Profile of a Cone

[DrBudman]

I'm trying to determine how pc-dmis calculates profile of a cone using form and location. I though I seen somewhere on this board that someone posted how all the calculations for pc-dmis are made but cannot find it.

Here is why I'm confused:

Measured cone has 8 hits at 4 levels for a total of 32 (cone is only about 1/4" long).

I report on the profile (form and location) and get a devistion of .0038", but when I do a xyza report I get:
meas nominal
X = 0.0026 0.0000
Y = 0.0023 0.0000
Z = -.4068 -0.4031
A = 60.1807 60.0000
also, roundness of cone checks .0026 (does'nt it take this into account?)

It just seems like my profile deviation should be higher than it is.

I have a model, but I justed entered required information manually, X, Y, Z, A, diameter 1 and diameter 2.

Thanks,
Dan
ver 3.6
measured in inches

[DrBudman]

No takers????

[John Riggins]

It looks like PC-DMIS is reporting the greatest deviation of the three axes.

(.4068-.4031 = .0037) rounding could account for the .0001.

John

[DrBudman]

Never thought about that...I'll have to keep searching for the formulas.

Thanks

[JamesMannes]

Yeah, I asked for the formulaes/algorithms from BNS and they never got back to me. The best I can do for you is to provide this lame axx link for definitions of dimensions. http://www.wilcoxassoc.com/Downloads...ionalfiles.htm

[DrBudman]

Things That Make Me Go.. Huh

From Wilcox Site:

8.2 Surface Profile

Any feature can be selected for a surface profile dimension. The surface
profile is calculated by computing the deviation of each actual measure point
from the corresponding theoretical measure point in the direction of the
theoretical vector for that point. Results can be reported for form only or form
and location. Form only will calculate the measured profile tolerance value
from max-min deviation, and the profile dimension will check the measured
value to see if it is out of tolerance. The form and location will calculate the
measured profile tolerance value from max-min deviation for information
purposes, but in checking whether or not it is out of tolerance it compares max
to the plus tol, and compares min to the minus tol. Both of them must be in
tolerance.
For V3.6, when form only is selected, the user also has the option of doing a
3D least squares bestfit of the theoretical and actual measure points before the
profile form is calculated. Of course, the user also has the option of doing a
2D bestfit alignment before the profile dimension if a bestfit other than least
squares is desired.


Thanks for the link.... I read it and go huh?? Maybe its my little pea brain that just does not put 2 and 2 together here. I understand what they are trying to say but it does not relate to the dimensions variation I get... So I guess I'm still at Huh?

Thanks
Dan

[DrBudman]

I think I have figured this out... hope so anyway. Per the profile definition, it is the min max error. I turned on the graphical and textual display of the cone profile so I could see it on the screen and see where the error is and to see each individual point errors. It appears the calculation it is making is for the "T" value of my points. My profile is checking .0026 and looking at the textual values I have a min deviation of .0012 and a max deviation of .0014 for a total of .0026. If I did the calculations for the "T" value correctly for my min and max point (and boy was that fun) they calculate out to be the same.

Just thought I would throw this out there if anyone would like or see if I'm doing something incorrectly.

Thanks
Dan

[kbotta]

Ok...
<soapbox starts>
This is another one of what I believe to be Pcdmis's GD&T inaccuracies. It calculates 'profile' as the diff. between Min and Max of your points. Although you will not find in ASME a direct way to report the 'total profile' this is how I believe it should be done. You take your worst deviation and mulitply it by 2. This is of coarse assuming you are talking bilateral equal distribution tolerance zones. As one side of the zone 'grows', the other must also follow suit. The same as a positional tolerance (x's 2). What we are reporting is a tolerance zone. Or, to out it in another term, we are reporting the tolerance zone it would take to encompass the feature, and all points within the feature.
Now, the argument can be made to show 'diff. between max and min' , but after talking to a committee member, he believes also that it should be the worst deviation X's 2.
If you decide you want to show it as the min-max way, a corporate policy should be drafted (per different resources) in case of desrepency.
<soapbox ends>

Sorry, had to get that off my chest...
Kev

[DrBudman]

No problem kbotta, makes too much sense....

The whole reason I brought this up to begin with was, I told someone their profile was .0016 out of tolerance, but that does not mean a hill of beans when I don't understand what is actually out of tolerance.

To report an a deviation is one thing, to understand what it means is another.
Is roundness causing our problem.... is the location of the cone itself... is the angle causing it... etc... When I train someone I want them to understand what is going on and not learn by repetition.

I can't fix a problem when I don't understand what the error is but atleast now I can put some numbers to the error.

This whole profile thing is easy when uderstanding that the surface has to be in a given tolerance zone, but what the cmm is seeing and correcting the error is a whole nother issue.

[Jeffrey Lovely]

Lets say profile measured -.05 lowest pnt and + .14 the higest pnt
then profile = .19

lets say you are allowed .30 (bilateral) +/ .15
someone will say you are out of tol - "u r not"
"dont use 2x"
so I will report individual points or "T" values at +/- .15

[kbotta]

Yep, I still report my 'T' val...but for the 1 inclusive answer "total profile"...well...see above.

I agree with Bud's assesment of finding out what is wrong. You cant fix it if you cant see it. kudos, words to program by....
Kev