About Parallelism Measurement

[[Za...]]

Hello Everyone,

Im trying to measure parallelism 0.05 on my part. When I check the parallelism dimension according to the main reference surface (Plane), of the cylinder that I have created theoretically by combining certain circles, I see 0.008. When I create a 3d Line from the same circles and the value i see is 0.057. What could be the possible reason for this? Which measurement method should I use?

[[Da...]]

can you show a sketch or a section of the print?

[[Ja...]]

Make sure you are using recall and not recall feature points. Recall will use the center points of the circles. Recall feature points could be off if there is more points on one side of the circle.

[[Ja...]]

Disregard that reply. I thought we were talking about perpendicularity.

[[Za...]]

Here is the section of parallelism of the drawing.

[[Ri...]]

Man drawings just keep getting worse and worse these days.

-removed

[[Ri...]]

Hold on - wait - you cannot do parallelism of a cylinder/axis that is perpendicular to the datum.

Are you sure the parallelism isn't pointing to the lower surface??

[[Ün...]]

Merhaba Zafer,

Çizimde hata olmalı, Paralellik sembolü mesafenin ve yüzey pürüzlülük değeri altında verilmiş.
Delik içindeki alın yüzeyinin (19,00 mm ) paralellik değeri isteniyor olmalı.

Tasarımcı yada müşterinle iletişime geçmelisin.

Kolay gelsin.

[[Za...]]

The section that I sent is not understandable. Parallelism measure of the cylinder (or 3Dline) generated from Circle1,Circle2 ,Circle3 ,Circle 4 is required according to Datum Z. As i mentioned,

If i create Cylinder, parallelism value is 0,008
If i create 3Dline, parallelism value is 0,057.

[[Dy...]]

looks like circle 2 is at a different height then the others try removing that from the 3d line and see if it matches better to the cylinder readout?

[[Da...]]

WOW Circle 2 should not be included but a new circle between circle 1 and the existing circle2 should be part of the check. use 1 and new 2 and 3 and 4 for a cylinder then ask for parallel to the plane (Z). I Think !?
report the diameter sizes also, if they are different they may affect the parallel reading. Also use recall feature points for the cylinder, and don't forget to apply your filters to the cylinder.

[[Mi...]]

Like others have said, what you labelled as circle 2 doesn't look right. Did you mean to label the next "step" above that one, that looks more like circle 3?

Of the two options, I would use a cylinder; that way it will constrain the evaluation of each circle to force them to be concentric.

HOWEVER, that will evaluate the cylinder's axis - are you sure that's what you want? I can't really tell from the drawing snippet you provided, but could it be referring to the surface itself instead? As in, scan a 2d line (or 4 disjoint 2d lines) along the bottom of the trough, and use that instead.

[[Za...]]

Circle 2 is at wrong side at the picture i sent. I use these circles. Also scanning whole circle form, and applied gauss filter. I am not sure if the customer wants parallelism measure from bottom side of form, or center of cylinder (3dline).

Inked222222_LI.jpg

[[Dy...]]

ok that makes more sense, to get closer results with 3d line project the 3d line against another flat face in X should solve the issue, the 3d line will average out your X,Y,Z and is they are even slightly off to each other in X it will effect your parallel so constraining it with projection will eliminate that variable. idk how you have this fixtured but this is the direction i mean for X

Inked222222_LI.jpg

[[Ro...]]

I wouldn't recall the circle segments to create a 3d line because you're going to get a line calculation that's trying to find the center of small circle segments that have no opposing points. I would opt for creating line segments across all 4 flats at various locations and checking parallelism of those.

To be more specific: I would either create lines that measured all 4 flat sections in 1 line, or measure individual segments and recall those feature points into lines.

Thinking about it more, I think I would recall those line sections into a cylinder to get my axis. But I would definitely opt for using line segments, as I think it represents the actual geometry better.

[[Za...]]

Is there any way to create a line from bottom of the circles in Calypso? I could not find.

[[Ri...]]

I know this part - or one extremely similar.

Constructing that Centerline isn't easy.


For me - I like to create Curves on the Circle section - create a Best Fit from the Curve - now measure a 3/4 touch-point Circle on the Circle section measured to the Best Fit of its respective Curve. Then you can constrain the radius of the Circle. Once you have both you can construct a 3d Line with the Recall function.

There are other options involving Free-Form if the results produced above aren't desirable.