Material condition on datums

[terry]

Can any one explain to me how using datums on could increase my positional error? Below is the same feature true positioned three ways, first is with datums off (.112), next datums on using MMC (.130) and finally datums on at RFS (.177). If PC-Dmis is best fitting the datums to simulate a hard gage check how can my true position error be greater by using datums @ MMC? Current alignment is -E- -C- -M-, bonus does not apply to hole because it is a threaded hole.

DIM LOC119= TRUE POSITION OF CIRCLE E_CIRCLE3 UNITS=MM ,$
GRAPH=OFF TEXT=OFF MULT=10.00 OUTPUT=BOTH DEV PERPEN CENTERLINE=OFF DISPLAY=DIAMETER
AX NOMINAL +TOL -TOL BONUS MEAS DEV OUTTOL
X 400.800 400.826 0.026
Y -86.500 -86.450 0.049
DF 10.500 0.100 0.100 10.592 0.092 0.000 --------#
TP RFS 0.800 0.000 0.112 0.000 -#-------
END OF DIMENSION LOC119
DIM LOC120= TRUE POSITION OF CIRCLE E_CIRCLE3 UNITS=MM ,$
GRAPH=OFF TEXT=OFF MULT=50.00 OUTPUT=BOTH DEV PERPEN CENTERLINE=ON DISPLAY=DIAMETER
AX NOMINAL +TOL -TOL BONUS MEAS DEV OUTTOL
X 400.800 400.738 -0.063
Y -86.500 -86.517 -0.017
DF 10.500 0.100 0.100 10.592 0.092 0.000 --------#
D1 PLANE DATUM_E_DCC AT RFS
D2 12.763 0.020 0.020 0.014 12.757 CIRCLE DATUM_C_DCC AT MMC
D3 82.626 0.020 0.020 0.026 82.632 CIRCLE DATUM_M_BORE AT MMC
TP RFS 0.800 0.000 0.130 0.000 -#-------
END OF DIMENSION LOC120
DIM LOC121= TRUE POSITION OF CIRCLE E_CIRCLE3 UNITS=MM ,$
GRAPH=OFF TEXT=OFF MULT=50.00 OUTPUT=BOTH DEV PERPEN CENTERLINE=OFF DISPLAY=DIAMETER
AX NOMINAL +TOL -TOL BONUS MEAS DEV OUTTOL
X 400.800 400.717 -0.083
Y -86.500 -86.530 -0.031
DF 10.500 0.100 0.100 10.592 0.092 0.000 --------#
D1 PLANE DATUM_E_DCC AT RFS
D2 12.763 0.020 0.020 12.757 CIRCLE DATUM_C_DCC AT RFS
D3 82.626 0.020 0.020 82.632 CIRCLE DATUM_M_BORE AT RFS
TP RFS 0.800 0.000 0.177 0.000 -#-------
END OF DIMENSION LOC121

Regards
Terry

[J Temmen]

I believe 0.130 is less then 0.177.

[terry]

yes but .130 is more than .112. I agree that when using datums the MMC (.130) error is less than the RFS (.177) error. What I dont understand why when using datums off the error is less that using datums on.
Regards,
Terry