Subject: Shaft Tolerance Question

James,

A customer supplied me with a shaft print with the following straightness and circularity call out. The tolerance in question is circled in red.

I feel that this callout is incorrect and not representing what the customer is actually wanting. The final assembly of this design will have two bearings pressed into place at the .392 and 1.957 dimensions. The customer wants these two diameters to be in line so that the bearings are properly aligned.

I think a better option would be to make the .1876 diameter on the left Datum A and the .1876 diameter on the right Datum B. Then we would specify a runout of .0002 to Datum A-B. This would allow us to remove the straightness tolerance and keep the circularity tolerance. If the diameter and roundness are within tolerance then the straightness has to be good. At the same time we will be control the location of the two diameter axis with the runout tolerance.
What are your thoughts on this?


Stuart

Stuart,

According to Rule #1 in the Y14.5 standard on Dimensioning and Tolerancing the size tolerance controls the surface form on all rigid parts. Since the size tolerance is a diameter of .0002, it already controls cylindricity (roundness, straightness and taper) to within a diameter of .0002. This is less than the straightness control of .0003 and the roundness control which is .0002 per side or a diameter of .0004. So, both the straightness and roundness geometric controls should be removed.

Making one of the datum target diameters A and the other B, then controlling more diameters to A-B would mean the same as calling both diameters datum targets (A1 and A2, as it currently does) and controlling other diameters to datum axis A. It wouldn't mean anything different than is currently on the drawing.

The real problem with this part definition is that there is no coaxiality control between the two .1875-.1877 diameters. They could be infinitely out of coaxiality to one another. They need one diameter to have a runout control to the other or they both need to be held to a runout control to their common axis. This could be accomplished by: 1) naming one of them A and the other one B and then having two leader lines from one runout control point to both of the diameters and say runout to A-B, or 2) Making one diameter A and allowing its size to control its form, then making the other diameter B and giving B a runout control back to A or 3) Give both diameters a runout control to datum axis A (which is currently formed by datum targets A1 and A2).

Another interesting fact is that they gave the surfaces on which the datum target circular line elements reside a straightness control. Maybe they thought that this would somehow align the two surfaces to one another, like a runout control would. It doesn’t. It just makes each individually straight.

Hope this helps.

Jim

Subject: Book differences

Mr. Meadows,

A co-worker of mine has your yellow hard cover book "Applications and Techniques for use in Design, Manufacturing and Inspection", I believe it’s based on the 1994 standard, 624 pages... What is the difference between that book and the new 2009 standard book "Applications, Analysis, and Measurement book? 574 pages... (Other than the new standard information). I do like the content of the "Yellow" book. I look forward to your response.

Best Regards,

Daniel


Daniel,

The yellow book was very successful and I’m proud of it. It went into 15 printings and sold all over the world. It is thorough and, if you have a little knowledge already, is easy to follow. The only knock on it that I heard over the years was that it didn't begin with the most basic principles and progress slowly to the more advanced principles. It explained everything in the beginning and then dissected the topics in greater detail as it progressed. Some of those who had absolutely no knowledge were intimidated by that. Others loved it.

When I wrote the new grey book, I kept that in mind and tried to progress more slowly from basic to advanced topics. I would also say that the sections of the yellow book that I got the most praise on were the sections that dealt with detailed explanations of how to progress through tolerancing mating parts and how to read geometric controls as though they are sentences. So, in the new gray book, I expanded those sections. I also added chapters on how to do a tolerance stack-up analysis in a couple of different ways and how to calculate statistical tolerances. The grey book, therefore, has addressed some of the harder GD&T principles (toward the end of the book) that were not addressed in the yellow book. There are also a few color illustrations in the grey book.

I believe either book will teach the reader GD&T or act as a good reference to answer any questions one might have. But GD&T has evolved since I wrote the yellow book and my knowledge has expanded. So, if you want the latest information, written to the latest standards, explained in a slightly clearer manner, by a slightly more knowledgeable person, I'd suggest the grey book.

Thanks for writing.

Jim

Subject: Basic GD & T question

Jim,

I have a basic question; however I would like your thoughts to reinforce my interpretation. Picture a cube with a hole though the center. The primary datum is the face that is pierced with the cylinder, the secondary datum is the cylinder, and the tertiary datum is one of the sides that are perpendicular to the primary datum. If a feature were true positioned using the above datum structure, would the "X,Y" dimensions originate from the cylinder (Secondary datum) or would the tertiary control the dimension that is perpendicular with it. I believe the cylinder would control X and Y and the tertiary is merely controlling rotation, as you said in the class I took, if the secondary gets there first, it gets the job, what is the true interpretation..

Best Regards,

Steven

Steven,

Your interpretation of this situation is exactly correct. The tertiary datum is only controlling rotation. It is what we call an angular orientation datum.

Jim
 

Subject: Mr. Meadows - Help please regarding functional gage definition and gaging to mating component?

Mr. Meadows,

I'm sorry to bother you, but I have an extremely important question possible pertaining to gaging that I'm seeking final resolution to.

Per ASME Y14.43-2003:

Functional Gage: A fixed limit gage used to verify virtual condition boundaries (MMC Concept) generated by the collective effect of the feature's maximum material condition and the applicable geometric tolerance at MMC size.

Many people seem to regularly misinterpret the definition of a functional gage to mean the mating part, or a gage based on the mating part.

Per everything I've read and been taught, a gage should be based on the features and tolerances of the component or assembly being gaged, and NOT based a mating component features or tolerances.

A typical example I see is an instrument that mates with/engages a screw implant.

From what I know, the instrument gage should be based on the instrument itself - GD&T, tolerances, boundaries, etc.

Correspondingly, a gage for the implant should be based on the implant GD&T, tolerances, and worst case boundaries.

Many times though, requests are received to make a gage for an instrument based on the MMC, LMC, and/or virtual conditions of the mating implant and vice-versa.
(Usually due to the part prints not having any GD&T and nobody having the time or resources to do a correct revision)

If you would please clarify once-and-for-all what gaging practice is correct and where it is clearly stated in any standards or examples shown anywhere, I'll be forever in your debt.

Thank you very much!

Sincerely,

Joe

Joe,

Anyone who thinks a functional gage is supposed to base its design on the mating part drawing isn't reading the Y14.43 standard on Dimensioning and Tolerancing Principles for Gages and Fixtures. Mating part design (not gage design), dimensioning and tolerancing has everything to do with the configuration and numbers on the mating part. Gage design, dimensioning and tolerancing are solely based on the part definition you are trying to gage.

There is no mention in any pertinent standard (or non-pertinent standard) that says the gage is based on the mating part. It is based on the part you are gaging. Often the gage ends up looking like the mating part because the gage is designed to be the inverse of the part being gaged. That just means that a hole being gaged is gaged with a gage pin. Datum features and features being gaged are represented by a gage element that is the shape of the feature it simulates, but holes are gaged by pins and pins are gaged by holes.

It's kind of shocking that anyone believes the gage is designed from the part that mates with the part that is being gaged. I can't imagine where they would get such an idea.

Even though ASME Y14.5 is not a gaging standard (which it says on the very first page), it shows a few examples of gages and fixtures. All of these examples of gages and fixtures are based on the shapes and sizes of the part being gaged. The ASME Y14.43 standard (which I chair) is the standard on gages and fixtures and every section shows that the gage is based on the virtual condition of the features being gaged (when the part features are referenced at MMC). This is also true of the datum features of size on the part, unless the datum features are referenced at regardless of feature size (what it was called prior to Y14.5-2009) or as it's now called RMB (regardless of material boundary). Then the gage pins expand and the gage holes contract to engage the datum features.

Anyone who doesn't know this simply hasn't read the Y14.5 or the Y14.43 standard in any depth or maybe hasn't read them at all. All I can tell you is to give them a copy of Y14.43 and ask them to look at every illustration in Appendix B where the part is depicted and the gage follows beneath designed from the part it is gaging. They will see that the gage is designed at the virtual conditions derived from the part being gaged. Then they can reinforce that by going back and reading the words in the text of Y14.43.

I'm always surprised when professionals in an area like gage design, dimensioning and tolerancing are unaware of even the most basic of rules that are stated in standards and apply to exactly what they are doing.

If part prints don't use geometric tolerancing as defined in Y14.5, I can understand more the ignorance of those about the rules. Some companies continue to turn out poorly defined products because they just refuse to learn the rules, principles and formulas in our standards for tolerancing parts with less ambiguity. It takes time, but it's definitely worth it for the specificity it gives manufacturing and inspection and (to address your situation) gage designers.

James Meadows
Chairman ASME Y14.43-Dimensioning and Tolerancing Principles for Gages and Fixtures

Mr. Meadows,

Thank you very much for the reply.

I never encountered any misinterpretations in the auto industry, but in the medical industry where instruments and implants are supposed to interface and work together, it seems that the term "functional gage" is what many engineers misinterpret.

Many times I'll get a gage request where an off-the-shelf implant has been used to gage the instrument, and now they want a gage to replace it.

What they'll ask for though is a gage designed to the MMC of the mating implant and not the instrument.

This is usually because the instrument print is old with no GD&T at all, and there are not resources to update it.

I encounter situations like this on a regular basis.

What I'm still searching for that "golden" paragraph or illustration in any GD&T or gaging textbook or standard, that clearly defines or states what a gage should (or should not) be based on in situations like shown in the attached pictures.

I want to be able to say to somebody "Absolutely, positively, definitely, irrefutably, a gage should (always?) be based on "X" and never based on "X" because _____________.
My being able to resolve this issue would greatly impact my company in a positive way, and I'd very much appreciate any clarity you could lend to the situation.

Thank you again for your time Mr. Meadows.

Sincerely,

Joe

Joe,

No matter what passage you find and no matter how definitive it is, people will try to interpret it to mean what they need it to mean to prove whatever misguided idea they are promoting.

The following definition is from Y14.43:

4.2.2 Functional Gages. Functional gages are made relative to the virtual condition (MMC concept)of the feature(s) they gage. Functional gages check for a violation of the virtual condition boundary (MMC concept).

There are many pertinent passages and illustrations in Y14.43 that say and illustrate the same thing. But I imagine those you are talking about can pretend it means something else. I'm the chairman of the Y14.43 committee, the only standard in the world that deals with Dimensioning and Tolerancing Principles for Gages and Fixtures. I'm the final arbiter when it comes to gage questions. If you write a letter to ASME with a gage question, they send it to me. With all of that, if they still refuse to believe what I've said and what Y14.43 clearly states, then nothing will convince them.

James Meadows
 

Subject: Straightness

Hi, Jim.

I am in a quandary. I have and engineer who wants to apply straightness to a centerplane (2 parallel sides of a rectangular part.)

6.4.1 states straightness applies to 1- a surface and 2- an axis.

I am looking at it strictly from the standard-surface and axis.

Could I be persuaded differently?

Best regards,

Ted

Ted,

I don't know if you can be persuaded differently, Ted. But the answer to your question is that from 1966 to 2009 the Y14.5 standard allowed straightness of the (centerplane) "derived median plane" (a term introduced in the 1994 standard) to be used for widths. In the Y14.5-2009 standard, the concept was switched to flatness of the the derived median plane. Currently, per Y14.5-2009, straightness is only used for surfaces and (axes) "derived median lines".

Jim

Subject: Composite vs. Two-Single Segment

James,

There is a slight confusion going on here with composite tolerancing, single segment vs. two single segments. In the example below I'm showing a two segment composite tolerance control. Someone is saying that I'm showing the wrong control it should be a single composite. What I wanted to do is have a tighter control to datum A. Could you please help me?

Alvin

Alvin,

It would appear to me as though you are using the correct control. A composite control used instead, would only be able to tighten the orientation (angle) relationship to datum A when used in the Feature Relating Tolerance Zone Framework (FRTZF) to within a diameter of .010. Composite controls lose their ability to locate to datums in the FRTZF.
Two single segment controls, such as the one you've used, do not lose the capability to locate in the second level of control. So, the control you have used tightens the orientation and location relationship to datum plane A to within a diameter of .010 (provided the dimension given from datum plane A is a basic dimension). If that's what you wanted to do (and the dimension from datum plane A is a basic dimension), then that has been accomplished with this control.

I hope this helps.

James Meadows

Subject: Datum Targets

Hi James,

I thank you for any response you might have to this question. Either I'm misunderstanding something or this is a bad drawing.

I'm to design a fixture to hold this welded Arm so that the 5 holes can be bored.

My question is concerning datum target A1. It appears on Datum Surface A, (front view toward the right), but it is underneath an 8mm plate welded to the outside of surface A.

How can this be inspected?

Rick


 

Rick,

If it's underneath the plate that is welded on, the only way to set up on A1 absolutely correctly would be prior to welding the plate on. It looks like the person who toleranced this was trying to say that datum plane A is constructed by the targets A1 through A3. It appears in one view that A1 is shown beneath the plate, but in the other view that it is on top of the plate (since they didn't use a dashed leader line in that view). Whichever it is needs to be clarified with the designer and if it ends up that it is beneath the plate, ask him if he will move it to a place that is accessible. If not, build your fixture the best you can to simulate these targets. A3 is also a concern. It appears to be on the surface opposite A2. That would imply a step datum, which is perfectly legal, but harder to design a fixture for, since they come in from opposite directions and would fight one another, even cock the part. I hope I'm wrong about A3.

Again, the best solution to these problems is always to talk to the designer and find out if these things can be clarified and/or changed to be more manufacturing and inspection friendly. Most of the time, people tell me that they don't have that kind of relationship with the designer. I think, if that is the case, that it is just bad business. We're all part of the same team. We need to work together.

I don't know if that helps or not, but it's the only suggestion I have.

Jim

Subject: Two-Single Segment vs. Three Segment Controls (Two Position and One Parallelism)

Jim,

Thanks for your quick reply. It has helped for sure. I will ask the designer as you have suggested.
I am not used to seeing 'True Position' with two single segment controls with the bottom one NOT having the datum-letter. Is it implied? I'm also a little confused by the "three-tier" method used here. (Seems redundant).
Thanks, Rick

Rick,

The two-single segment control without a datum reference in the lower control makes sense if there is more than one hole being controlled. In these cases, the position without the datum reference is to keep the two holes in alignment with one another (coaxial). It would have been clearer if they had stated that there were two holes within each pattern, but you can see from the views that they are controlling the coaxial holes. No datum is implied.

The three tier methods used here are not redundant. The last control just appears to be trying to hold a tighter orientation (parallelism) relationship to one of the datums and has referenced that datum feature as primary in the refining control.

Jim

Subject: Tolerance Stack-up Question about Dowel Holes

Jim-

I took your tolerance stack-up class last year and have a follow-up question.

I am working on a tooling issue and am considering changing two pressed in dowel pins (.0625in.) and replacing them with screwed in dowel pins (dowel pins that are threaded at one end and replace the press fit hole with a threaded hole). The threaded hole would then location the position. I know that the stack-up will increase, but could you give me a suggestion on how to analyze the differences:

The problem with the pressed in pins is they fall out over time.

Thanks,

Carl

Carl,

As long as the tolerance zone is projected on the threaded hole, as it should have been on the dowel hole, they are treated the same. The threaded pin centers itself in the threaded hole in such a fashion that any additional clearance created by the difference in the pitch diameters of the pin's thread and the hole's thread is negligible and not quantifiable. So, just treat the portion of the pin that projects from the threaded hole as you would the portion of the dowel that projects from the press fit hole.

Jim
 

Subject: Datum Reference(s) for Parallelism Control Tolerances

Dear Jim,

I have a question regarding a requirement in the standard, if you have the time to reply:
If I have two parallel surfaces and I want to control their parallelism to each other, why should it be necessary to anoint one of them a datum for the sake of having a datum reference for the other? Why is not permissible/sufficient to point to both surfaces with a parallelism control tolerance sans datum reference?



In the example above, I would think selection of one end as a datum would only create clutter and confusion-i.e., which end (of the symmetrical physical part) is meant to represent the datum? The intent here is that both ends have the same acceptable flatness, as well as being equally parallel to each other, so there can be no right/wrong way to assemble. Albeit not an orientation tolerance, the application of profile appears to be allowed for multiple surfaces without the benefit/necessity of datum references.

(Perhaps we have a situation here that deserves consideration the next time the standard is refreshed.)

Ted

Ted,

The text and illustrations in the Y14.5 standard define parallelism of a surface as a control of (flatness and) the zero degree angle held to a datum. We have to know what to set up on, to hold parallelism to the plane or axis it is referenced to.

If it had been defined differently by the Y14.5 committee, then who knows? The first thing I learned when I started going to Y14.5 meetings is that what one person thinks is perfectly logical, others disagree with, and they disagree with the same conviction.

We make our suggestions, hear the arguments and either win or lose the vote. If you go off the grid by putting something on your company drawings that is not supported by the standards, you are on your own and have no legal leg to stand on unless you can back it up by creating your own company standard with different rules (which is almost always a really bad idea).

The truth is, your concept would have to be defined to be understood. I have to admit if I saw it on a drawing, I wouldn't have a clue as to what you were talking about, simply because it isn't defined in the standard. As currently defined and understood by all, surfaces can't be parallel only to each other. Surfaces can be parallel to planes or axes. However, if two surfaces are parallel to the same datum, then they are parallel to one another to within the sum of their tolerances to the datum they are both parallel to. This is a statement supported by good math.

The problem is that (unlike something like profile of a surface or even position, where datum references are optional) parallelism must be referenced to a datum. Why? Because the rules say so. If the rules change, then it's a whole new ballgame.

Jim

Subject: A New Twist on an Extrusion

Jim,

Have GD&T question for you! I need to know how to specify the maximum amount of twist on an aluminum extrusion? Attached is the drawing. I could care less about a little bit of twist back-n-forth up and down the profile. What I need is for the ends to be aligned when assembled. Straightness will not get me there. The only one that might is composite profile, unfortunately I do not want the thickness of the tube to vary as much as I will let it twist. This is a handle for a vacuum, and I want the handle at the top aligned with the powerhead at the bottom of the vacuum when assembled. Any ideas? Seems there needs to be a new control for extrusion twist?

Tim

Tim,

If you only need the ends controlled, only control the ends. One option is to make the dimensions for each end basic for size and shape. Use chain lines with basic dimensions that state how much of the end is being controlled (where it is [if it's not on the very end] and how long it is). Profile one end and reference no datums, but label it as a datum feature. (You said we needed a new control for extrusions. In fact, in Y14.5-2009, a new type of datum feature is called a Constant Cross Section/Linear Extruded Shape. (If you have my new Gray text book, look at page 242.) Now that a certain length of one end of the extrusion is a datum feature, profile the other end and reference the datum feature you've created on the first end. This will control shape and size on both ends and align them to one another.

If you want the size and shape controlled to within a tighter tolerance than the alignment, profile the first end (the datum feature) to within a tight tolerance, then profile the other end with composite profile tolerancing with a looser tolerance in the upper level of the control that references the datum, and then use a tighter profile tolerance in the lower level of the composite control that references no datums.

If, for some reason, you want a looser size and shape tolerance, than an alignment tolerance, profile the first end and label it as a datum feature (but reference no datum in the profile control). Then go to the other end and profile it to within the same tolerance and reference no datum. Beneath this profile control, put a positional (BOUNDARY) control with a tighter position tolerance that references the datum.

You can also augment or replace the chain line with a "BETWEEN" symbol, that is used below the profile and position controls (if the positional boundary concept is used), that states the control applies between x and y (or whatever you want to call where the control starts and where it ends). See my books on Profile controls for examples of the use of the "BETWEEN" symbol.

The middle of the part that does not constitute the important ends of the part can be controlled with just plus and minus tolerances for size and shape control, or if they are important enough, just define them with basic dimensions as well, but tolerance them with a looser profile tolerance.

The really important part of this is to make sure you define what control applies to what portion of the part. As I said, chain lines, between symbols and basic dimensions should suffice. If not, add a detailed note (local-near the geometric control, if it's short, or put it with the general notes if it's long).

I hope this helps.

Jim
 

Close Window