Roundness issue

[jacyEd]

I noticed that every feature i measure at 30º indicates .001" minimum out of round on the CMM. I double check with manual gauging and find it to actually be tenths. I run 3.7 MR3 with no CAD files on a RS-670 with PH10MQ and TP200b. I have also had this same issue using a TP200 and any of 6 modules i have on hand reguardless of stylus size. Anyone else see an issue like this or have any suggestions as to what i can do?

[craiger_ny]

3D feature or 2D? The reason I ask is if you are not in a level scenario that sets you up well to measure normal to a 2D feature then your vector error may show as a roundness problem. First susspicion should be things like part and styli cleanliness. Also what type of manual gage are you using 2 point or 3 point contact and what kind of process is generating that feature and how many hits are you taking? ? Lastly a CMM is not the best for roundenss but I think it is possible to be more effective than what you are getting if you take all these things into account and prepare for them.

[John Kingston]

To add to Craig's comments, by 30 deg, do you mean the diameter is at 30 deg to your workplane? That could cause a roundness issue. Can you post the section of code for us to peruse?

[jacyEd]

Here are the last 4 points of a 36 point pattern at 30º probe angle. Also should be noted that i use a 100mm extention and TP200b with standard force module and 2BY30MM stylus in this program. The Datums were set in DCC as well.

MOVE/POINT,NORMAL,6.0324,-3.4189,5.4531
TIP/T1A30B-90, SHANKIJK=0.5, 0, 0.866, ANGLE=-90.0005
PNT105 =FEAT/POINT,RECT
THEO/2.5247,-2.1185,5.4531,0.7660442,-0.6427879,0
ACTL/2.5248,-2.1095,5.4533,0.7660442,-0.6427879,0
MEAS/POINT,1
HIT/BASIC,NORMAL,2.5247,-2.1185,5.4531,0.7660442,-0.6427879,0,2.5248,-2.1095,5.4533,USE THEO = YES
ENDMEAS/
PNT106 =FEAT/POINT,RECT
THEO/2.8543,-1.6479,5.4531,0.8660252,-0.5000004,0
ACTL/2.8532,-1.6409,5.4534,0.8660252,-0.5000004,0
MEAS/POINT,1
HIT/BASIC,NORMAL,2.8543,-1.6479,5.4531,0.8660252,-0.5000004,0,2.8532,-1.6409,5.4534,USE THEO = YES
ENDMEAS/
PNT107 =FEAT/POINT,RECT
THEO/3.0971,-1.1272,5.4531,0.9396925,-0.3420205,0
ACTL/3.0955,-1.1229,5.4536,0.9396925,-0.3420205,0
MEAS/POINT,1
HIT/BASIC,NORMAL,3.0971,-1.1272,5.4531,0.9396925,-0.3420205,0,3.0955,-1.1229,5.4536,USE THEO = YES
ENDMEAS/
PNT108 =FEAT/POINT,RECT
THEO/3.2457,-0.5723,5.4531,0.9848077,-0.1736486,0
ACTL/3.2447,-0.5703,5.4534,0.9848077,-0.1736486,0
MEAS/POINT,1
HIT/BASIC,NORMAL,3.2457,-0.5723,5.4531,0.9848077,-0.1736486,0,3.2447,-0.5703,5.4534,USE THEO = YES
ENDMEAS/
DIA -8- =FEAT/CIRCLE,RECT,OUT,LEAST_SQR
THEO/0,0,5.4531,0,0,1,6.5918
ACTL/0,0.0092,5.4529,0,0,1,6.592
CONSTR/CIRCLE,BFRE,PNT73,PNT74,PNT75,PNT76,PNT77,PNT78,PN T79,PNT80,PNT81,PNT82,PNT83,PNT84,PNT85,PNT86,PNT8 7,PNT88,PNT89,PNT90,PNT91,PNT92,PNT93,PNT94,PNT95, PNT96,PNT97,PNT98,PNT99,PNT100,PNT101,PNT102,PNT10 3,PNT104,PNT105,PNT106,PNT107,PNT108,$
PNT108,,
OUTLIER_REMOVAL/OFF,3
FILTER/OFF,UPR=0
DIM DIAMETER -8- MIN. 6.5913)= LOCATION OF CIRCLE DIA -8- UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=1.00 OUTPUT=BOTH
AX MEAS NOMINAL +TOL -TOL DEV OUTTOL
D 6.5920 6.5918 0.0000 0.0000 0.0002 0.0002 -------->
RN 0.0011 0.0000 0.0000 0.0000 0.0011 0.0011 -------->
END OF DIMENSION DIAMETER -8- MIN. 6.5913)
DIM DIAMETER -8- (REPAIR -07)= LOCATION OF CIRCLE DIA -8- UNITS=IN ,$
GRAPH=OFF TEXT=OFF MULT=1.00 OUTPUT=BOTH
AX MEAS NOMINAL +TOL -TOL DEV OUTTOL
D 6.5920 6.5918 0.0005 0.0005 0.0002 0.0000 ------#--
RN 0.0011 0.0000 0.0000 0.0000 0.0011 0.0011 -------->
END OF DIMENSION DIAMETER -8- (REPAIR -07)

[jacyEd]

I actually just realized that some of my A0B0 patterns display the same out of round. This part has been checked on a runout table using dial indicators, showing less than .0002" out of round.

[John Kingston]

Is the hole at an angle to your datums? If so, the circle doesn't show that. Your vectors show this to be perpendicular.

[jacyEd]

Its a bearing journal OD that should be perpendicular to the datum.

[John Kingston]

Is the probe pointing straight into the hole? If not, the probe compensation maybe off, due to the probe hitting above its equator. Also, the CMM is not necessarily the best way to check roundness, especially if it is tight tolerance. There have been numerous discussions here about this very subject.

[jacyEd]

I definetly understand what your saying and you may be right on. What i have been doing is measuring at an angle that is not perpendicular to the surface of measure. _|_ / <---- basically like this with the / being the probe. I am forced to do this due to constraints of the CMM and the size and shape of the part. We do have a 2nd CMM with a TP20 and no 100mm extention that does not have the same roundness issue measuring the same way. My guess is the less sensitive probe and no extention help it to remain consistant.