Concentricity

[ag162]

Good Morning,
I need verification on the method that PC DMIS uses to calculate the concentricity of a circle. This will be printed out for the Upper Mgt in an attempt to acquire a tolerance deviation. While I gave an answer, I need the support of this posting for my answer.

The question is:
When calculating concentricity is the deviation (ie.. dev = .900mm) a unilateral or a bilateral value? As a linear distance or a range?

Thanks in advance,
A.Gore

[craiger_ny]

It is from zero (the target, also the low limit) to the limit specified in your FCF. I may not understand your dillema and exactly what you need to know though.

[Wingman]

PC DMIS does not do a true Concentricity. PC DMIS basically reports TP. You input a unilateral tolerance just as in TP. PC DMIS will report a TP style Deviation (radial deviation x2).

[ag162]

What the situation is, I was in a conversation with our water walkers in Engineering about the concentricity of one of our parts. Although I stated that the results were a linear distance from the center of the datum a (cylinder), he wanted me to investigate this a little more thoroughly. The result is a hypotenuse (sp), or a tp value from the theoretical zero. so the supporting answer is:
a unilateral dimension from the theoretical zero. Am I right?

Thanks,
A.Gore

[Wingman]

What the situation is, I was in a conversation with our water walkers in Engineering about the concentricity of one of our parts. Although I stated that the results were a linear distance from the center of the datum a (cylinder), he wanted me to investigate this a little more thoroughly. The result is a hypotenuse (sp), or a tp value from the theoretical zero. so the supporting answer is:
a unilateral dimension from the theoretical zero. Am I right?

Thanks,
A.Gore

It is diametric. The result is the linear (radial) distance times two if I am not mistaken. You have a Cylindrical Tolerance Zone.

[spazus_maximus]

I am pretty sure it is the average range of all the diametrically opposing points back to centerline

Kinda Like

d1-d2=v1 (d1 being distance from point to centerline & d2 being the opposite point to centerline)

Then the range of all v's

Atleast thats how i understand it......??

[cmmguy]

It is supposed to be what Spaz describes but CMMs cant measure that way, so it ends up being the radial deviation of centerlines TIMES 2 as Wingman describes. TP and Conc will give the same results.

(It is the diameter about the datum that is defined by the centerpoint of the measured feature)

[Wingman]

Well, according to the Spec all median points must fall within a cylindrical tolerance zone. Median points are diametrically opposed points. PC DMIS does not calculate Concentricity according to spec. PC DMIS treats Concentricity as a True Position. The Centroid or Axis of the feature muct fall within the cylindrical tolerance zone.

Not exactly correct but the true value of concentricity is suspect anyway (if you ask me).

BTW, the 2nd feature is the PRIMARY DATUM feature. If you specify only one feature then the Workplane is the Primary Datum Feature.

[ag162]

Thanks for the input. I learned something new today, while I was incorrect about how this is derrived, it is always good to ask and learn the correct answer.
However, if I put a part on a V block, used an indicator, zeroed out on the dat a cyl, rotating it, then moved to the area to be measured. Lets say that i had .15mm dev on the indicator, and I had a .1mm tolerance to dat A. How would I calculate that? Just for my curiosity.

Thanks,
A.Gore

[cmmguy]

The answer you get with your method is a good answer - no doubling as a CMM does. However, checking concentricity in a vee block presents other problems due to form...

[tking]

Linear distance from center to center X2.

I put a 20mm gage pin in a V block and made it origin, replace it with a 25mm gage pin and probed a circle. The distance between the 2 centers was 3.548, the concentricty result was 7.097 (3.548 X 2).

Experiment finished,

TK

[cmmguy]

Linear distance from center to center X2.

I put a 20mm gage pin in a V block and made it origin, replace it with a 25mm gage pin and probed a circle. The distance between the 2 centers was 3.548, the concentricty result was 7.097 (3.548 X 2).

Experiment finished,

TK

Not so fast... he said indicator not CMM... you know, the old-fashioned way.

[spazus_maximus]

Why not use variables to calculate concentricity?

[Wingman]

Why not use variables to calculate concentricity?

Why not throw concentricity out the window as the bad idea it really is.

1) I doubt the Engineers that use it really know why they are using it.
2) What does it 'really' return? When is it 'really' ever needed.
3) True Inspection is BIG $$ and there is dubious return.

Check with the Eng but I can almost guarantee they just want the two features to be on the same centerline. TP/ Coaxiality - call it what you want.

[tking]

Not so fast... he said indicator not CMM... you know, the old-fashioned way.

So fast, what? If I understood what Mr. Gore asked it was what did the demon spit out when the concentricity function was used. That is all MY experiment provided. Did I miss something? As to the indicator I was not commenting, K?

TK