Line vs. Cylinder Controversy

[Big Red]

Guru allert! Please help settle our conflict...

Is it better, worst or the same to use a Constructed Line or a Cylinder?

Some of us measure bores as Cylinders. We use them to level, check concentricity, datum alignment, etc...

Some of us measure 3 or 4 circles down the lenght of a bore then construct a line using these circles. We use them to level, check concentricity, datum alignment, etc...

Some say that if we measure a bore as a cylinder, we cannot see if it is tapered (big on the top, small on the bottom) or if it is bell mouth at any point. The circles will tell us this information.

Some say that only a cylinder is good for checking concentricity, runout or using as a datum alignment as the axis of a cylinder is better than a constucted line.

So, you see, there is controversy and I look to the gurus of PD-DMIS user forum for input.

Thank you so much.

[Matthew D. Hoedeman]

Well, if you measure it as a cylinder, ALL the points get 'averaged' to create the cylinder. If you measure them as circles, then each circle's group of points gets averaged to creat that circle, then the average is what is used when constructing the line through the circle centers. So, I am pretty sure that you WILL get 2 different results, one for each method. What you can do is measure all the points as points, then construct circles at each level and also construct a cylinder using all of them. Then construct the line through the circles' center points and see if the line is different than the cylinder. AS for which way is best, ya got me!

[tking]

While far from being a "guru" I will contend that a cylinder is best for alignment purposes (3D). It's form will give an indication of the other conditions. To know the particulars of the other conditions circles at various levels would be required. By the by, circles can be constructed from the points of a cylinder.

.02 or .03,

TK

[bw_bob]

Like the others said either way you do it you will come up with different results. The best way to do it ???

[george frick]

the "best" way is whatever gives you a "clean" cmm report. makes everyone happy. if you really want to analyze the "true" condition- well that's metrology and no one is really interested or will just undermine your results. ship it.

[bw_bob]

the "best" way is whatever gives you a "clean" cmm report. makes everyone happy. if you really want to analyze the "true" condition- well that's metrology and no one is really interested or will just undermine your results. ship it.

George you should be in management if your not already there.....

[Jan d.]

Here's my $0.02. I heard that older versions of PC-DMIS had trouble (or flat could not) with the construction of a cylinder from circles. I was told that this is not an issue anymore in V4.0 or later, although it seems that even within the hexagon application engineering corps, they are not entirely sure construction of a cylinder by means of circles is done correctly.

For me: I NEVER measure an auto-cylinder. I only measure auto-circles and construct. Gives me far more information. And based on the numbers I get, I think the cylinder construction actually works just fine now.


Jan.

[#2#]

Just an echo of what has already been said. If it is your boss you are arguing with, do it his way unless you can prove one way is better than the other. Just talk about it in an e-mail and save that e-mail as a cya in case someone higher up starts asking questions about why you do it that way.

[than_357]

The question would arise;

if the bore you're measuring is at a significant enough angle to the work plane- will the circle method (and resultant ovality of measure) show correct dia. sizes, and would it show true centerline deviation (to the work plane)?

I use cylinders as a matter of most contact with the surface and remeasure as circles if the min/max looks scary.

Tom