how would I constrain 6 DOF of a cylindrical part with no clocking features?

[ProbFlex]

Hi Guys !

i have been scratching my head on this one. How would I make sure to constrain rotation ( All 6 DOF) even if I am measuring a cylindrical part with no clocking features?

[JacobCheverie]

You could use one or both of the faces of the part to constrain translation along the axis of the cylinder and you can construct an "Alignment" line that will arbitrarily constrain rotation about the axis of the cylinder.

The cylindrical part itself will naturally constrain the other 4 DOF.

[ProbFlex]

sorry to ask but, could you please put that in CODE. i am a visual person. thank you.

[JacobCheverie]

cyl1 =feat/contact/cylinder/default,cartesian,in,least_sqr
theo/<0.187,-6.64039,0>,<0,-1,0>,0.188,2
actl/<0.187,-6.64039,0>,<0,-1,0>,0.188,2
targ/<0.187,-6.64039,0>,<0,-1,0>
start ang=0,end ang=360
angle vec=<1,0,0>
direction=ccw
show feature parameters=no
show contact parameters=no
pln1 =feat/contact/plane/default,cartesian,none,least_sqr
theo/<0.187,-6.64039,0>,<0,-1,0>
actl/<0.187,-6.64039,0>,<0,-1,0>
targ/<0.187,-6.64039,0>,<0,-1,0>
angle vec=<1,0,0>,square
show feature parameters=no
show contact parameters=no
lin3 =feat/line,cartesian,unbounded,yes
theo/<-0.5,0,0>,<1,0,0>
actl/<-0.5,0,0>,<1,0,0>
constr/line,align,1,xplus
a1 =alignment/start,recall:auto_2,list=yes
alignment/level,yminus,cyl1
alignment/rotate,xplus,to,lin3,about,yminus
alignment/trans,xaxis,cyl1
alignment/trans,zaxis,cyl1
alignment/trans,yaxis,pln1
alignment/end

[DAN_M]

CONSTRUCT a line, select ALIGNMENT in the dropdown box, click OK. When the code is in your edit window, OVERRIDE the vectors and make it the direction you want. In my example below, I am leveling to the Z+, which means my rotation line MUST have a "X" or "Y" vector, I chose XPLUS (1,0,0).

Like so:

WORKPLANE/ZPLUS
LIN_WCS_XP_POS1 =FEAT/LINE,CARTESIAN,UNBOUNDED,NO
THEO/<-0.5,0,0>,<1,0,0>
ACTL/<-0.5,0,0>,<1,0,0>
CONSTR/LINE,ALIGN,1,XPLUS
MOVE/CLEARPLANE
CIRXY1 =FEAT/CONTACT/CIRCLE/DEFAULT,CARTESIAN,IN,LEAST_SQR
THEO/<0,0,2.6>,<0,0,1>,1.85075
ACTL/<0,0,2.6>,<0,0,1>,1.85075
TARG/<0,0,2.6>,<0,0,1>
START ANG=0,END ANG=360
ANGLE VEC=<1,0,0>
DIRECTION=CCW
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
FLY/ON
MOVESPEED/ 270
TOUCHSPEED/ 5
PREHIT/0.125
RETRACT/0.125
CHECK/0.125,1
CIRXY2 =FEAT/CONTACT/CIRCLE/DEFAULT,CARTESIAN,IN,LEAST_SQR
THEO/<CIRXY1.TX,CIRXY1.TY,2.6>,<0,0,1>,1.85075
ACTL/<0,0,2.6>,<0,0,1>,1.85075
TARG/<CIRXY1.X,CIRXY1.Y,2.6>,<0,0,1>
START ANG=-120,END ANG=240
ANGLE VEC=<1,0,0>
DIRECTION=CCW
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
MOVE/INCREMENT,<0,0,1>
PLNZ1 =FEAT/CONTACT/PLANE/DEFAULT,CARTESIAN,NONE,LEAST_SQR
THEO/<CIRXY2.TX,CIRXY2.TY,3.027>,<0,0,1>
ACTL/<0,0,3.027>,<0,0,1>
TARG/<CIRXY2.X,CIRXY2.Y,3.027>,<0,0,1>
ANGLE VEC=<-0.8944272,0.4472136,0>,RADIAL
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
CLEARP/ZPLUS,PLNZ1.Z+0.5,ZPLUS,PLNZ1.Z+0.5,ON
A1 =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
ALIGNMENT/LEVEL,ZPLUS,PLNZ1
ALIGNMENT/ROTATE,XPLUS,TO,LIN_WCS_XP_POS1,ABOUT,ZPLUS
ALIGNMENT/TRANS,XAXIS,CIRXY2
ALIGNMENT/TRANS,YAXIS,CIRXY2
ALIGNMENT/TRANS,ZAXIS,PLNZ1
ALIGNMENT/END

I have given you examples of this before, if you need anything re-explained reach out my friend! You've got the #

[ProbFlex]

Dan, my homie..
I have text you on your company #. are you on vacation still? hope your doing good!

[Mike Ruff]

Based on my personal experience, it seems unnecessary to constrain the final rotation of turned parts. I leave the 6th degree of freedom unconstrained and PC-DMIS seems to treat it as though it is clocked to the CMM axes. People say that you should always have rotation constrained, but for the special case of turned parts, I have never encountered a single problem with only 5 DOFs constrained... Just my .02

Let me add that I always program turned parts parallel to the X, Y, or Z axis of the CMM! If your part is at a compound angle, you will probably need to constrain all 6 DOFs

[sealevel]

I leave the 6th degree of freedom unconstrained and PC-DMIS seems to treat it as though it is clocked to the CMM axes.

Fair enough. Yet, a rotation generic line (or Dan's constructed) costs nothing, and removes any doubt of the possibility Demon acting up without 6 DoF. (unless in a loop)
my 2 pesos

[bfire85]

I create a generic line, like Dan, that aligns the part with the CMM's axis.

[dph51]

Create a generic line.

[dph51]

Fair enough. Yet, a rotation generic line (or Dan's constructed) costs nothing, and removes any doubt of the possibility Demon acting up without 6 DoF. (unless in a loop)
my 2 pesos

+1

[dph51]

cyl1 =feat/contact/cylinder/default,cartesian,in,least_sqr
theo/<0.187,-6.64039,0>,<0,-1,0>,0.188,2
actl/<0.187,-6.64039,0>,<0,-1,0>,0.188,2
targ/<0.187,-6.64039,0>,<0,-1,0>
start ang=0,end ang=360
angle vec=<1,0,0>
direction=ccw
show feature parameters=no
show contact parameters=no

pln1 =feat/contact/plane/default,cartesian,none,least_sqr
theo/<0.187,-6.64039,0>,<0,-1,0>
actl/<0.187,-6.64039,0>,<0,-1,0>
targ/<0.187,-6.64039,0>,<0,-1,0>
angle vec=<1,0,0>,square
show feature parameters=no
show contact parameters=no

! create a generic line, type alignment
lin3 =feat/line,cartesian,unbounded,yes
theo/<-0.5,0,0>,<1,0,0>
actl/<-0.5,0,0>,<1,0,0>
constr/line,align,1,xplus

a1 =alignment/start,recall:auto_2,list=yes
alignment/level,yminus,cyl1
alignment/rotate,xplus,to,lin3,about,yminus
alignment/trans,xaxis,cyl1
alignment/trans,zaxis,cyl1
alignment/trans,yaxis,pln1
alignment/end

+1 JacobCheverie

I just highlighted the code in red to make it plain to see.

[sealevel]

Forgot you could use the Rayco magnetic rotation vector serpent. Calibrated to 90 deg increments. (shown used at -1,0,0)

rayco_vector_snake.jpg

dph51 ... Am I wrong or did you mention you were retiring?

[vpt.se]

Very hard to find batteries for though...

[JEFMAN]

In your code, the theo plane vector is the same than the cylinder one, so you can't use the second as a rotation. In this case, the demon will project the cylinder actual axis on the plane, and use the projection as x axis.
It can't be accurate or repeatable.
You have to choose leveling between plane and cylinder (I would use the biggest surface between them, or look at the datums, if they exist)
I would also measure a line along a cmm axis to ensure that the head rotation is in the right alignment. Generic line could work also, only if you create it in the start up alignment.
In addition, I would create a pierce point between cylinder and plane to set the origin.

[dwade]

One thing that has not been brought up about constraining all 6 DOFs is that PC-DEMON CAN hose your nominals throughout your program if you do not rotate (even if it is done by creating the generic line as mentioned above)

ALWAYS construct your alignments in this order
Level
rotate
origin
translate (if needed)

[LaserJay]

You don't,
You can align without a rotate. I know technically you will be deviating from the golden rule of controlling all DOF but with a part that has no rotational features adding a rotate would be just a formality, you really don't need it.

[LaserJay]

In this case I would

LVL Z
ORG Z
ORG X
ORG Y

[neil.challinor]

You do need to constrain all 6 DOF. I've attached a (very old) tech sheet explaining everything.

[JacobCheverie]

Thanks for that Neil, I didn't know that an alignment line was not sufficient and that a plane should be used instead.