Can a profile tolerance used to control a true position? 2

[jcastaneda79]

A current production cam cover has 11x mounting holes. This holes true position is not in drawing, instead we got a 1.0mm profile tolerance to control this position... can this be done? it is a mistake? thanks

 

[KENAT]

It's perfectly legal per ASME Y14.5M-1994 although whether it's most appropriate based on function, other users of the drawings etc. may be an issue.

Profile tolerance controls size, form & location.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

What is Engineering anyway: faq1088-1484

 

[PaulJackson]

Is the cam cover a casting or molding and are the holes "as cast or as molded"?

paul

 

[drawoh]

jcastaneda79,

If a profile tolerance is applied to a hole, the thing can be interpreted as per ASME Y14.5M-1994, and inspected. I have even done this on one or two occasions.

Whether it is good practise, or the optimal way to dimension your drawing, is more open to question.

I like to apply to holes zero positional tolerances at MMC. An alternative to this would be to apply a unilateral profile tolerance with the guide line outside the hole. See Figure[ ]6.5.1 in the standard.

Either way, you are clearly defining an MMC condition, which is the important thing for a clearance hole.

JHG

 

[jcastaneda79]

The cam cover is as cast, so in this case, the 1.0 mm can be interpreted as I can move the hole´s center 0.5 to one side and 0.5 to the other side? thanks

 

[axym]

jcastaneda79,

That's not the correct interpretation. The 1.0 mm profile tolerance doesn't directly control the hole's center, it controls the hole's surface.

The tolerance zone is a tube-shaped shell 1.0 mm thick, straddling the nominal surface of the hole. The OD of the tube will be the hole's basic diameter plus 0.5 mm, and the ID of the tube will be the basic diameter minus 0.5 mm.

In this case, the profile zone controls the size, form, orientation and location of the hole.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[PaulJackson]

drawoh,

If one was to relate a variable position tolerance to a profile tolerance of a hole... the position tolerance would have to be zero @ MMC, zero @ LMC, and maximum at nominal.

See the attached paper.

paul

 

[drawoh]

PaulJackson,

I agree that the two dimension conditons I described are not identical. In most cases, the MMC state is much more important than the others.

The accurate position at LMC might be what I actually want! Think about a hex socket head cap screw clamping down on the maximum sized hole.

JHG

 

[dingy2]

Jcastenida79:

Firstly, are the diameters of the holes in question shown in a basic dimension? If they are not, then I think we may have a design error here since the true profile of each hole should be basic. I would then ask for clarification from the Designer.

Please suggest that holes that have a function or relationship with the mating part should be shown in positional rather than with profile of a surface and also suggest at MMC. It just is more appropriate and as you can see from some of the answers, it can get a bit problematic using profiles.

Dave D.

http://www.qmsi.ca

 

[PaulJackson]

Jcastenida,

There is nothing wrong with having the cored holes of the casting toleranced the same as the cast surface contour.

Dave is right that if the hole sizes are toleranced rather than basic then there is a problem... but... I don't know what would lead him to suspect that?

The only thing "a bit problematic" as Dave puts it, about measuring cored hole's is that people typically want to see the cored hole's size and coordinate location data rather than its contour points data. Not that it would do them much good in modifying the mold contour or core pin sizes to achieve contour conformance post shrink.

Once they have that size, X and Y data they are perplexed about how to relate it to the profile tolerance. I have seen quite a few negotiated substitute specs to constrain size and position but none of them captured the liberties that size and location have within the profile boundaries. All of them proportioned the profile tolerance into fixed amounts for size and either fixed or variable amounts for position. That is why I created that paper to show how to figure a position tolerance for any measured core feature size.

paul

 

[CheckerRon]

I would stick to the recommendation in paragraph 2.1.1.1, and only use position for features of size like round holes.

AXYM (Evan J.) has a good point in noting that you are controlling the tubular tolerance zone of the holes and ignoring the centering of the feature as position would do

 

[MechNorth]

For what it's worth ...
Some people have prophilaphobia ... the fear of profile controls. I tend to lean the other way, and try to use profile controls whenever possible. I'm repeatedly faced with situations in a broad spectrum of industries where people default to a size tolerance and position control at MMC, when they are actually concerned about boundaries rather than individual attributes. By that, I mean a separation of the size & position aspects of the feature.

In so many cases, a profile boundary that effectively controls the functional limits (as pointed out above, not the exact equivalent of the size & MMC position boundary) is far easier to verify than separate size & position controls. I haven't come across an example yet where using a surface profile in place of size & MMC tolerance cost more or was more difficult to verify. On the otherhand, some people want SPC data and aren't adequately familiar with their CMM or digital inspection equipment software to draw the data out of a profile inspection protocol. Don't just shy away from profile because you aren't well versed in it. Take the step to get better versed & gain the value of profile controls.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Jim:

So one has, maybe, 8 holes shown in positional tolerances at MMC with a size tolerance. A checking fixture of at virtual conditions size can easily check the features with a minute or two and could be included in a Control Plan for the Operator.

You state that having a profile tolerance on each hole would not cost more to confirm. Could you please reflect the checking method and estimated cost?

Thanks

Dave D.

http://www.qmsi.ca

 

[PaulJackson]

It is always best to consider function when selecting the controls and tolerancing the features. There is certainly merit to your counsel Dave since the 11 holes are most likely fastener clearance holes and consequently do tolerate greater position deviation as their size increases. If the holes were processed for size and location following the casting process I would agree with you wholeheartedly. Aggressive attribute gauging likewise could be appropriate as a “goal post” check on the process since the variable limit position tolerances (MMC) are not typically addressed in statistical conclusions from variables data measurement.

Are there however other considerations that would lead the designer to functionally choose one common profile tolerance “all over” to control the cam cover casting? Jcastenida79 said “The cam cover is as cast” and if we can assume that little or no machining is required beyond gate snag grinding and trimming then the profile that results from the die casting along with its blow, shift, shrink variation, and tooling wear is all we have to measure. The inner and outer contours of the cover along with the fastener hole cores are produced in one operation and result in one contoured profile.

If I were designing this part so that it required limited or no subsequent machining to the processed casting then I would probably do it the same way with a single profile note “all over” tolerancing the contour +/- from its basic. If the gasket surface (possibly with its “as cast” channel to accommodate the RTV sealer) required a refinement of that “all over” tolerance I would probably constrain it with a separate profile callout controlling its flatness. If a surface refinement operation was necessary so be it. The advantage to limited or no machining is $$$$... you don’t have to buy, facilitate, staff, or dedicate process time to machines to do subsequent processing. At the same time you don’t risk exposing acceptable levels of porosity and therefore risking leakage. The measurement of the profile could be confined to strategic points on the casting, measured on a CMM, and the deviations illustrated with “wisker plots.” Each of those points could easily be evaluated for its statistical control and capability to satisfy quality requirements, and if the tool makers wanted size and position inspection results to modify/adjust core pin sizes or locations they could use the simple equation that I posted above to examine the difference between measured position and the allowable position deviation based on the measured core size.

The only way to legitimize the use of a hard attribute gauge would be to divvy the profile tolerance into proportions for size and position @ MMC and then employ an attribute process sampling strategy which means enormous sample sizes to detect “goal post” conformity. Also divvying up the profile tolerance so that an LMC hole and its maximum position deviation is equivalent to the LMB (least material boundary) of the profile tolerance means that the separate tolerances have to be reduced which may force subsequent machining (not feasible for “as cast”) and cost more $$$$.

paul

 

[dingy2]

Paul:

An attribute gauge on profile of a surface is not appropriate since it only covers outer boundary and not perpendicularity to datum A or straightness on the "Z" axis.

The best method with a profile of a line or surface is the use of CMM and plotted points covering all the requirements.



Dave D.

http://www.qmsi.ca

 

[MechNorth]

Dave,
Depending on datum precedence, a vision system is pretty quick & easy for inspecting such features. Actual inspection time is a fraction of what it takes to engage a hard gage, and is less susceptible to wear and misuse, plus it can give you back attribute data, not just go/no-go. A hard gage, to moderate gage maker tolerances, would still be several thousand dollars, plus regular calibration & maintenance costs. A hard gage can only be used on that one product or those products with identical features and datums. The vision system may cost into the tens of thousands (there is great variance on price & abilities) but it is usable for a greater variety of parts & features by programming, and more easily integrated into production for every-part inspection.

Contact CMMs can provide greater flexibility than vision systems, which again may be useful depending on the datum precedence, however their costs are typically orders of magnitude higher still, so the payback period is significantlly longer. Also contact CMMs typically take far longer for data acquisition than vision systems.

Each inspection / qualification method has its merits and its drawbacks. Increasingly I see that companies want maximum flexibility in their system rather than tying it into a specific product / manufacturing process. In this particular case, there are too many unknowns to establish any single process as best.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Jim:

Vision systems appear to be 2 dimensional while profile of a surface is 3 dimensional. In the example of 8 holes using profile of a surface, I can see the vision system (template as an example) controlling the inner boundary of a hole but not the outer boundary.

I also have a hard time envisioning a vision system controlling the angularity of a hole to the primary datum. Well, maybe it could by reflecting half a hole rather than the full circumference but it is still suspect.

Do you think that there would be a vision system on the shop floor for the Operator to use? One by every machine? I am afraid that this type of equipment is more lab equipment rather than for shop floor use just like a CMM.

Dave D.

http://www.qmsi.ca

 

[drawoh]

dingy2,

I can work this one out. Take the case that I have two plates to be attached by M6 screws with positional tolerances of Ø0.3mm.

My first plate's holes are dimensioned at Ø6.9/6.3 and positioned to[ ]0 at[ ]MMC. I am ignoring the possibility that the tapped holes are out of perpendicular, so hopefully it is a thin plate. If the Ø0.3mm positional tolerance of the tapped hole is a reasonable expectation of what the fabricator can do, then the hole will have to be drilled at least Ø6.6mm. The fabricator will have a tolerance on their hole of Ø6.9/6.6mm. If they can locate the hole more accurately, they can open up their drill tolerance.

My second plate has Ø6.6mm holes and a profile tolerance of 0.3mm. The maximum allowance for diameter if Ø6.9/6.3. The sloppiest allowable positional error is at Ø6.6mm, and it works out to Ø0.3mm. A fabricated diameter of Ø6.65/6.45 will meet specification if it is positioned within Ø0.15mm. Again, the fabricator can balance positional accuracy against drilling accuracy for optimal tolerances.

The positional tolerance, not surprisingly, requires more accuracy by the fabricator. Then again, these are two different geometries, and they satisfy different requirements. If I am using large headed screws and/or flat washers, I am going to call up the positional tolerance. An M6 hex socket head cap screw has a head diameter of 10mm, so I would stick with the positional tolerance. If this were a 1/4-20UNC hole and cap screw (Ø.375"[ ]head) I would consider a profile tolerance. If we were using shoulder screws, we need the profile tolerance to control the LMC state.

You can always go back and position the tapped holes more accurately.

I am a designer, not a machinist. If any machinists would like to suggest real numbers for fabrication capability...

Designing GO gauges to test the MMC of these holes is easy. Fabricating them is more of a challenge. The NOGO gauge for the LMC profiles could be interesting.

JHG

 

[MechNorth]

Dave,
For a thick casting, a vision system is definitely limited. For relatively thin castings, moldings, extrusions, and where good control of the casting/forging/molding is present, vision systems are commonly used. They are best at gaging inner boundary, which is comparable to an MMC modifier on the tolerance. Neither vision nor contact CMMs do particularly well with LMC when there are decently formed surfaces (i.e. no significant voids, porosity, sinks), though contact systems do approximate the outer boundary better when there are poorly formed surfaces.

For holes designed perpendicular to the primary datum, vision systems more closely approximate the Y14.5 standard which specifies the use of the true geometric counterpart, perpendicular to the datum. Contact systems tend to have significantly higher variance in this situation, depending on whether your probe tip type (ball vs cylinder).

Vision systems are very common on production floors, not just in labs, though not on a 1:1 availability ratio. Granted, I haven't seen them in an automotive casting facility, but I have seen them used in plants machining automotive, aerospace, electronics and consumer goods, and they're used extensively in molding companies. I've seen a single system used for 4 to 6 machines, each producing different parts and therefore running different programs. Even when inspecting every part, the production rate per machine (in those situations) allowed individual part inspection by all users. Increasingly, batches of parts are sent thru an automated vision system to check a variety of features without any human intervention beyond loading & unloading pallets. I'm also seeing greater integration of vision & contact systems so that you get the speed & flexibility needed.

I've also seen contact systems out on the work floor, though nominally isolated from the surrounding environment by heavyweight plastic curtains, isolation bases and overpressurized a/c systems.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com