Parallelism on per unit basis

[powerhound]

Hey folks,

Another question has come up. I found another thread on this subject but didn't find a definitive answer in it. I don't find direct support for parallelism on a per unit basis. This seems to be a legitimate extension of principle but I'd like to get others opinions or maybe point me to the place in the 2009 standard where it exists. In reality we'd like to add a tangent plane modifier to it but that shouldn't make a difference in the answer.

Thanks,

Powerhound, GDTP S-0731
Engineering Technician
Inventor 2010
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[MechNorth]

Powerhound, your definition of tangent plane is intuitive, but not supported by the standard. I've attached a Tec-Ease tip as reference for (T) modifier with a datum reference. Note that the tangent plane is oriented wrt the datum references, not per the 3 high points; it is located at the single highest point on the surface, while the plane is oriented.

http://www.tec-ease.com/gdt-tips-view.php?q=218

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[powerhound]

Jim,

I agree with the tip; however, it is using a basic dimension along with a profile of a surface callout. That is not the same as using parallelism with a toleranced dimension.

Powerhound, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[pmarc]

Jim,
I do not think the tip proves that tangent plane is oriented wrt the datum references. Actually it clearly says that the tangent plane is established by high points of the feature. This plane is simulated by surface of flat plate and that surface must lie within tolerance zone that is oriented (and in this case also located) wrt datums.

 

[Belanger]

Jim, the tangent plane doesn't have to be oriented wrt the datum ref frame. It must merely fall within a given tolerance zone, and that tol zone must be oriented to the datum ref frame. I think that's what pmarc was getting at in the post from 15:50.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[CheckerHater]

Pmarc, Belanger, Powerhound

Yes, indeed, it is about geometry

 

[pmarc]

CH,
Cut off everything that is on the left and on the right of the local area, and the one that you are saying is not tangent will become tangent. Not tangent in geometrical understanding, but tangent in a way that the plane will be contacting the high points of the surface within considered area. Again, forget about math, just follow Y14.5.

J-P,
Yes, this was exactly my point.

 

[CheckerHater]

And where exactly Y14.5 supports your interpretation?

 

[pmarc]

And what else, besides definition given in paras. 1.3.45 & 6.5 and fig. 6-18 (all in Y14.5-2009), do you exactly expect to see? It was stated at the very beginning of the thread that the issue is a typical "extension of principles" stuff.

 

[CheckerHater]

Pmarc,

I didn’t find where it mentions “surface within considered area”, only the entire surface.
Illustration doesn’t show it on “per unit basis” either.
Y14.5-1 is no different.

About being “rather worried if the local area was convex”, I may assure you that your “plate” will rock anyway on either convex or concave surface, because there is no guarantee that surface will always match all 4 corners of your 1 x 1 area. So why not to extend principle even further and specify 1 x 1 x 1 triangular zone?

Let’s face it: Parallelism per unit basis is not supported by the standard. There might even be a reason for that.
Also, Parallelism when specified with Tangency does not automatically control Flatness, so it is only natural to add separate Flatness control, if you are worried about your surface being smooth.

To me Parallelism per unit makes as much sense as applying Tangent requirement to Flatness. Could that be considered “extension of principle”?

I also liked how you changed your argument from “from purely geometrical point of view there is a difference” to “Again, forget about math”. Well, whatever helps you thru the night.

Sorry, I promised to get out of this argument, but someone woke up my inner jackhole.

 

[powerhound]

Well I sure didn't mean for this thread to take this kind of turn. Let me add this:

1. We have a 48 X 48 base plate.
2. On this base plate sits a 20 X 20 plate with the functional assembly mounted to it.
3. All I care about is that the 20 X 20 plate sit parallel within .005 to the datum plane which is established by the bottom of the base plate.
4. There is no reason for the entire 48 X 48 surface to be parallel within .005 to the datum plane.
5. There is no reason any part of the surface to even be flat within .005.

I feel that the tangent plane modifier on parallelism is a logical extension of principle for this situation.

If there is a way to do what I'm trying to do that doesn't require an extension of principle like this, please clue me in. I'm all for using the available tools in the standard.

Powerhound, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[CheckerHater]

Tangent plane on parallelism is not extension of a principle, it is standard.

Parallelism per unit is not. But you can add flatness and /or surface roughness to demand your surface to be as smooth as possible.

It looks to me that you are just fine.

Sorry for unnecessarily heating up the discussion

 

[Belanger]

If something can be flat on a per-unit basis, why doesn't it make sense to allow parallelism on a per-unit basis? Sure, the flatness aspect is gone, but orientation of the imaginary, finite, local plane can still be toleranced.

This all reminds me of another thread from this past January where we debated the meaning of 2D and 3D relative to profile of a line -- including discussions of topology and references to obscure math definitions. Shrug.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[Belanger]

CH -- in your graphic posted at 6:24 this morning, the smaller plane that you've drawn is indeed a tangent plane.
But it's just the tangent plane within the local area that you've defined by virtue of the size of the small plate sitting down in there. That's the meaning we've been trying to say is probably OK, though not spelled out in the Y14.5 standard.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[CheckerHater]

No, it is not.

I would agree on the opposite, that part is tangent to the highest points of plate, but it’s not the same; it’s like swapping datum and related feature.

And why am I arguing anyway? I already have ANSI / ASME on my side

 

[pmarc]

CH,
I will try to answer to some of your statements/questions, though I have a feeling this will not convince you even a bit (but well, it wouldn't be me, if I hadn't try):

I didn’t find where it mentions “surface within considered area”, only the entire surface.
Illustration doesn’t show it on “per unit basis” either.
Y14.5-1 is no different.

You could not find it because it is simply not there. How can "extension of principles" be defined in the standard? If it was there, it would not be "extension of principles" but a rule.

About being “rather worried if the local area was convex”, I may assure you that your “plate” will rock anyway on either convex or concave surface, because there is no guarantee that surface will always match all 4 corners of your 1 x 1 area. So why not to extend principle even further and specify 1 x 1 x 1 triangular zone?

The surface does not have to match the corners. The surface has to rest on at least 3 high points. And it will always rest on at least 3 points - the problem is that those points may not always be the high ones. As for convex surfaces, it is interesting that both sources - the standard (para. 6.5) and the Tec-Ease's tip - make remarks that for this type of surface tangent plane concept has to be treated with some kind of caution.

To me Parallelism per unit makes as much sense as applying Tangent requirement to Flatness. Could that be considered “extension of principle”?

No, it couldn't. Tangent plane, by definition, is perfectly flat, so Flatness requirement would be useless, illogical, illegal or whatever you want to call it.

I also liked how you changed your argument from “from purely geometrical point of view there is a difference” to “Again, forget about math”. Well, whatever helps you thru the night.

Flatness or Flatness per unit tolerance zone is not tied in rotation to anything. Parallelism or Parallelism per unit tolerance zone is always parallel to datum plane. This is why I said that from purely geometrical point of view there was a difference. And I stick to it. If you do not like the wording, change it to "Y14.5's point of view" or "my point of view". I do not care. The clue was: "THERE IS A DIFFERENCE". No offence, but in my opinion you are simply nitpicking on a single word which isn't really crucial for the meaning of my response. You are just attempting to make it important for the purpose of proving that I am wrong or inconsequential. (Or maybe for the purpose of hiding that you are incapable to admit you are wrong in this case :-])

 

[Belanger]

CH -- I don't know how ANSI/ASME supports the exact point you're making. You claim that tangent plane is the highest points of the part. Fine; I agree. Now what about the tangent plane within a local area? You seem to be stopping short of that statement, which is the whole point of the discussion.

If you take the large block in your sketch and slice it down to a block of only the smaller size, then can we discuss a tangent plane? If so, what's the difference if all the other stuff on the block is still connected to that local area?

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[MechNorth]

I apologize for my error re tangent plane, gents (Evan, tks for contacting me). I've re-checked the standard and it does specify the 3 high-points; not sure where I got the other idea, but it does still make sense and has merit. In the T-E tip, the disc will move axially along the guide pocket until it hits the high point on the drive shaft and then be fastened down; where the 3-point tangent plane is may not be relevant. That, of course, presumes that the guide piece is adequately sized wrt the disc so that there isn't a lot of pitch & yaw. If there is more clearance on the guideway, then the 3-point tangent plane makes sense. Unfortunately there isn't any guidance on how to use a single-point / oriented tangent plane ... so far!!!

As I was looking over CH's graphic, I see it as a great support for "per unit area" application. Again, I know that it is being used in practice already, and it is very practical and necessary for the application.



Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[DeanD3W]

I too have run into an application that "Unit basis" Parallelism worked very well for. It's a perfectly valid extension of principles IMO. Six or eight years ago, when I encountered the application, I resisted its application a bit, but I got over that.. .

A tangent modifier can be applied for an entire feature or for the incremental control that a "Unit basis" tolerance provides. Any issue with a tangent plane for an incremental portion of a feature could also exist for a tangent plane from an entire feature.

Dean

 

[CheckerHater]

Pmarc,

And you call ME nit-picking? I just noticed that you embraced math when it suited you and then threw math into the window both in the same forum.

And about admitting being wrong; everyone is entitled to their opinions.
I am simply in agreement with the guy who happened to say:

“ASME Y14.5-2009 also mentions only about flatness & straightness applied on a unit basis.
My opinion is that only these two controls can be applied with this concept at all. I can imagine a lot of difficulties in correct interpretation of such concept when it is applied to a geometrical tolerance that has got any datum reference (e.g. parallelism, runout or whatever).”

I am not hiding anything.

 

[pmarc]

CH,
Of course everyone is entitled to their opinions, no doubt about it. After all, this is what this forum is for.
And believe me, I truely respect your opinion.

I have just one more question to you, if you don't mind:
If you agree with the opinion of the person you cited, could you tell me/us what are these "difficulties in correct interpretation of such concept when it is applied to a geometrical tolerance that has got any datum reference"? I think this is the clue here. I have an impression that so far we have not heard from you any argument that would really support the statement. I am not telling that there are no difficulties at all. Maybe the difficulties are so serious that the concept indeed cannot work with characteristics other than flatness or straighntess. I would simply like to hear about those difficulties from you.