"Per Unit Basis" For Roundness? 4

[courtjester140]

I have a problem with press fitting a circular component into another circular piece. They need to be as flush as possible, but I'm not sure how I can control it most effectively. I'll try to explain my problem as simply as possible.

Suppose I have two toilet paper rolls (and lets assume they're rigid). I want to press fit one toilet paper roll into the second toilet paper roll. I want the fit to be as flush as possible, so I don't want any gaps between the two rolls once they're mated.

I can use a total runout callout on the ID of the female toilet paper roll, but the tighter I make the tolerance, the more expensive it gets. There's also the risk that all of the runout will be on once side of the ID, while the rest will be fine, causing the roll to pass inspection but create a leakage path.

What I am looking for is something like total runout per unit basis, where that unit would be some number of degrees. So say the nominal total runout value would be 2.5, but for each 15 degree segment it can't exceed .5. That would solve all of my problems, but as far as I can tell it isn't valid.

What else can I do here? I want to minimize gaps between the two components. I also can't change the male component, so it has to be something for the ID of the female component.

Thanks for taking the time to read this!

 

[CheckerHater]

Circularity, or Cylindricity, depending on shape and size of the parts.

The tolerance zones are formed by concentric circles or cylinders respectively, so that prevents “bulge” from passing inspection. Cylindricity will also control Straightness.

If you are in ANSI/ASME part of the world you can rely on Envelope rule: your surface should always lie between LMC and MMC. So you just specify tolerance tight enough.

In ISO part of the world you can use (E) in circle symbol to invoke Envelope requirement.

I hope this will help.

 

[courtjester140]

Thanks for the reply! Circularity or concentricity would be fine, except I would need to use a very very tight tolerance to prevent any major bulges.

I'm looking for something that looks like the attached image. I'm sure this doesn't exist, but maybe there's something similar that I can use? I know it's possible while inspecting to check a tolerance for a specific range of degrees. Doing this would ensure the part follows a smooth path within the tolerance zone.

 

[ptruitt]

I agree with CheckerHater. You can probably specify ASME Y14.5-1994 or -2009 in your notes and give the size you want because of Rule #1 in ASME Y14.5-1994 or -2009. Have you read Rule #1? You could add 'Profile' to Circularity and Cylindricity as acceptable alternatives, although most folks on this forum will probably say that you only need to specify size if it is defined per a recent ASME Standard.

You can look in the Machinery's Handbook for recommended fits.

You might also need to add ASME B46.1-2002 or -2009 to your notes and specify a surface finish per that Standard that will prevent leaks. The surface finish may need to be developed through testing.

Peter Truitt
Minnesota

 

[Madhu454]

Hi,
I agree with ptruitt, surface finish of the parts have greater segnificance in providing proper seal between mating surfaces.

Also in the drawing attached by you, u have used datum for circularity control which is illegal. Roundness does not have any datum references.

Also there is no option in ASME to specify roundness on unit basis as shown in your drawing.
This is an illegal control.


Madhusudhan Veerappa
Mechanical Engineer

 

[CheckerHater]

Real Puzzler

Normally part surface can do anything as long as it stays within tolerance zone.

The only control more restrictive then that is Controlled Radius, it actually prevents “reversals”; but I have never seen it used on cylindrical parts.

I am 99.9% sure there is no standard symbology for what you are trying to accomplish, so you may have to create a verbal description like “Surface curvature not to be less than so-and-so”, but it may be difficult to actually measure.

I am out of ideas by now, but will follow this thread with great interest.

 

[courtjester140]

Madhu,
My mistake, I had concentricity in my head while writing that.

Surface finish won't solve the problem. The bulges would still be there, but with a smooth surface. The magnitude of surface finish compared to the tolerance zone doesn't compare enough.

If there are no "per unit basis" tolerances for circular surfaces, then I'll have to settle of a tight total runout. It will be expensive, but it would have to do.

ptruitt,
Are you saying that I should use MMC on the ID of the female part?

 

[ptruitt]

To the MMC question: No.

I found an example of a method which is probably legal, although I could not find anything quite like it in ASME Y14.5-2009. It is from a book written by a member of the ASME Y14.5-2009, I believe.

Peter Truitt
Minnesota

 

[Belanger]

The 2009 standard now allows something called NON-UNIFORM. By putting that word in the feature control frame, it opens up unlimited possibilites for whatever crazy tolerance zone you want. You just have to be very careful in defining the zone with basic dimensions. (I don't have Y14.5 in front of me so I don't know the paragraph or figure number.)

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

 

[ptruitt]

J-P,

My example is an equal-bilateral tolerance (uniform). I do wonder how it could be measured, though. As far as I know, SmartProfile is the only GD&T analysis software that conforms to the ASME Y14.5.1 math standard (according to Medtronic), but it does have limitations. I am guessing that it would not be able to do it.


Peter Truitt
Minnesota

 

[pmarc]

courtjester140,

Out of curiosity. You said male part can't be modified, but could you share with any simplified sketch of it so we could see how it is dimensioned and toleranced?

I am asking because if dimensioning on male part is not grasping intended function as well as on female part or - even worse - if the dimensioning is anyhow vague or incomplete, whatever you specify on female side will not make much sense since the failure (leakage) will most likely be caused by male side.

 

[courtjester140]

pmarc,
I don't have the drawings in front of me since I'm at home at this point, but attached is a finely crafted piece of art from MS Paint that illustrates what's going on.

The intent here is for the softer graphite to conform to any imperfections in the shell. The back surface is bonded with a non-porous epoxy. The problem with this particular piece is that it's larger than anything else we've done, so it seems small leakage problems that we didn't detect before are now amplified.

changing the male part won't really help here, since it's role is to conform to the shell's shape anyway. Without going into too much detail, leakage through the mating surfaces needs to be minimized.

 

[DeanD3W]

The concentricity on the male part will allow a lobed shape, as long as there are an even number of lobes, so that would be one concern.

It has been pointed out that Circularity cannot be applied on a unit basis (incrementally)... That is not true. Y14.5 provides concepts and those concepts can absolutely be extended to applications other than those shown in examples provided by the standard. I think cylindricity applied on a unit basis may work well. A size or profile tolerance can be used to control size and a form refinement on a unit basis, even with an angular increment, should be fine. Something like |Cylindricity|Tol Value/x degrees|. I think the measurement results may require some manual analysis, but it is possible that asking for cylindricity with only a limited data sample, from a given increment, may work.

This could even be done with a linear increment, so with a distance in the axial direction. The effect could be combined and turned into a square or round patch on the surface, with the cylindricity value within any patch required to be within the specified very tight tolerance value. If you want a fixed curvature tolerance zone a unit basis profile of a surface would provide this. If you want a variable curvature tolerance zone the unit basis cylindricity can be used.

Please trust me on this. There is absolutely no problem with extending a principle shown in the standard to other applications.

Dean

http://www.d3w-engineering.com

 

[SeasonLee]

courtjester140

We have seen straightness, flatness and profile application on a unit basis, and we haven’t seen any “per unit area” used on roundness. The “per unit area” applied on a profile surface means to refine the original requirements, it will limits abrupt changes on the surface such as bumps, dimples, but this application may not solve your leakage problem between the mating surfaces.

I will not to settle a tight total runout as you mentioned, since total runout is a surface to axis control only, it will control the form (cylindericity), orientation and axis offset, but not for the size control, so I will recommend to use multiple single-segment profile control on your application :

1. The upper segment is a profile of a surface control, it controls size, location and orientation relative to the DRF.
2. The lower segment is a profile of a line control, it controls the line elements for form and orientation relative to datum axis, it’s a refinement of surface profile control, the tolerance is less than the upper segment.

This design, the profile of a surface controls limits the size and location of the cylinder surface, the profile of a line control limits the form and orientation of the line element of the cylinder.

SeasonLee

 

[ctopher]

But the two toilet paper rolls are the same size...




Chris
SolidWorks 10 SP4.0
ctopher's home
SolidWorks Legion

 

[DeanD3W]

Season Lee,
Multiple single segment feature control frames really shouldn't be referred to as an upper and lower segment. They're not segments. Each is a feature control frame which imposes a tolerance just like any other.

The upper feature control frame you describe, with profile of a surface, will control size, form, orientation, and location, assuming capable datum features are referenced. I don't know why any datum features would be referenced for courtjester140's case though, since I believe he is dealing with a single cylindrical feature's form and size only.

The lower feature control frame you describe, with profile of a line, will impose tolerance zones on each cross section of the feature and those tolerance zones will control size, & form of those cross sections... You left off the size control that will be present. Also, since profile of a line cannot detect steps in a the considered feature from one axial location to the next, it doesn't seem like a good candidate to achieve the seal courtjester140 needs.

A unit basis cylindricity or unit basis profile of a surface would likely address the need though.

Dean

http://www.d3w-engineering.com

 

[SeasonLee]

Dean

Multiple single segment feature control frames is very common on GD&T, its easy for us to say the upper or lower segment, I just follow what Alex said, you may find out this term from Advanced Concepts of GD&T by Alex Krulikowski on oage 25-2, you may also find out the same usage on page 20-2, I can send you a snapshot if you need them.

On the upper FCF, you are right , relative to “DRF” better update to relative to “datum axis”.

But, I am interesting on what you said “profile of line cannot detect steps in the considered feature from axial location”, each cross section will limited the tolerance within the specified tolerance zone, it’s a series of cross section to control the size to a tighter tolerance zone, I can’t image why cannot avoid the steps.

If the step is a concern on the axial direction, we may use straightness control instead on the cylindrical surface, my multiple single-segment FCF will updated to :
1. The upper segment is a profile of a surface control
2. The lower segment is a straightness control

SeasonLee

 

[courtjester140]

All,
I should clarify that the concern is bumps in the radial direction, not axial. Inspecting a component, that's where these bumps come from. Really I'm just having trouble putting this on a B/P.

Even though I say that a radial direction is my concern, that doesn't mean that I can omit the axial direction as well...although controlling axial imperfections is much easier.

 

[DeanD3W]

SeasonLee,
I'm completely aware that some people use the terms "upper and lower segment" with regard to stacking the feature control frames. I'm telling you that it's a poor practice to do so. The point is that there is no special meaning at all with regard to this stacking... One could "un-stack" the feature control frames and apply them with separate leaders and they will have the same meaning as when they were stacked. The requirements for each feature control frame are completely independent of the others. You may choose to create a series of controls that have tighter values and fewer degrees of freedom constrained as you progress down the stack, but that doesn't mean that you should refer to any of those feature control frames as a "segment". To do so implies a systematic link, when there is none.

With composite feature control frames there are restrictions regarding what may be specified with the 2nd or lower segment(s) that are based upon the upper segment, so the segments are not independent of each other. So, for composite feature control frames the "upper segment" and "lower segment" terminology has some meaning, but if you're not using a composite feature control frame no such meaning exists.

Regarding the inability of profile of a line to detect steps, that is true if no datum feature which constrains translation normal to the considered feature's surface are referenced in the particular call-out... The tolerance zone for each cross-section is fully independent, wrt translation, of those for adjacent cross-sections, so there is no ability to detect a step. If you reference datum features such that the tolerance zones don't have that freedom, then the tolerance zones form a fixed stack that creates the equivalent of profile of a surface. So, when profile of a line is used in a way that justifies its existence (when the tolerance zones have some degrees of freedom remaining, so they don't stack into a profile of a surface equivalent) then steps in the considered feature will not be detected.

Another problem for the specific case proposed is that the profile of a line in the lower feature control frame will control size, so the effect intended will not be present.

Dean

http://www.d3w-engineering.com

 

[Belanger]

Dean, I think the question arose because of a slight inconsistency in your earlier post. You wrote: "Multiple single segment feature control frames really shouldn't be referred to as an upper and lower segment. They're not segments."
Why can't we called them segments if it is a multiple single segment feature control frame? And if they are indeed called segments, it's not a huge leap to resort to the terms of upper and lower simply for ease of discussion, though you are correct that there is no hierarchy.

When different symbols are involved (the example given was profile of a surface and then profile of a line), they really shouldn't be called multiple single segment feature control frames anyhow... I think this was your main point.

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