How to persuade engineers to abandon silly tolerances? 3

[616Nickel]

I'm the Senior Manufacturing Engineer for a mid-size electronics manufacturer, and our Mechanical Design Engineers have recently started to put GD&T tolerances in all sorts of strange places:

- 0.02" perpendicularity specs on PEM studs up to 1" long.
- 0.005"-0.015" flatness specs on machined 14" x 10.5" aluminum bezels.
- 0.010" parallelism specs between the front and rear faces of a bezel.

Now, these bezels are assembled to a sheet-metal PCB mounting plate and enclosure, and the entire thing mounts into a console. We have never had problems with poor flatness on our parts, and there is simply no reason to ask our machine shops to hold parallelism with a surface that is effectively just waving in the air. As for the PEM stud perpendicularity spec, I am told that this is to allow us to reject parts if they are damaged in shipping, and to ensure proper packing by our vendors!

Now, I used to be a Mechanical Design Engineer not so long ago, so I didn't come down in the last shower. But the trouble is that these engineers are largely insulated from the cost implications of these specs. Plus, they're quite jealous of their territory, and it's notoriously difficult to persuade them to change their drawings. So what I think I need to do is demonstrate in a pretty unambiguous way the cost impacts of adding these kind of superfluous tolerances. Are there any good sources for analyzing the cost impacts of tighter tolerances, or is it really just a matter of asking our vendors for case-by-case estimates?

 

[MikeHalloran]

You haven't demonstrated to me that the tolerances are superfluous. The overprotective engineers can probably cite an expensive instance at the root of every one.

Here's a clue.
If you have a 'Buyer' dealing with your vendors, maybe you can have a good, friendly, cordial relationship with them. Understand that the relationship is inherently adversarial, but can also be collegial and cooperative.
If you have a 'Supply Chain Manager' dealing with your vendors, then your relationship with your vendors will be evolving from adversarial to openly hostile as each of them is systematically screwed over. In this case, you will probably need even more tolerances, because your vendors will be trying to ship you garbage and your inspection will be trying to sort it out.






Mike Halloran
Pembroke Pines, FL, USA

 

[616Nickel]

"You haven't demonstrated to me that the tolerances are superfluous."

No, I haven't. Nor have our engineers attempted to justify them to me. But I was the original designer of the precursors to these enclosures in a previous stint with the company, and since they have changed remarkably little over several years I can state with confidence that there are "no expensive instances at the root of each one." For instance, what justifies a 0.010" parallelism spec on a display bezel? Why is it so important that the cosmetic, non-functional front face of a console-mounted display is everywhere within 0.010" of the rear? Heat dissipation ? As I wrote above, the only justification for the perpendicularity spec on the PEM studs is that they might get bent in shipping. But this is using a sledgehammer to swat a fly--incurring an additional fabrication and inspection expense on each unit in order to deal with a hypothetical problem; that is, a vendor refusing to take responsibility for damage resulting from poor packaging.

We have a Buyer dealing with our vendors, yes, but I am the main technical point of contact, and our relationships are generally friendly, especially with those vendors I have known for several years. I could certainly ask them for estimates, but I'd like to save them the hassle if there is a good reference available.

 

[ctopher]

Have the engineers give you a tolerance stackup to prove their tolerances work.
If they can't, they are probably only trying to show their ego's.
Usually when I see ".010" and ".005" typical all over the designs/drawings, it's a red flag of ignorance.

Chris
SolidWorks 11
ctopher's home
SolidWorks Legion

 

[616Nickel]

"Have the engineers give you a tolerance stackup to prove their tolerances work."

Nope. These particular GD&T tolerances don't stack: Flatness specs on machined parts that mate with sheet-metal parts that do not have matching flatness specs, parallelism specs on surfaces that don't mate to anything besides air, and perpendicularity specs on mild-steel PEM studs (just increase the hole size!).

I'm on the review list for the drawings, and I could just be a jerk and reject them until they either stop doing it or give me an ironclad justification. Maybe it'll come to that, but for now I'm trying for a gentler and more persuasive approach.

 

[MintJulep]

Just look at all the silly questions about "default tolerances", "title block tolerances", "standard tolerances" and "Standard tolerances" here on Eng-Tips.

Clearly demonstrates that there is widespread lack of understanding of the purpose of tolerances within the broad engineering and design community.

 

[Jboggs]

Ah, you have hit on one of my pet peeves. I find that many young engineers, fresh from school where they were continually told how smart they are, have never really had to justify why they make certain decisions, and are often completely unable to do so, other than "I just wanted to do it that way." And they just cannot resist the opportunity to splash a lot of technical looking symbols all over a drawing to convey to all that see just how "smart" they really are. They usually don't fully understand the symbology and the gemoetry behind it themselves, but it sure looks impressive. Its sets them above the common person who is confused by this magical language. So it makes it easy to say "I just wanted to do it that way." I've been in this field for a LONG time and all I see is an ego that needs "adjustment".

Can you institute some sort of "work sharing" scenario where some of the engineers that cause the problem have personal face-to-face encounters with the costly results? Can they be involved in the interface with vendors? Can someone show them that "additional specifications = additional costs"?

Another idea I have used to great benefit in the past is this: identify the offender that is the most respected by his peers. Take him/her aside and ask for his/her help. Tell them that you have seen a great benefit from the use of GD&T, and give concrete examples that you really believe in. Tell him that you also are afraid that its value is being harmed and that's where you need his help. Compliment him. Tell him you are approaching him about this because you have seen how the others respect him and will follow his lead. You might tell him you repsect his professionalism and judgment.

Explain how overuse of any tool wears it out. It becomes useless and you don't want that to happen to GD&T because it is too important. Explain how when the symbols appear all over a drawing its hard for the regular guy on the shop floor to know which ones are really important. The end result is either a part that does not meet the need, or one that costs much more than it should because of all the unnecessary inspection time. That means that not all GD&T symbols were created equal. Ask him if would agree that there are examples of GD&T being overused in some cases, and that overuse in effect devalues it. Ask him what his suggestions are to help make sure that the GD&T axe remains sharp. Ask him if he could be your go-to guy to help the guys in the shop using the drawings and demonstrate to them how to inspect the parts for all the GD&T symbols. That will bring him face-to-face with the issue. Ask him if he thinks he could help keep the GD&T tool sharp for the benefit of the company.

His subtle change in priorities can spread through the whole group. I've seen this work more than once.

 

[Belanger]

One other thought -- when implementing GD&T, that means that datums will be assigned at various places on a part. And when datums are placed on the print, it's good design practice (and this protects you as the mfg engr) to use GD&T to relate the datum features to one another, even if they seem irrelevant at first. This could be another reason to maintain things that seem unnecessary to you.
For instance, I've seen perpendicularity specs on holes through very thin parts. Most of us would say, "Who the heck cares about perpendicularity in a 2 mm metal plate; it is negligible!" But suppose that hole is a datum feature -- we want to relate it back to the primary datum, which is most likely the face of the metal plate. This also ensures that any gaging or fixtures will have everything tied together.

I don't know if this is justification for the items given in the OP, but it's something to keep in mind.

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

http://www.gdtseminars.com

 

[Madhu454]

Hi Belanger,

Even I faced the same problem, we recently introduced GD&T in our organization. When we need to relate the datum features in the part, mainly while qualifying the datum features usually with form controls. There were many questions raised form sourcing about the introduction of new geometrical controls on the part. Even suppliers quoted more money for the same parts.

Assume a case of a rectangular block with only size dimensions. The same part is now with some datum features, flatness tolerance and some perpendicularity tolerance to relate the datum features etc., the supplier says that now he wants to measure all these new geometrical symbols which are put on the part which causes extra money for it. By introduction of GD&T don’t you think measurement cost will increase to some extent?.

Also while qualifying the datum features, to what tolerance it should be? One of the GD&T teacher suggested that the datum features to be qualified with the flatness or appropriate geometric characteristics to 10% of the desired tolerance of the features we are controlling? Is that true?

 

[CheckerHater]

How to persuade engineers to abandon silly tolerances?

You do realize you are asking people who apply tolerances for a living?

 

[TheTick]

I am an engineer who does his own drawings and also used to be a machinist. I consider all tolerances carefully.

Engineers need to understand that every tolerance on a print adds cost to a part. That cost is exponentially proportional to the tightness of the tolerance.

Also, there needs t be understanding of the tolerances that are possible with a given process. To simply mandate a tighter tolerance and demand that manufacturing "be more careful" is not enough. Also, manufacturing needs to provide these process tolerances.

Like most people, engineers are pressed for time and will copy specs without really thinking about them. Sometimes (like w/ PEM studs), these tolerances may come from the manufacturer of the purchased part.

In any case, the beginning of the conversation should be an objective examination of the need for the tolerance and how it got there.

 

[616Nickel]

"You do realize you are asking people who apply tolerances for a living?"

Haha, yes. I was one of those people until a relatively short time ago, and I may well be one of them again one day. I understand the point of tolerances and wholly support their prudent use, but friends don't let friends apply pointless tolerances that only serve to drive up cost without adding value to the product.

Oh, an anecdote that halfway answers my own question. When we got quotes for a prototype run on one of these parts, the price came in shockingly high. When I asked our vendor what was driving the cost up, he cited these GD&T tolerances. I requested authorization from the project lead to waive those tolerances for the prototypes, and the price dropped by 45%. Of course, the parts were perfectly fine; as I mentioned, we've been making variants of this enclosure since 2004 without these tolerances. OK, point proved, I thought, only to see the exact same GD&T tolerances turn up on the next drawing revision.

Hence my inquiry about the availability of estimating tools, so I can make the point in a more general fashion. Our vendor has also suggested we switch from 6061 to MIC 6 for these parts specifically to cope with these GD&T specs (since it has very good flatness and parallelism), but I am reluctant to incur the increased material costs just so we can continue to use pointless tolerances.

 

[616Nickel]

Here's another related question: Our drawings specify a blanket ±0.5° angle tolerance in the title block, and my understanding is that it implicitly applies to all perpendicular surfaces on the part that are not explicitly dimensioned. In your opinions, would the same blanket tolerance also apply to perpendicular pressed-in components like self-clinching studs and standoffs?

PEM, for their part, claims a 0.5° standard perpendicularity tolerance on their studs and standoffs, and a 2° worst-case tolerance.

 

[Belanger]

When I asked our vendor what was driving the cost up, he cited these GD&T tolerances.

Nope. That's a myth. Your parts might seem less expensive when stripped of the GD&T, but here's the real question: If you take off the GD&T stuff (position, profile, perpendicularity), guess what -- those qualities don't cease to exist! They will just have to be stated in a different way. And I'll bet that when the GD&T is stripped off the drawing, your designers don't bother to address those qualities any other way, except perhaps through the title block tolerances which is fraught with dangers except in simple cases of length/width/height.

So the bottom line is that GD&T is a language allowing you to control certain characteristics based on the function of the part. If you think the GD&T is making the cost go up, your design engineers need to evaluate the GD&T numbers, not chuck everything out the window.

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

http://www.gdtseminars.com

 

[616Nickel]

"If you take off the GD&T stuff (position, profile, perpendicularity), guess what -- those qualities don't cease to exist! They will just have to be stated in a different way. And I'll bet that when the GD&T is stripped off the drawing, your designers don't bother to address those qualities any other way, except perhaps through the title block tolerances which is fraught with dangers except in simple cases of length/width/height."

Right, of course--as you may gather, I understand that objects retain qualities of flatness, perpendicularity and parallelism, to a greater or lesser degree, even when we don't explicitly specify them using GD&T callouts. And, as I've already stated, I think GD&T is a wonderful tool when used appropriately--I have used it many times myself as a designer.

But these are parts that were manufactured for years without these GD&T specs, sourced from around seven different vendors over that time, with the price remaining roughly consistent, all else equal. I assembled some of the early units myself when I was the designer, and they did not present any issues of this nature ... nor did any of the subsequent units, to the best of my knowledge. After all, these are electronics enclosures made mostly of sheet metal, made to fit into cutouts in sheet-metal consoles or into 19" electronic racks--they do not generally have stringent fit requirements or tight assembly tolerances. Then all of a sudden, the mechanical designers started adding GD&T tolerances to the machined parts (already our most expensive metalwork by far), and behold! The price went up! I just drew the natural conclusion.

 

[Belanger]

That's true -- but the reason is that the GD&T drew their attention to stuff that hadn't been there previously. So they think it's more expensive because there's more callouts to inspect. IOW, if those callouts weren't given on the print, they weren't checking things that were really meant to be checked. (The only exception I suppose would be form controls on a feature of size, because "Rule #1" already controls form, so a GD&T callout there could be seen as more expensive.)

My hesitation: If a drawing gets hauled into a courtroom, we have to be sure that all the possible characteristics are adequately addressed.

Here's an example -- I was helping design a flat pipe flange once. The outline (perimeter) of the flange was a triangular shape, but the shape wasn't critical because nothing touched that perimeter; it just saw air. When I tried to put a GD&T profile tolerance all around that perimeter, there was great resistance from the others: "That's not a critical feature." "Now I have something else to inspect." Etc.

But I pointed out that if we didn't use the profile tolerance, then the triangular shape still had to be toleranced somehow, and the plus/minus method would lack many of the advantages of GD&T. So I showed them how a profile tolerance could still be used, but the tolerance number we agreed upon was something like 3 mm! Not expensive at all.

In summary, you might have a point about things like flatness and parallelism, because those characteristics were already implied via the size (unless it's a non-rigid part). But I'm just cautioning against thinking that GD&T is only to be used on critical features.

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

http://www.gdtseminars.com

 

[looslib]

While the price myth may be that, it does exist. We sent basically the same parts with +/-, ASME Y14.5 and ISO standards called out for tolerancing to a mixture of our normal vendors and got back different price quotes depending on the tolerancing method on the print. All of the tolerance methods would make the part acceptable, but the shops perceived a higher cost when we specified ISO tolerancing over ASME which was higher than the +/- method.

At another company, we took a standard drawing and sent it out for a quote. We then took that same drawing and retoleranced it to reflect the machining processes used. The revised quote was 30% less for basically the same part. These were molds for glass ware, so we had also looked at the final product cesign and its tolerancing. When we went to the mold designers to ask them to loosen the tolerances, they bulked and would only loosen about 20% of those that we asked them about changing.



"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[powerhound]

It's not the GD&T that is raising the part cost, it's that part requirements that do it. When you dimension a pattern of holes at 1.0" from the edge of a rectangular plate, all the way around, with no other controls than a general tolerance block, there's nothing that says you can't use caliper to pinch the material between the edge of the part and the edge of the hole and just add the radius. Now if your plate is out of square, your hole pattern will be too. Now you have a nice cheap part that doesn't work. Add GD&T to specifically define your requirement and the price goes up because you have told the manufacturer that what you buy actually has to work. That price increase is not because of GD&T, it's because you no longer have an ambiguous requirement that could force you to buy parts that you can't use.

Powerhound, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[616Nickel]

All excellent points and I agree 100% but believe me, that's not what's going on here. Our enclosures are pretty low-precision (and frankly, pretty crude) pieces of equipment. Does the PCB fit? Are we chafing any cables? Are we able to dissipate heat? Do we comply with our client's EMI/EMC requirements? Do holes line up with obrounds? And so on.

No, I think this is just the common engineering practice of adding specs because it doesn't cost the engineer anything to do so, plus it adds a little of that "real engineering" flavour to a job that is, in reality, pretty unchallenging--designing one sheet metal box after another. I did that job for 31 months and found it grindingly tedious after a short while, jumped at the first proper mechanical design job that came my way, then came back to the company four years later as a Manufacturing Engineer.

 

[MikeHalloran]

So, if it's _just_ a tin shop, one has to wonder what GD&T brings to the party, and why the engineers started using GD&T.

You won't likely change the engineers' minds; you have to find and re-educate whichever Heavy Hitter got talked into it.



Mike Halloran
Pembroke Pines, FL, USA