ACTL/<0,10.933,-27.229>,<-1,0,0>
              MEAS/POINT,1,WORKPLANE
                HIT/BASIC,NORMAL,<0,10.933,-27.229>,<-1,0,0>,<0,10.933,-27.229>,USE THEO=YES
              ENDMEAS/

A3_DCC3      =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
                ALIGNMENT/LEVEL,ZPLUS,PLN1_DCC
                ALIGNMENT/TRANS,ZAXIS,PLN1_DCC
                ALIGNMENT/ROTATE,XPLUS,TO,LIN1_DCC,ABOUT,ZPLUS
                ALIGNMENT/TRANS,YAXIS,LIN1_DCC
                ALIGNMENT/TRANS,XAXIS,PNT1_DCC
              ALIGNMENT/END

            ENDGROUP/ID=DCC_ALIGN

[QA NINJA]

How do you start your alignment to where it will pull in a DCC automatically? why does yours say engroup?

[acgarcia]

He has his manual align and dcc align in groups. Groups are like exapndable folders, You can store program code in them and then select the group to not show and it hides your code from the edit window. Its really good for organizing and making it easier on the eyes

[acgarcia]

You need to have done your manual alignment at lest once and have the part located in a repeatable fixture to skip the manual alignment features. For example, once you do your manaul alignment "A1", you can enter the goto/ command lets say for practice, right after your probe and tip selection. GOTO/A1 and your program will start at A1 which is your manual alignment, That is the rough location of your part, then you go MODE/DCC, maybe a movepoint and then start probing your dcc alignment features.

[Roberto]

If you use lines and/or circles in your alignment, you need to do the level before measuring the lines and/or circles, as they are 2D features and take the info for the 3:rd dimension from the current alignment. This is mostly not a large problem on a CMM where you put the part on a table with a known [good] relation to the machine axes, but with a measuring arm (for example) there is no known correlation between the part support and the measuring unit, and the resulting errors in circles and lines measured before levelling can be substantial.

This is a good point.
I always do a level-rotate-origin in a single block. It never gives any issues. But portable arm it's a very different animal.

[Matthew D. Hoedeman]

is the 'quick align' different from the "view ~ other windows ~ quick start" option and selecting the alignment option you want to make? (plane line line, plane circle circle, plane line point, etc.)

[AndersI]

is the 'quick align' different from the "view ~ other windows ~ quick start" option and selecting the alignment option you want to make? (plane line line, plane circle circle, plane line point, etc.)

Yes, it's a button in the tool bars (and a menu choice). I don't think the "Quick Start" part of PC-DMIS has seen much new features for a long time.