...from my understanding, it reports the vector deviations from nominal/thoretical vector values...this deviation is on the print out...useful for profile/form error evaluations...I don't use it too much tho...others probably have better stuff for you...
Anyone using this regularly. I'm just trying to get a better hold on this. any help of explaining this in simple terms would be greatly appreciated.
T = TOTAL vector deviation or deviation along the vector from the nominal point location.
This is for
SURFACE
VECTOR and
EDGE points.
For SURFACE & VECTOR it is the surface deviation, for EDGE it is the EDGE deviation.
When doing EDGE points, if you take a surface sample hit, you can report S in addition to the XZYT and the S will be the SURFACE deviation from the surface sample hit.
Heres the deal I get allot of prints that look like this the I report XYZ of these points and the T value also. just making sure T value or the S value was the right thing to do.
Also, when reporting Surface Profile (Form and Location) it is the value listed in the deviation column when textual analysis is applied.
Like this:
You are correct, sir.
FORMULAS:
T = SQRT ( XDev^2 + YDev^2 + ZDev^2 )
It is actually a little bit more than that if you are NOT using SNAP for your vector points or if you are using SURFACE points or EDGE points. If not using SNAP points, the vector values are included in the formula, BUT, the T value will ALWAYS show as if you used SNAP.
If you don't use snap, you COULD see something like this:
XNOM 0 XVector 0
YNOM 0 YVector 0
ZNOM 0 ZVector 1
This would be a point, straight up and down. Now, NO MACHINE is perfect, they ALL have some amount of drift, so your touch COULD be this (grossly exaggerated and in METRIC):
XACT 0.125
YACT -0.105
ZACT 1.124
No matter SNAP or no, your T deviation in this case will be 1.124, HOWEVER, if you trig it out (as some will do), you would think it would be 1.135, but that would NOT be along the vector of the point
SQRT (0.125^2 + -0.105^2 + 1.124^2) = 1.135
When done using the vectors, well, ZERO times anything is zero, so:
SQRT ( (0.125*0)^2 + (-0.105*0)^2 + (1.124*1)^2) ) = 1.124
the 1.124 is what the T value will ALWAYS be in this example, BUT, by using SNAP (which put the actual reading back on the perfect vector line), it removes the drift from all 3 axis. Now, some will say this hides a sloppy machine, and yes, it can, BUT, this example is a VERY exaggerated example of how much drift you might see. On my old (OLD! 20+ years old) machine, I get less than 0.002" of drift and that really won't show up unless you are trying to hold micron values.
Heres the deal I get allot of prints that look like this the I report XYZ of these points and the T value also. just making sure T value or the S value was the right thing to do.
If you create a set from those points you are measuring and then dimension it with profile, you can get a report that looks like what I posted above...a little simpler than reporting each point individually.
If you create a set from those points you are measuring and then dimension it with profile, you can get a report that looks like what I posted above...a little simpler than reporting each point individually.
...some people ("Engineers") want to scrutinize e-ver-y sin-gle point into oblivion...then when they see what's right/wrong...they'll tell you they were right anyway...most times...at least here..."Profile" is the do-all when engineers can't tell you what they mean/really want on the print...
It is actually a little bit more than that if you are NOT using SNAP for your vector points or if you are using SURFACE points or EDGE points. If not using SNAP points, the vector values are included in the formula, BUT, the T value will ALWAYS show as if you used SNAP.
If you don't use snap, you COULD see something like this: XNOM 0 XVector 0 YNOM 0 YVector 0 ZNOM 0 ZVector 1 This would be a point, straight up and down. Now, NO MACHINE is perfect, they ALL have some amount of drift, so your touch COULD be this (grossly exaggerated and in METRIC): XACT 0.125 YACT -0.105 ZACT 1.124
No matter SNAP or no, your T deviation in this case will be 1.124, HOWEVER, if you trig it out (as some will do), you would think it would be 1.135, but that would NOT be along the vector of the point SQRT (0.125^2 + -0.105^2 + 1.124^2) = 1.135 When done using the vectors, well, ZERO times anything is zero, so: SQRT ( (0.125*0)^2 + (-0.105*0)^2 + (1.124*1)^2) ) = 1.124 the 1.124 is what the T value will ALWAYS be in this example, BUT, by using SNAP (which put the actual reading back on the perfect vector line), it removes the drift from all 3 axis. Now, some will say this hides a sloppy machine, and yes, it can, BUT, this example is a VERY exaggerated example of how much drift you might see. On my old (OLD! 20+ years old) machine, I get less than 0.002" of drift and that really won't show up unless you are trying to hold micron values.
I was explained in PCDMIS CAD++ class that snap point is to correct machine drift mathematically not hiding machine drift (at least that the impression that I got). Example:Blue print defined a point on a complex surface at XY exact position (basic dimensions) and the third axis (Z) must meet blue print requirement at that position. The snap point does just that by re-calculating the actual Z value at exact position (theoretical position).
I remembered the instructor said he was helping one of his customers measuring a complex curve surface using CAD and the DCC of the machine is very sloppy, the part they were checking is a known good part but the result proved otherwise he then suggested to use snap point and everything fall in.
Your explanation is exactly the same as stated in PCDMIS help file.
I have read many of your postings and I know you are among the very best in the country, I have no doubt of your assessment. I am just confused!!!
