GD&T question - Profile Again...

[c_bearden]

Have a look at the following website and tell me, should the SIZE of the Hex be toleranced ±.1 or ±.05

http://www.tec-ease.com/tips/september-00.htm

Thanks,
Doberman

[Jan d.]

You can't say that. Profile controls form, orientation AND location. So when you go further off location, there is less left for form and orientation, etc.

If you had perfect form and orientation, then the location could theroretically vary by up to +/-.05 . But perfect form and orientation never occurs on my parts....


Jan.

[c_bearden]

Thanks for the reply, Jan.......Me, My boss and another supervisor at work have had a few LOUD discussions regarding this callout.....They both are wanting to apply the profile callout as a ±.....meaning, when you see a profile callout and its "ALL AROUND"..... they want to check the wall thickness of the part and Apply ALL of the profile tolerance as a ±.
Heres an example.....the Profile callout is (Profile of a surface @ .005 to Datum -A-B-C- All Around).......Say there is a post or boss on the part called out DIA .250........They want to apply ±.005 to the DIA. and they say the Profile callout lets them do that.......I disagreed with them and we had some heated discussions.......I told them if thats what you want to do, go ahead but don't expect me to buy it off........
Can someone help me explain this to them.......

Thanks,
Doberman

[Jan d.]

As usual, heated discussions regarding this topic are very normal. I have had my share of them. And been wrong...

They are wrong though... unless otherwise specified, it is +/- 0.0025.

I have had a lot of luck using James D. Meadows' book "Geometric Dimensioning and Tolerancing" (Marcel Dekker Inc). This stuff is arcane, so is James' book. But he does a good job trying to explain these fundamentals.

Let me try to give short stab at it: Your datums A, B and C set up a reference frame. Using the basic dimensions, you create a theoretical part location from those ACTUAL (and this is important) datums. So you create a nominal surface locations from actual datums. Or the location of multiple surfaces, if you have an all-around call-out. Then you apply the profile (and if it is bilateral, you use half the tolerance up and half down).

What may help you a lot, is going to V4.1. Use the new XactMeasure GD&T and you do not have to worry about how to work around it: on you report it shows the FCF just as it appears on the print. Discussion closed. No more arguing what is right and what is wrong...

Be careful though: I remain of the believe that the way PC-DMIS reports profile is not intuitive. Make sure you always only report MIN/MAX. That's your ticket to the correct values, plus you can do realistic process control on them.



Jan.

[Mike Johnston]

I simply scan the surface using one of the scan tools and then click on "single hit" before running the scan. From there I dimension the "T-value" from all of the individual hits that are generated and then find the maximum T-value whether it is inboard (-) or outboard (+) and multiply it by 2. That's my profile.

Mike

[Paul Sarrach]

Not that I ever seen such a callout, I would assume that it is profiling all sides to itself. Where as once I have determined the center of the hexagonI would take two edge points per side and tolerance the T value to itself if its within +/- .05 I would call it good. Come on Matt whats your opion?

[spazus_maximus]

±???

Profile is an envelope. Whether form & location or form only. It could be + the whole tolerance, - the whole tolerance, or any combo in between.


Right?

[Matthew D. Hoedeman]

OK, Paul(a), here is MY take on that callout:

The profile of the hex hole is to be within +/-.2 to the datums on any and all points taken on it. HOWEVER, what I would then do is origin to it and all the points TO THAT ORIGIN must be within +/-0.05. In other words, it can vary TO THE DATUMS +/-0.2 but it's actual shape must be within 0.05 to itself. Try doing a BEST FIT alignment on the ABC check data, then dimension it all again using the tighter tolerance.