holding roundness control

[Paul H]

Hi, As you will see I'm no expert on GD&T but I have been involved in a discussion to try to reduce the number of concessions or scraped parts we produce as this has existed for many years as different manufacturing sites can interpret the drawing in different ways so the design intent isn't always clear. Some design offices ignore the 'rule' and allow a larger roundness tolerance than the size tolerance but other limit this to the max size control (which makes it harder to achieve!) and so we don't have a Standard practice across our design groups which is our ultimate aim.

A bit of history first so you know our issues. We produce mechanical seals and the sleeve component must comply to a very tight tolerance (API) over the pump shaft (typically F7 h6) so the bore size of our sleeve is often controlled to +/- 0.013mm. Controlling this size is not the issue but maintaining a reasonable roundness control is the problem. If I understand this correctly and if we follow the rules of GD&T (ANSI or ISO)the roundness tolerance cannot be > than the size tolerance so as in this example the roundness cannot exceed 0.013mm.
The actual components are typically quite thin (<4mm section) and are normally just turned (not ground) so manufacturing struggle to hold such tight tolerances. typically they can maintain a minimum of around 0.03mm tolerance on roundness so my question is: is it possible to apply an 'average' size control (as we know there is enough clearance between the shaft and the sleeve bore) and allow the roundness tolerance to be larger then the size tolerance (as per the attached)?
Appreciate any help or advice you guys can offer

 

[3DDave]

A better control is by perimeter and curvature. Perimeter tells what the average fit will be when it forms into shape and curvature tells whether the part will be over-stressed trying to bend to the mating curvature. ASME Y14.5 have a specification for either as part of its definition; I don't know if there's anything in the ISO standard

the 'Y14.5 does make mention of AVG for the diameter to allow a looser circularity than allowed by size but it assumes the variation is nearly nothing and does say that a peripheral measurement might be used. This leaves open the chance for a kink in the section which specifying curvature (1/radius) controls. I consider curvature an instantaneous property, unlike radius which applies over a span.

You didn't select the attachment correctly and when you do, omit special characters such as "&" from the file name.

 

[Paul H]

 

[pmarc]

This comes down to a question what standard governs the prints.

If it is ASME Y14.5, then by default Rule #1 controls roundness of a feature within the specified limits of size and any roundness tolerance greater than the size tolerance is redundant.

If it is ISO, then default rule is indepedency of feature's size and form and the roundness tolerance can be greater than the size tolerance.

 

[greenimi]

Paul H,
I would say that in your example,
“the roundness tolerance cannot be > than the size tolerance so as in this example the roundness cannot exceed [highlight #CC0000]2X[/highlight]0.013mm.”
Because that is the size tolerance (0.026).

"Free-state variation is the distortion of a part after removal of forces applied during manufacture. This distortion is principally due to weight and flexibility of the part and the release of internal stresses resulting from
fabrication."

Pmarc,
If the OP is trying to use AVG note as per “so my question is: is it possible to apply an 'average' size control” does it automatically mean that the part is subject to free state variation?
(free state= the condition of a part free of applied forces.)

So, the default rule#1 is nullified and ASME or ISO becomes irrelevant.

Just asking…… Please be gentle (as usually, you are)

 

[pmarc]

greenimi,
Although AVG diameter concept is explained in Y14.5 in the context of free state variation, I wouldn't say its use must automatically mean the part is subject to free state variation. Unfortunately, because it is shown in the standard in the context of free state variation, its applicability to "rigid" parts is kind of a gray area.

Good point on the total amount of roundness tolerance available in OP example (assuming the explicit roundness tolerance is not there). It is not 0.013, but 0.026.

 

[Paul H]

Thanks for the advise but I still need a couple of points clarified. In the example above the MMC for the sleeve bore diameter is 38.027 (38.04 - 0.013) agreed? If so is it fair to say that if a part was manufactured at MMC the form (roundness tolerance) would be zero i.e. perfect form.
If the part was produced at LMC i.e. 38.053 then the max roundness (0.026) could be applied?
If that is the case then the roundness tolerance cannot be a fixed value as it is dependant on the actual size?
As I stated earlier we are following ISO rules for GD&T so want to know if we can legitimately have independent tolerance for size and form BUT allow a greater tolerance for form as the size tolerance is so tight (typically +/- 0.013)that we struggle to hold that for roundness using conventional CNC equipment. To achieve this we would need to grind which obviously we want to avoid for time and cost reasons. We know we have a reasonable clearance between our sleeve component and the customer shaft so can get away with a slightly more generous roundness value but don't want to break any GD&T laws.

 

[3DDave]

AFAIK ISO doesn't depend on the Envelope Principle unless it is exactly specified. I believe this leaves the full roundness tolerance available regardless of feature size. To be sure you should have a copy of ISO 1101 and ISO 12181-1 & 2 at the relevant version.

 

[pmarc]

PaulH,
What you said is true for ASME.

You are following ISO, so the rules are different. In ISO there is no dependency between size and form of the feature of size in case of the callout you have shown, because the envelope requirement (known in ASME as Rule #1) is not default condition. This means that the roundness tolerance value can be greater than the size tolerance.

The callout you have wouldn't be legal if circle E modifier followed the size specification or if letter E was used somewhere near the drawing title block indicating that the envelope requirement was default for the drawing.

 

[Paul H]

Can I query what was said earlier. You wrote: "the roundness tolerance cannot be > than the size tolerance so as in this example the roundness cannot exceed 2X0.013mm.”
Because that is the size tolerance (0.026)." I can't see how that is true? If you look at the simple picture below:
Surely it can only be half the diameter tolerance i.e. 0.13mm and not 0.26mm as suggested(original example) or have I misunderstood?

 

[Belanger]

Paul, your picture didn't get attached. But to answer your question, in ASME the rule is that the roundness tolerance can't exceed the full diameter tolerance (0.026). See my attached graphic to show why: People often assume that the max and min size limits of the diameter must be coaxial, but that's not required. The squiggly line showing the actual part outline can use up the entire diameter tolerance on one side of the circle.

 

[greenimi]

Although AVG diameter concept is explained in Y14.5 in the context of free state variation, I wouldn't say its use must automatically mean the part is subject to free state variation. Unfortunately, because it is shown in the standard in the context of free state variation, its applicability to "rigid" parts is kind of a gray area.

Pmarc,
Interesting enough in the new Y14.5-2018 standard F symbol (circularity) has been removed from its equivalent figure (5-13 , 2009, page 97). Do you have any idea why? Probably because “belts and suspenders” concept? Just a guess.
Also the next figure (5-14, 2009 page 97) the F symbol is kept. And both of these above mentioned figures have AVG symbol along with their size dimension.
And even more interesting, new standard revision, stating that invoking average diameter constitutes an exception of rule #1, along with the fact that average diameter should be specified for parts that are flexible in a nonrestrained condition, but not limited to such cases.
So, my follow up question is: why F was not removed from the circularity FCF from fig 5-14?
What do you think?

 

[Paul H]

below is the example:

 

[Belanger]

Now I see your graphic. But read my post -- we can't assume that the MMC and LMC of the circle are coaxial. Thus, all of the roundness error could be lumped into one big lobe on one side.

 

[pmarc]

Paul_H,
To back up what J-P said, this is how actual roundess error can be equal to the total size tolerance.