【转帖】question about composite position

yang686526

question about composite position
i am inspecting a part on a cmm that has holes toleranced with composite true positioning.  the pltzf has an mmc tolerance of .030 and the frtzf has a mmc tolerance of .015 with no datums referenced.
reviewing data from the cmm report i am seeing that the pltzf positions are good with deviations from .004 to .008.  what confuses me is the frtzf positions are reported as not meeting specification with deviations from .018 to .025.  is this possible?  or am  i right to believe that there is an error in my cmm program.
first of all, the pltzf was pretty easy for your cmm to calculate even if it was at mmc. those are valid readings.
you really can't perform the frtzf on a cmm since it requires one of the holes as a datum. it then can calculate the positional tolerances of the other holes relative to the selected hole but that is not the requirement. if one changes the selected hole, then the result would also change.
the best way to check the frtzf at mmc is with a checking fixture made at its virtual condition size. it truly checks the relationship inside the pattern. i would not have any confidence in the frtzf result from your cmm.
  
dave d.
mcaruso,
there may be something wrong with your program or there may not! if your cmm software has the ability to do a pattern "best fit" to accomplish the frtzf portion of the analysis and you have not constrained it to datum features then it may be working correctly.
you said that the tolerance for the frtzf is 0.015(m)... just curious what is the tolerance for the pltzf ???.?(m).
i disagree with dave's statement
quote:
you really can't perform the frtzf on a cmm since it requires one of the holes as a datum.
paul   
0.030 i guess i overlooked it
paul
the cmm software i am using has a reporting statement for composite position that i would expect to properly calculate the both the pltzf and frtzf tolerances.  i align the part to determine that datum, then measure the holes as circles.  i then tell the program to first report the individual features where i define the in this case x and y coordinates and diameter.  i then ask for composite position of a group of holes.  i am allowed to either choose datum reference or datum features.  
i input the tolerance of .030 for the pltzf and .015 for the frtzf both with mmc modifiers.
the only other option is a box for the frtzf, which is titled 'alone'.  according to the cmm software help, "when checked, the alone checkbox indicates that the same geometric characteristic will be applied to both the pltzf and the frtzf,  the software will use the frtzf datum features to report on the relationship of features within the pattern.  clear the alone checkbox to verify the feature position relative to the datums, as well as relationship between the features"
i will not state how i have the alone box to get unbiased opinions.  also i can state that like most problems in life, this was simpler that i expected and i just found two holes from another pattern dragged into my composite position matrix thus throwing off the frtzf calculations.
i cannot say that i recognize your software options from your description but i'm sure that if you pose your question to one of the cmm forums...
i was involved in sourcing cmm software for offline programming, operation and post-processing about 7 years ago.  at that time, we found that very few software packages (oem or add-on) were capable of processing the frtzf segment, and we would have had to do as dave indicated above for many of the packages.  there were some packages that were properly and demonstrably y14.5.1 compliant.
i've had the same problem pop up, mcaruso ... hang in there.
jim sykes, p.eng, gdtp-s
mcaruso
   you have a problem with the cmm program. the positions among the holes, frtzf, will always be equal to or better than the positions as related to the datums, pltzf.
in reply caseynick's statement,
so lets say that my holes deviations are split between positive x and negative x values from true center.  wouldn't that calculate the frtzf values bigger than the pltzf values?
    the cmm software may calculate the frtzf as bigger than the pltzf. but if you were to position each hole at it's best location the postions would be the same.
    let's say you have a pair of holes with a composite positional tolerance. one hole is .004 short of the basic dimension and one is .004 long of the basic dimension. all the deviation is in the x direction. no deviation in the y. the postion of each hole, for the top part, pattern locating portion, of the composite callout would be .008. agreed? the postion for the bottom part, feature relating, portion of the callout would be the same. .008 for each hole, not .000 and .016. nowhere does it say that you have to set up one hole as perfect and check the rest from there.
    if you were to make a gage based on virtual condition it certainly would not be perfectly centering one hole and checking the other hole from there. the part with the holes would be rotated, slid sideways, etc. till it was determined if it did or did not fit on the gage regardless of the direction of deviation of the holes.