Constructing a Cylinder from 2 cylinders

[YGOLORTEM]

I have an 8" shaft on the machine long ways in the Y axis. I measure a Ø 3.00" X 2" deep cylinder on one side (Y+) and a Ø 4.00 X 2" deep cylinder on the opposite side (Y-). Is it possible to create 1 cylinder (centerline) out of the 2 cylinders? I've used all possible configurations (Auto, Best Fit, BF ReComp, Projection, Cast and Reverse) and it keeps telling me there is not enough features. Is there a way possible to create this?

These 2 cylinders create a datum (Datum -A- & Datum -C- or A-C as called out on the drawing) and there is a Total Run Out called out from a different diameter on the shaft back to the A-C datum.

[ThePudds]

Rather than creating 2 different cylinders, have you tried taking multiple circles on each end and creating a cylinder that way? I am not sure if that would work. That would be the first thing I would try. I'm still new though, I do things different. Sorry if it doesn't help.

[YGOLORTEM]

Rather than creating 2 different cylinders, have you tried taking multiple circles on each end and creating a cylinder that way? I am not sure if that would work. That would be the first thing I would try. I'm still new though, I do things different. Sorry if it doesn't help.

I use auto cylinders to speed up programming. I'm thinking in the future this will be the way to go

[ThePudds]

I use auto cylinders to speed up programming. I'm thinking in the future this will be the way to go

Auto features are pretty much all I use. I was just thinking that in this scenario, a couple Auto circles on each end and construct a cylinder that way. It may work, it may not. I'm not honestly to sure. Worth a shot I suppose.

[UKCMM]

As the two diameters are different then a 3D line is what should be used

[YGOLORTEM]

Unfortunately a 3D line doesn't help with the dimension

[YGOLORTEM]

Currently I have a 3D line constructed between the 2 and it's working for everything need (except TTL RO) but I believe I need the form error from the datum's to get this call out to work

[Mike Ruff]

As the two diameters are different then a 3D line is what should be used

+1

I would construct a plane at the top and bottom of each cylinder. Then construct an intersection point from the cylinder and each plane. You should now have 4 points, one at the top and bottom of each cylinder. Now you can construct a 3D line from the 4 points and use this line.

[DAN_M]

The first time I came across an "A-B" callout on a print, Hex remoted in and showed me how to do the following:

Create a 3D line using the start & end points of your desired geometry and use that line for level & origin setting. Like so:

LIN_XP_THEO=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
CYL_A =FEAT/CONTACT/CYLINDER/DEFAULT,CARTESIAN,IN,LEAST_SQR
THEO/<0,0,0>,<0,0,1>,10,1
ACTL/<0,0,0>,<0,0,1>,10,1
TARG/<0,0,0>,<0,0,1>
START ANG=0,END ANG=360
ANGLE VEC=<1,0,0>
DIRECTION=CCW
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
CYL_B =FEAT/CONTACT/CYLINDER/DEFAULT,CARTESIAN,IN,LEAST_SQR
THEO/<0,0,2>,<0,0,1>,5,1
ACTL/<0,0,2>,<0,0,1>,5,1
TARG/<0,0,2>,<0,0,1>
START ANG=0,END ANG=360
ANGLE VEC=<1,0,0>
DIRECTION=CCW
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
PNT_CYL_A_START=GENERIC/POINT,DEPENDENT,CARTESIAN,$
NOM/XYZ,<CYL_A.EX,CYL_A.EY,CYL_A.EZ>,$
MEAS/XYZ,<CYL_A.EX,CYL_A.EY,CYL_A.EZ>,$
NOM/IJK,<0,0,1>,$
MEAS/IJK,<0,0,1>
PNT_CYL_B_END=GENERIC/POINT,DEPENDENT,CARTESIAN,$
NOM/XYZ,<CYL_B.SX,CYL_B.SY,CYL_B.SZ>,$
MEAS/XYZ,<CYL_B.SX,CYL_B.SY,CYL_B.SZ>,$
NOM/IJK,<0,0,1>,$
MEAS/IJK,<0,0,1>
LIN_AB =FEAT/LINE,CARTESIAN,UNBOUNDED,NO
THEO/<0,0,-1>,<0,0,1>
ACTL/<0,0,-1>,<0,0,1>
CONSTR/LINE,BF,3D,PNT_CYL_A_START,PNT_CYL_B_END,,
OUTLIER_REMOVAL/OFF,3
FILTER/OFF,WAVELENGTH=0
PLN_TOP =FEAT/CONTACT/PLANE/DEFAULT,CARTESIAN,NONE,LEAST_SQR
THEO/<0,0,4>,<0,0,1>
ACTL/<0,0,4>,<0,0,1>
TARG/<0,0,4>,<0,0,1>
ANGLE VEC=<1,0,0>,RADIAL
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
A1 =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
ALIGNMENT/LEVEL,ZPLUS,LIN_AB
ALIGNMENT/ROTATE,XPLUS,TO,LIN_XP_THEO,ABOUT,ZPLUS
ALIGNMENT/TRANS,XAXIS,LIN_AB
ALIGNMENT/TRANS,YAXIS,LIN_AB
ALIGNMENT/TRANS,ZAXIS,PLN_TOP
ALIGNMENT/END

Also, check this out:

https://www.pcdmisforum.com/forum/pc...ting-a-b-datum

Note:
An A-B alignment can't have a material modifier attached to it because a line is not a feature of size.

[DAN_M]

you can do whatever you want depending on your situational needs

i just like the idea of the line being forced through multiple pnts as opposed to just through two points

[Jay_Q]

Makes sense!

[DAN_M]

Cris_C has better articulated what I am trying to say in Post#12

[LostL]

Personally I exctracts the points into a set and construct a cylinder from the set. That way you get everything.

SCN1 =FEAT/SET,CARTESIAN
THEO/<9770.13,-528.611,4717.251>,<0,0,1>
ACTL/<9770.13,-528.611,4717.251>,<0,0,1>
CONSTR/SET,BASIC,CYL1.HITS[1..CYL1.NUMHITS],CYL2.HITS[1..CYL2.NUMHITS,,
CYL3 =FEAT/CYLINDER,CARTESIAN,IN,LEAST_SQR,NO
THEO/<9770.26,-528.484,4717.005>,<-0.019565,-0.1045285,-0.9943294>,6.5,7.3
ACTL/<9770.26,-528.484,4717.005>,<0.019565,0.1045285,0.9943294>,6 .5,7.3
CONSTR/CYLINDER,BF,SCN1,,

[Jay_Q]

Personally I exctracts the points into a set and construct a cylinder from the set. That way you get everything.

SCN1 =FEAT/SET,CARTESIAN
THEO/<9770.13,-528.611,4717.251>,<0,0,1>
ACTL/<9770.13,-528.611,4717.251>,<0,0,1>
CONSTR/SET,BASIC,CYL1.HITS[1..CYL1.NUMHITS],CYL2.HITS[1..CYL2.NUMHITS,,
CYL3 =FEAT/CYLINDER,CARTESIAN,IN,LEAST_SQR,NO
THEO/<9770.26,-528.484,4717.005>,<-0.019565,-0.1045285,-0.9943294>,6.5,7.3
ACTL/<9770.26,-528.484,4717.005>,<0.019565,0.1045285,0.9943294>,6 .5,7.3
CONSTR/CYLINDER,BF,SCN1,,

Out of curiosity. Why construct into a set 1st? Does it not work the same if you edit the code after constructing the cylinder?
constr/cylinder,bf,cyl1.hits[1..cyl1.numhits],cyl2.hits[1..cyl2.numhits].

also would this still work with 2 different diameter sizes for the cylinders?

[A-machine-insp]

Personally I exctracts the points into a set and construct a cylinder from the set. That way you get everything.

SCN1 =FEAT/SET,CARTESIAN
THEO/<9770.13,-528.611,4717.251>,<0,0,1>
ACTL/<9770.13,-528.611,4717.251>,<0,0,1>
CONSTR/SET,BASIC,CYL1.HITS[1..CYL1.NUMHITS],CYL2.HITS[1..CYL2.NUMHITS,,
CYL3 =FEAT/CYLINDER,CARTESIAN,IN,LEAST_SQR,NO
THEO/<9770.26,-528.484,4717.005>,<-0.019565,-0.1045285,-0.9943294>,6.5,7.3
ACTL/<9770.26,-528.484,4717.005>,<0.019565,0.1045285,0.9943294>,6 .5,7.3
CONSTR/CYLINDER,BF,SCN1,,

I do this but I don't construct a set out of it then construct the cylinder, it can all be done in the cylinder construction without constructing a featureset first. I edited your code but this is how I always do it when this type of situation arises.

CYL3 =FEAT/CYLINDER,CARTESIAN,IN,LEAST_SQR,NO
THEO/<9770.26,-528.484,4717.005>,<-0.019565,-0.1045285,-0.9943294>,6.5,7.3
ACTL/<9770.26,-528.484,4717.005>,<0.019565,0.1045285,0.9943294>,6 .5,7.3
CONSTR/CYLINDER,BF,CYL1.HITS[1..CYL1.NUMHITS],CYL2.HITS[1..CYL2.NUMHITS],,

Jay_Q The NUMHITS code will not give you the desired results using 2 different diameter cylinders. If that is the situation, I would do an auto circle at the end of each cylindrical feature, do a cast point for each circle then construct a line. Using the line, you will obviously not be able to use MMC/MMB with the line but it will help establish a centerline inline with the cylinders. If a line is used, I would recommend also doing an auto cylinder on each cylindrical feature to verify that they are inline/concentric.

[Cris_C]

+1

I would construct a plane at the top and bottom of each cylinder. Then construct an intersection point from the cylinder and each plane. You should now have 4 points, one at the top and bottom of each cylinder. Now you can construct a 3D line from the 4 points and use this line.

I too prefer a 3D line for this application. A 3D line using all 4 end points will account for any deviations between the bores when fitting to the datums. If there is a slight shift or angular difference between the centerlines of the bores, a 3D line will find the average center. It would best approximate what a shaft would do when placed through both bores.

[Jay_Q]

I do this but I don't construct a set out of it then construct the cylinder, it can all be done in the cylinder construction without constructing a featureset first. I edited your code but this is how I always do it when this type of situation arises.

CYL3 =FEAT/CYLINDER,CARTESIAN,IN,LEAST_SQR,NO
THEO/<9770.26,-528.484,4717.005>,<-0.019565,-0.1045285,-0.9943294>,6.5,7.3
ACTL/<9770.26,-528.484,4717.005>,<0.019565,0.1045285,0.9943294>,6 .5,7.3
CONSTR/CYLINDER,BF,CYL1.HITS[1..CYL1.NUMHITS],CYL2.HITS[1..CYL2.NUMHITS],,

Jay_Q The NUMHITS code will not give you the desired results using 2 different diameter cylinders. If that is the situation, I would do an auto circle at the end of each cylindrical feature, do a cast point for each circle then construct a line. Using the line, you will obviously not be able to use MMC/MMB with the line but it will help establish a centerline inline with the cylinders. If a line is used, I would recommend also doing an auto cylinder on each cylindrical feature to verify that they are inline/concentric.

I didn’t think it would work out very well with 2 diameters.

I have in the past used a cast point from each cylinder (different sizes) then constructed the 3D line. Not from the far end of each cylinder but the center essentially.

Dan M above has a good point above if you construct a point from the start and end of each. That seems like it would give you optimal results. Kind of like if you had a cylinders the same size and used numhits.

I know there is 1000 ways to skin the cat but I like getting other peoples thoughts on the best/most accurate way of doing things.

[DAN_M]

I didn’t think it would work out very well with 2 diameters.

I have in the past used a cast point from each cylinder (different sizes) then constructed the 3D line. Not from the far end of each cylinder but the center essentially.

Dan M above has a good point above if you construct a point from the start and end of each. That seems like it would give you optimal results. Kind of like if you had a cylinders the same size and used numhits.

I know there is 1000 ways to skin the cat but I like getting other peoples thoughts on the best/most accurate way of doing things.

The only other way to break my way down EVEN more would be to NOT use cylinders...use circles. Then you could force your line through all of those center points at each level (now you've got an 8 or ten pnt 3d line vs one forced through the 4 start & end pnts)

[A-machine-insp]

I didn’t think it would work out very well with 2 diameters.

I have in the past used a cast point from each cylinder (different sizes) then constructed the 3D line. Not from the far end of each cylinder but the center essentially.

Dan M above has a good point above if you construct a point from the start and end of each. That seems like it would give you optimal results. Kind of like if you had a cylinders the same size and used numhits.

I know there is 1000 ways to skin the cat but I like getting other peoples thoughts on the best/most accurate way of doing things.

DAN_M Does make a good point, as rare as it is (JK) . I didn't think about a 3d line between 4 points to get the average of the points. I would still do 2 individual cylinders as well to verify concentricity/inline condition though but just as a verification. You could dimension the cylinders as verification but set the dimension output to none so it does not show up in the report or any dimensional output that you may be using. It will only show up in the code in the edit window.