roundness / circularity 1

[josson]

I am measuring a large diameter of 100mm on a CMM. The tolerance on the interal diameter is extremely tight +/- 0.05mm. A lot of my measurements have had roundness values greater then the tolerance range > 0.1 which I have considered to be out of spec. Is this correct? Also, some have roundness values less then the tolerance range, like 0.09 for example, but the nominal reported by by the CMM is at or just below the LMC of 100.05mm. Would this be considered in spec? It seems to me that not likely within the size limits in this senerio. My measurement experience is limited and there is no meteorologist in house. If anyone could explain how to interpt such results or suggest another approach it would be appreciated.

 

[alexit]

You have 100mm +/-0.05 bore, is there additional geometry spec or is roundness/cylindricity also assumed as +/-0.05?

What is important (you measuring size or roundness?)

Of course CMM is not best equipment for measuring size / geometry in this tolerance range (air gauge / form measuring machine from companies like Mahr is better) but you work with what you have.

How is CMM output set? Are you getting point cloud or surface trace?

 

[josson]

The roundness is not specified although I am assuming it should not be greater then .1.

I am dealing with an o-ring groove for a seal designed for minimal o-ring compression so both the size and roundness is of importance, especially if the roundness is throwing the groove out of size.

The roundness values I have seen have been anywhere from 0.050 to 0.250 but I am not sure how to interpret these when the CMM reports the diameter to be in spec especially when the bore is close to LMC.

I am not sure what the output set up is. I will have to check. Is one preferable for the my senerio.

 

[MikeHalloran]

<rant>
The phrase "minimal o-ring compression" should be a red flag, as should "maximal..". Any deviation from o-ring manufacturer's handbook specs is a sign that a dilettante has been messing with the design, and the whole thing will be a headache, forever.
</rant>

What's important to the o-ring is the radial depth of the gland, as assembled. The groove can be elliptical, or look like a polydyne cam, and still function, so long as the radial depth of the gland, not the groove, at every point, gives the right radial compression, and the perimeter gives the right circumferential compression.

<rant>
A CMM is the the worst, most error- prone, most labor- intensive possible choice for the task of measuring an o-ring groove.
</rant>
You have clearly figured that out already.

Unfortunately, the depth of the gland that's important to the o-ring is determined by two (or possibly more) parts, and you only measure one at a time, so you have to apportion the available tolerance (not much) among the features that define the real dimension of interest.

In this case, the real limit on roundness is partially determined by the clearances between the locating features of the parts in the assembly, and their allowed eccentricity relative to the o-ring faying surfaces. We can't evaluate that from here, but someone needs to do it. You can't just pull a number out of the sky.

A meteorologist can't help you. A metrologist might. He'd probably say that you need to be measuring runout relative to whatever feature locates the mating part(s).

If you have the time, put the problem into Excel, and work out what happens as the dimensions and eccentricities change. You'll probably find that the whole assembly needs to be retoleranced.








Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

I've had big problems with static and dynamic o-rings assembled in properly dimensioned grooves when the grooves had rough or threaded finishes.

 

[ccw]

If you are using ANSI Y14.5 dimensioning and tolerancing rules then by Rule 1, in the absensce of other modifiers, everything must lie within the +/- .05 theoretical lines, and can range all around the 0.1 space in any manner.

As Tmoose points out there are other parameters to be considered in the o'ring dimensioning world, surface finish not being the least of these. In addition to diam tol., runout, and surface finish, I have also used profile per ?? degrees to control critical seal dimensions. I think this spoke to Mike's "polydyne cam" problem, where detrimental rates of change of surface contour were inside runout control but outside surface finish control. This fourth parameter, by the way, caused great consternation amongst the dimensional verification folks. ;-)