Readers - As many of you know, I have no real schedule as to when a new newsletter comes out. It has a lot to do with a mix of my time in the office and the batch of letters written by you with clear, important questions others might be interested in reading. Here is the latest installment of the continuing saga of The Tolerancing Engineer (serving the GD&T community since 1982).

~ James D. Meadows

Subject: Tapers and Two Single Segment Profile Controls
 

Jim,

Thanks for taking my call this morning. I'd like to continue our discussion on tapers by sending you these hand-drawn pictures. You'll notice the top half image represents the current dimensioning of the taper. The bottom half represents the proposed changes based on Fig 8-26 in your GDT 2007 book. Can you please review the validity of these changes? I have a few more questions also.
1. How would I dimension the depth of this taper? Plus/minus or basic?
2. Would the perpendicularity to A allow the taper to lean? Wouldn't datum A move with the lean? I'm having a hard time visualizing this. Any suggestions?

Brian





Brian,

The proposed change looks good. The depth of the taper can be dimensioned with plus and minus tolerancing. For this dimension, I would recommend an arrowhead at the bottom of the taper and a dimension origin symbol (a circle) at the top.

The perpendicularity control to A provided in the profile control would allow the entire cone to lean .002. Datum A would not move with the lean, since the profile control is measured from datum A (not vice-versa).

I don't have any suggestions, but I have questions. Will datum feature B still be necessary on the drawing? If nothing is referenced to B, it should be eliminated. Shouldn't the cone be located to the rest of the part in some way? That would mean that although datum B wasn't the proper location datum for the cone, the cone still needs to be located somewhere on the part to other datums. One option is to use more datums than just A in the upper profile control (that uses .002). Another option is to lead in with a third control that uses an even greater tolerance to A and other datums, then refine with the two controls you currently propose.
Another option is to make the cone a datum feature and locate (with position or profile) other features on the part to datum A and the cone.

Just one more thought, Brian. Consider giving datum feature A a 3-D form control of its own. If it is a planar surface, a flatness control would be appropriate. If features are to be measured from it, it should be well formed enough to make those measurements repeatable.

Hope this helps.

Jim



Subject: GD&T in Kuala Lumpur

Hi Jim,

How would you like to go to Kuala Lumpur? Do you think these topics are broad enough for 2-day workshops?
1st workshop: Intermediate level (Fundamental lessons in proper interpretation of engineering drawings used in the design, manufacture and inspection of parts; Measurement Applications; Also a discussion on differences between the ASME standards and the ISO ones, and trends towards simplification.)
2nd workshop: Advanced level (Advanced application and analysis; Advanced Optimization strategies; Advanced Tolerancing; Implications of GD&T on product reliability, cost savings and six sigma.)
Gene

Hi, Gene. I do think these topics are broad enough to sustain a 2-day course. And, no, I don't want to go to Kuala Lumpur. I'm certain it's a very nice third world nation, but I have enough trouble finding my way around Milwaukee. I don't want my body found covered in Ox manure in a ditch in a country I couldn't locate on a map with a flashlight.

If you ever need me to do a job in Disney World, let me know right away. I'm here for you. I'm just not there for you.

Jim

 

Subject: Just Grinding My Profits Away

I was consulting for a company, teaching them how to tolerance parts and looking at their existing design drawings. At one point, I was given a tour of the shop by the manufacturing manager. As we walked through the facility, I could see assemblers putting piece parts into the machines that make the packaging equipment. I noticed that many of the parts simply didn’t fit. When that happened, which was most of the time, I saw them making notes of the problems and then taking the parts over to an area where a huge group of employees were using grinding machines. These machines were throwing up such a stream of sparks that it looked like a fourth of July celebration.

The manager spoke on at length, obviously proud of the company. He said, “We make products that are unique. We don’t make hundreds of thousands of parts that have to interchangeably fit into any machine we make. The parts just have to fit the one machine they will be used in. Therefore, it not only seems to me that we don’t need to know what you are trying to teach us about tolerancing parts, but we don’t need tolerances at all. When we get a part in from one of our suppliers, we just grind it until it fits into the assembly. What benefit is there to us to learn to design and tolerance parts in what you call the “correct way”?
I asked him, “Just how much time is spent grinding things to fit?”.

He answered by saying, “A lot. That’s why we employ all those people over there. They grind and grind, until the parts fit like gloves. The parts become dedicated to that one assembly and wouldn’t fit another. It’s a great system. The machines work wonderfully.”

I asked, “Wouldn’t it be more efficient if you were able to just take the parts as they come into your facility and assemble them without all of that grinding?”
He looked blank, and responded, “Why would I want that?”

I plodded on, saying, “You could save a lot of money and you could get rid of all those people that do the grinding.”

He face changed to a horrified expression and his voice went up an octave. “I like those people. Why would I want to put them out of a job?”

“Well,” I said, “You could always reassign them to do something else, I guess.”
“Like what?” he responded, his voice even higher. “There wouldn’t be anything else for them to do.”
I nodded, showing concern for the workers. Then I asked, “What if one of your machines breaks down when your customers are using them? How can you ship them a replacement part?”

His voice went back down into its normal range, and he smiled at me like he was about to explain something to a small child. “We just bring the broken part back to this facility and give it to the grinders. They measure it, and grind another one that size. Then we take it back and put it into the machine.”
I said, “Wouldn’t it be faster, if you could deliver to them a replacement part that just fits, you know, without all that grinding?”

He was dismissive about the idea. “I guess that might be better in some cases. Let’s move on and I’ll show you the scrap area.”

He did, and let me tell you, it was impressive. It reminded me of the time I was standing next to a general foreman on the shop floor of another company, when one of his employees came running up and said, “Boss, the scrap area is full!” And the boss responded calmly, “Just build the walls higher.” I remember thinking at the time, “This is a management decision? Wow, I could do that job.” It gave me managerial aspirations.

It seems to me that although having parts fit interchangeably offers many benefits to those making thousands of the same thing, it would beat the heck out of all those “grind to fit” or “just throw it in the scrap area” situations people are so satisfied with (or maybe just so used to) even if they do make only one of each.

Recently, I was called by the design manager at that company and invited back to take another stab at teaching them how to dimension and tolerance parts that just fit together the first time out of the box. He said he had to get permission from the manufacturing manager first, since the money to pay me would come out of his budget. When he called back, he said we would have to cancel the trip. The manufacturing manager had told him that he just didn’t want to spend his budget money that way. I guess he needed to hire some more folks to work in the grinding area.

...James D. Meadows



Share this newsletter with friends or colleagues...



Subject: As cast bolt holes on W5A580 gear box case part# 449AA & 103AA

James,

I have been having trouble holding true position on two as casting assembly bolt holes ( thru holes) . I don't know why the general rule for MMC can't be applied to these two as cast bolt holes? See the areas painted blue below. The core holes have a dia 10.23 max, with a min of 9.93 with the following:

The mating bell has these two holes drill and tap all the way thru the bell. See the bell machining print below. Steve, our resident engineer said that I don't have any bonus tolerance on these two as cast thru bolt holes. I want your opinion on this.

Thank you,

David




Bell Machines Drill and taps two holes

 

David,

Listen to Steve. There is no MMC symbology given in the control on these holes. Without these feature positional tolerances being referenced at MMC, bonus tolerance does not apply.

If the drawings had been created per the 1966 or 1973 versions of the Y14.5 standard on Dimensioning and Tolerancing, then for positional tolerances (only) the MMC symbol would be implied and bonus tolerancing could be considered.

But in 1982's version of Y14.5, position without MMC, LMC or RFS references is incomplete. One of them must be referenced.

However, these drawings appear to use 1994's revision of the Y14.5 standard wherein position is treated like all other geometric characteristic symbology. It is implied Regardless of Feature Size (RFS), unless the MMC or LMC symbology is used after the geometric tolerance or after datum features of size. RFS controls allow no bonus tolerance.

The giveaway is that the international datum feature symbol is used on these drawings. That symbol wasn't adopted by the ANSI approved ASME Y14.5 standard until the 1994 revision was issued. In this revision, RFS is implied for all geometric characteristic symbols, including position. It is called Rule #2.

Jim Meadows



James D. Meadows & Associates, Inc., will provide your facility print review services and/or on-site training for courses 2½ to 5 days in length---look at our generic course outlines for Basic GD&T, Advanced GD&T, Tolerance Stack-Up Analysis, or Gaging (any of which can be tailored to your design, manufacturing and inspection requirements, if requested, and to your budget and time constraints) at www.geotolmeadows.com



Subject: Co-Datums

Jim,
I was looking through your geometric dimensioning and tolerancing book trying to search for an example that references two different size features that are used as co-datum. I was wondering if my application of GD&T (on the attached drawing) is correct. We are trying to establish a co-datum between the two holes that are placed at a distance of 8.478 inches away. I have come across examples where features that have share a co-axis but none that are not aligned with each other. Could you please provide some information about the how my dimensioning scheme would be interpreted and where I can located the text and example in ASME 14.5 that explain my situation.

Thank you for your assistance,

Otha



Otha,

It is acceptable and common to make a compound datum pattern out of two holes. First the holes must be positioned to each other. This can be done by making one of the holes a datum feature and giving it a perpendicularity control to the appropriate datum (such as H). Then make this hole a datum feature (such as G). Then position the other hole to the same primary datum (H) and to the first hole (G). Then make this hole a datum feature (F).

Subsequent features on the part may then be positioned or profiled back to H, G-F. It would be advisable to reference G and F at MMC in these controls. We could then picture what we are measuring from with a fixture that consists of a plate to simulate H and two pins at virtual condition to represent G (MMC)-F(MMC). Once the part is mounted on this fixture it is immobilized, making additional datum features unnecessary. To do this, a basic dimension between G and F is required. The datum axis formed by these two holes would be perpendicular to H and halfway between the two gage pins.

If other surfaces are located from these datums, they would have to use profile of a surface and use basic dimensions that would be toleranced by the profile control. The profile tolerance can be large or small as needed.

If widths are located from these datums, we would need to show their relationship to the datum axis. If they are offset from the datum axis that is formed by the common plane that runs between G and F and the plane that is formed halfway between the two virtual condition boundaries (embodied by the gage pins), the offset must be given as a basic dimension. If they widths are centered to the datum axis, then no basic dimensions are needed.

There are other methods open to you, but I can't easily explain them without going on for pages. If you want to talk more about this, please call me on Friday. It is my only day in the office. (615)824-8644.
Jim
 

Subject: Question on Wall Thickness
 

Hi James

I attended your course Dimensioning and Tolerancing Principles for Gages and Fixtures.
I have a question with a wall thickness calculation you apply on your book Geometric and Dimensioning Tolerancing.
This is the image



What I’m looking for is the minimum and maximum wall thickness from the right side of the part in the front view and the 4.0-4.3 hole closest to it.

So in your book you use for your calculations only the pattern shift;

4.3 = LMC Hole
+ 0.3 = Geo. Tol. At LMC
4.6 = Outer Boundary
+ 0.4 = Pattern Shift (B at MMC)
5.0 = Outer Boundary with Pattern Shift

The question is why you didn’t apply also the Position Tolerance of B (which will increase another 0.4)?

Thanks and Regards,
Luis

 
Luis,

The edge of the part is measured from B. B is not measured from the edge. If B is out of perpendicularity to datum A, then the wall could thin, but any out-of perpendicularity of the B hole would negate an equal amount of pattern shift experienced by the profile control. So, you could ignore the MMC symbol after the B in the profile control and assign all of the impact on B being out of perpendicularity to A, or you could assign all of the tolerance given by B to the pattern shift for the profile control. But it is one or the other, not both. Either way, the wall thickness is as given in the book.

Jim

The full calculation and subsequent tolerance stack-up follows:







Subject: RE: Follow-Up Question
From: Luis

Hi Jim,

 
And what difference does the separated requirement label? Are all of these changes are due to it?

Luis,

If there was no separate requirement note, there would be no pattern shift between the 21 hole pattern and the profile, because if either shifted both would have to shift together and their relationship to one another would not alter. However, with the separate requirement note, the pattern shift of the 21 holes would be allowed to shift in one direction and the profile would be allowed to shift in an opposite direction. So, the pattern shift would be allowed for the 21 holes and it would also be allowed for the profile control. The 0.4 shift tolerance would apply to both (as shown in the calculation in the above).

This question is different than your original question.

The question you've asked in this email gets asked a lot and is addressed in many of the newsletters on my website. Please take a look if you need further clarification.

Jim

Close Window