Not using GD&T? 10

[ctopher]

My last company, and current, have argued with me about using GD&T.
Engineers, purchasing, and managers have argued that it makes the parts more $$.
I tell them no it doesn't.
Last company had parts machined in China, current in Thailand and in USA.
China and Thailand have said not to use it because they don't understand it.
Often I see parts made that don't meet print with GD&T, but are bought off anyway.
I'm at wits end, tired of arguing with everyone.
I also find more people here in USA that don't understand it.
Anyone here run into this? If so, what do you do?

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[ctopher]

Burunduk,
That's in a perfect world. Doesn't always work like that.
That last time I saw any type of in-house training was when I worked at Boeing years ago.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[3DDave]

Here's the sort of thing I used to work on; not this particular system or company, but on electro-optical units and missile aiming platforms for ground vehicles.

https://www.thedrive.com/the-war-zo...the-u-s-is-sending-to-ukraine-can-actually-do

Also did airborne radar systems for missions that aren't declassified and never divulged to us and a bunch of vehicles and vehicle mods as well as a stint in carrier aircraft test systems and one hand-carried test box.

The last electro-optical package I worked on was for the Australians. I did the IR and laser designator module.

Then off to a contract to fix an Israeli home-turf design for a trailer; it needed a lot of work as they hadn't put all the required parts on the drawing and were half-assing it to get it to pass the IDF requirements while telling the IDF it was ready for production. The IDF didn't care too much because it's all US foreign aid money so for them it's basically free money to spend to keep the Israeli defense contractors getting money for screwing up. The IDF colonel cared because half-assed or not his task was getting the trailers produced. I felt for that guy. The people in Haifa were nice enough, but they were terrible at design and documentation. IDF is still ordering them, a decade later, after I repaired their drawing package and they should have done the hydraulic and electrical mods to fix self-inflicted problems, but who knows.

Over the years products were for TACOM, USAF, USN, Australian Army, Kuwait, USMC, US Army, Boeing, and the USPS. IDF was via TACOM who managed FMS (foreign military sales) contracts on behalf of the IDF, but we did a number of direct to TACOM contracts that I worked on. One I wanted to work on was E-FOGM launcher systems, but it got killed by the DoD before it got out of the early development stage.

 

[Burunduk]

3DDave,
In order to "put into words" a relationship between two features or between a controlled feature and two or three reference features, the designer needs first to understand how this relationship is established as a requirement and how the variation is controlled. If it is expected from the designer to describe the conventional method involving a tolerance zone, a datum reference frame, and basic dimensions, the company should at the very least allow the designer to access either the Y14.5/ISO GPS standards, an internal standard based on any of these, alternatively some professional literature, or arrange some kind of training. If the designer never learned it because the company's management is "smart" and so good at cutting expenses, it is expected from the designer to come up with a method of his own.
Then:
Firstly, I'm all for encouraging creativity and originality, but - good luck No. 1.
Secondly, If he actually managed to come up with something that makes sense and is truly effective at controlling the variation AND providing enough allowance, will the person reading the note be able to understand it as intended? Good luck No. 2.

Suppose that they managed with the above, by extensive direct communication between the designer and the in-house manufacturing+inspection or external vendor. Now the company is going to "rapidly collect" these notes, as you suggest, to create a note bank to be used as a sort of company standard. Wouldn't it soon be required to train people to understand and implement these notes?
What else will come along in the documentation that has this list of notes? Maybe a "terms and definitions" section? Examples and illustrations "means this" style? Who will be editing and maintaining this document? Deciding whether new proposed notes are worth adding, or maybe there are already equivalent ones in the list? Sorting them by categories? What will be the cost of all this to the company?
Wouldn't it already make more sense to train the employees according to a national or international standard and spare all these note writing, collecting, and clarifying efforts?

"Suppliers who have expressed a desire to increase prices or refuse to work if they see even one geometric tolerance" are expecting to see only old school directly toleranced dimensions on the drawings they work by. That includes holes located by directly toleranced dimensions with no detailing of what the tolerance applies to. That supplier doesn't give a damn if you mean to apply the tolerance to a straight axis derived from the largest inscribed perfect cylinder ASME style, an imperfect line representing a center which is influenced by the form of the hole ISO GPS style, a center point of a projected circular contour optical comparator style, the center point of the near/far end of the hole, a center point in a circular cross-section in the middle of the feature that can be probed by a CMM, or anything else. That supplier just wants to work by an ambiguous drawing that doesn't bind him to any verification method or requires him to hire people capable of interpretation of any robustly defined requirements. How will this supplier react to a drawing cluttered with lengthy notes, possibly supplemented by portions from that company standard note-bank documentation (intended to add clarity to some of the notes)? Do you expect him to say, "yeah, that's nothing special... I shouldn't charge a lot for this part since the print doesn't have any GD&T in it..."? You got to be kidding me.

Maybe a better approach would be that if a supplier "expressed a desire to increase prices or refuse to work if they see even one geometric tolerance", that supplier should be avoided?

 

[Burunduk]

3DDave, regarding your latest post:

"they hadn't put all the required parts on the drawing..."

"IDF is still ordering them, a decade later, after I repaired their drawing package and they should have done the hydraulic and electrical mods to fix self-inflicted problems..."

So was the problem that they made an incomplete assembly drawing?
How well were their component drawings made? Was the dimensioning and tolerancing on the component drawings adequate?
Do you feel they could do much better had they been more competent at tolerancing practices such as defining functional DRFs to relate features to each other, specifying geometrical tolerance controls that minimize tolerance stacks etc.? Was the problem with their design or was it related to knowing how to communicate the design intent on the prints?

What's the moral of the story?

 

[3DDave]

The moral is - yet another organization that doesn't fit your explanation of everyone you have met, and in your own back yard.

And that's the last question I answer until you fix the backlog of questions I've had of you.

 

[Burunduk]

Nice try at evasion.
You asked many questions and those of them that make sense were answered either directly or you could easily read them between the lines. For example, your question about ISO GPS - you could conclude what my position is from my recent post from 06:26 where I mentioned it more than once (and I'm expecting you to deal with points I raised in that post). You could also understand it before you even asked the question, from posts such as the one at 30 Aug 22 11:49 - the clarification of what is referred to as "GD&T", in the parentheses at the beginning of the post.

What are the questions you think I didn't answer?

Your story is another reinforcement to my point. Most likely another company where the people who are in charge of how product definition is made think that standards are just for the formality of how things should look on a drawing and engineers don't need them to understand how to assure functionality. Therefore, the failure you told about. It doesn't mean that the engineers don't want to learn, they do. But they only learn what they are taught or told to learn, and in the long run they forget what they don't regularly put to practice. Only the right management strategy can change that situation. Do you truly think notes could solve their problems? I know my "backyard" and one thing is for sure - last thing they'll be doing is cluttering the drawings with literal notes. If they do so, few users will pay attention to those notes, and many won't even read them. People go straight for the nitty-gritty, and they look for it in the orthographic views, not in the notes section. Perhaps to a lesser extent, the same applies to your backyard too:

From my experience, people have become lazier. Add a bunch of notes on a dwg, they don't read them. It's like sending someone an email, they only read the first sentence. More people have shorter attention span.

But motivate them, and many of them will learn work methods that once assimilated make the completion of work assignments faster, easier, and with better results.

 

[ctopher]

Burunduk,
Only curious, are you in management?

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[Burunduk]

cthoper,
I'm a senior engineer. I see how problems are being solved through implementation of the principles that are taught in GD&T training. I also do some in-house mentoring and consultation. I witness how all this contributes to improvement.

 

[3DDave]

I know about 50 guys with that exact same job title, including myself. Consultation "in-house" is what the typical engineering job is. That's supporting purchasing, QA, QC, manufacturing engineering, the workers on the manufacturing floor, the documentation group, customers, suppliers, and so on. "Improvement" is a great word. Doesn't mean anything on a resume.

 

[Burunduk]

"Improvement" may be a great word that doesn't mean anything on a resume, but it can be witnessed by observing small advancements such as reduction of inspection time and cost for a product without compromising functionality, as can be concluded from customer satisfaction and continuous demand for the item.

 

[drawoh]

Since the previous 10,000 years of civilization managed without it...

During WWII, Great Britain expected its factories to all be bombed to crap. As a result, all sorts of Canadian and American manufacturing firms had to make sense of British drawings and specifications. The Japanese, Italians and Swedes manufactured various Daimler Benz engines under license. I was under the impression that GD&T came out of that in the late forties.

Back in the day, parts were not interchangeable. People made parts, put them together, and they used hammers, files, and rework to make everything assemble and function.

I have a SPI standards book here on plastic injection moulding. They claim they can mould ABS plastic to [±]0.15mm at 100mm length. This is their "fine" tolerance. Their "commercial" tolerance is more like [±]0.27mm.

I look at a couple of cellphones here, and there is no way things are that sloppy. Somebody was handed parts and told to make their process fit something perfectly. I have just written a DFMA manual in which I recommend not designing things with curvy edges and surfaces that must line up. My rule, however correct it is, is violated with gay abandon. It can be done. Over very large production numbers, a lot of NRE can be amortized. Can these parts be second sourced?

--
JHG

 

[Burunduk]

I have a SPI standards book here on plastic injection moulding. They claim they can mould ABS plastic to ±0.15mm at 100mm length. This is their "fine" tolerance. Their "commercial" tolerance is more like ±0.27mm.

I look at a couple of cellphones here, and there is no way things are that sloppy. Somebody was handed parts and told to make their process fit something perfectly

+/-0.15 mm is for 100 mm dimension. What are the dimensions of the cellphones components you are looking at?
Also, the assembly process of plastic parts takes a lot of advantage of flexibility.

There is way too much talk about how processes nowadays are very accurate, and how this supposedly means that the tolerances on the print don't matter that much. If that was the case, no one would experience excessive scrap problems, and design engineers wouldn't get requests to loosen tolerances.

 

[ctopher]

Burunduk,
It's not that there is "too much talk", that's real world. I can't tell how many times I have heard from management and senior engineers "It's in CAD", it's "made on CNC", etc. So, they think loosening the tolerance will lower the cost because it doesn't matter.
You may not see it at your company, but past few companies I have been at, I see it. The are large aerospace companies.
The drive is to cut cost, this is their way to do it. Then, later they wonder why it's not working and why parts are scrapped.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[3DDave]

Beginners analyze the parts.
Intermediates analyze the assemblies.
Journeymen analyze the inspection.
Masters analyze the manufacturing process.
Top masters look very carefully at the requirements.

If they are looking only at tolerances, they are just guessing with regards to cost control. Parts will either cost more than they should or leave open variation that could have made other parts less expensive. What's really bad is when tolerances are increased to accept greater variation and then the assemblers get stuck spending more time aligning parts and users see greater variability in performance.

Excessive scrap problems? It sounds like a confession that someone isn't paying attention to the process or the requirements. If they were they could say why the parts were being scrapped rather than just hand waving in that general direction. The truth is that a bad process isn't made better with a better part description, so they can still generate scrap in that case.

Those experiencing it might eventually be desperate, but they won't change until they are.

I'm sure there are plenty of case studies that have been done on the subject. Not seen even one. Anyone using them please link to the ones that you have used.

 

[CWB1]

Dave,

IME not using GD&T is extremely rare today. I have been employed by mega-corp OEs and a mid-sized consultancy working for small companies and even startups. If its powered by a recip engine, hybrid setup, or battery I have probably designed, sourced, and/or owned the manufacturing and quality otherwise - light-HD vehicles, trains, ag/construction, military/offroad, aircraft, oil&gas, power gen, etc. Million+ volume production programs and many more one-offs. I have no issues buying COTS parts manufactured without GD&T, but I'd RUN from any supplier that wanted our prints reworked bc experience and supplier audits have shown 1. there's no good reason for it and 2. those suppliers are usually shitshows otherwise. Ultimately GD&T is a language like any other - folks either understand it or they dont. The communication doesn't drive the cost, the work does. A lousy engineer could approve an unnecessarily vague/complex/ridiculous print using GD&T or any other language. I'd guesstimate <1% of the parts I've released ever required a CMM or uncommon inspection equipment, tho I have done many complex castings that did.

Granted, I have known many who started on the shop floor with no knowledge of GD&T, but I dont know many fabbies/machinists/welders/etc that didnt learn it well within a few years either via OJT or formal training. Among the technical office staff its just a given, not sure how any team member could suggest a prototype improvement/optimization without it regardless if they're a design engineer, CAE analyst, or lab tech.

I'm sure there are plenty of case studies that have been done on the subject. Not seen even one.

I have not seen any public studies (nor looked), but can attest that the large OE supplier quality systems will ding suppliers for not following Y14 or other odd practices.

 

[Burunduk]

cthoper,
I heard claims like the ones you do ("loosening the tolerance will lower the cost because it doesn't matter") but since most of our parts are manufactured in-house, and the cost does not depend on how suppliers bid, I soon make the people claiming this realize that they don't make sense. If the process is so accurate because "it's made on CNC" to the point that the tolerance doesn't matter I could as well tighten it and manufacturing would still be producing only conforming parts. Then how loosening the tolerance reduces cost? The truth is that it does, but that's only because the processes are not as perfect as people think them to be, and because the tolerance does matter - but it also matters to function.

If the cost is related to pricing done by suppliers, then those suppliers who are adequate look at the tolerance value because this is what will drive the cost of their process. Converting drawings to GD&T allows increasing tolerances. I'm sure you know the classic example of the cylindrical tolerance zone for position that gives manufacturing an extra 57% tolerance compared to +/- locating dimensions, while maintaining the same functional limits.

 

[Burunduk]

Excessive scrap problems? It sounds like a confession that someone isn't paying attention to the process or the requirements. If they were they could say why the parts were being scrapped rather than just hand waving in that general direction. The truth is that a bad process isn't made better with a better part description, so they can still generate scrap in that case.

A bad process isn't made better with a better part description, but it produces even more scrap and/or nonfunctional parts when the part description is also bad. Someone's job is to improve the part description, and another someone's job is to improve the process. Management's job is overseeing both.

 

[drawoh]

I'm sure you know the classic example of the cylindrical tolerance zone for position that gives manufacturing an extra 57% tolerance compared to +/- locating dimensions, while maintaining the same functional limits.

I looked into this.

Positional Tolerances

I have done my analysis assuming that Gaussian distribution [&pm;]3[&sigma;] is equivalent to the positional tolerance circle. As noted, this is not a good assumption. If my machining process reliably does [&pm;]0.2mm, there is no cost reduction for calling up [&pm;]0.4mm. On the other hand, if I design a casting, careful design and tolerancing might eliminate all the machining.

On my new job, I am doing DFMA studies. Usually, I import a STEP file, and I create an assembly drawing. Can it be assembled easily? Is the precision required for product functionality within the capabilities of the manufacturing process? If tolerances look nasty, I create fabrication drawings and apply tolerances. If something is easy assemble and fabricate, I am not interested. In my report, I may express an opinion on the original designer.

The drawings I have seen from vendors are all over the place in terms of quality.

If I am submitting drawings with GD&T, there are two cost issues. My drawings may be very badly prepared, and the vendors will cost for the inevitable confusion and free rework. High quality GD&T may clearly and accurately describe a difficult requirement. If you call up 0.1mm[&nbsp;]profiles all over your parts, you are making most fabricator's lives difficult.

--
JHG

 

[3DDave]

(started writing before Drawoh's reply)

Without quantitative information you are guessing. Management is expecting the engineers to understand the entire process so that the engineering design decisions are the lowest cost for the required performance, not to act as experts at every technical aspect.

"Then how loosening the tolerance reduces cost? The truth is that it does..."

A statement without quantitative support and, when it comes time to add in the potential increase in cost to align sloppy parts at assembly, it may be even more costly than not loosening them. It is not certain to be true and can easily be false.

Most suppliers, when trying to sell an improvement, have case studies to back them. ASME and the Y14.5 Committee, on their claims of reducing cost, simply don't. If it was clearly advantageous then it would be very simple to accumulate that evidence. They haven't. 50 years and nothing. I think the reason is that Y14.5 isn't the product. Training is and equipment sales is.

Perhaps if there was some documented savings over many industries the OP management would have an easier time of it. And if the underlying document for some rather simple geometric concepts wasn't drenched in complex jargon and sold for text-book prices it would have greater penetration. If that value was really clear that would be seen by CAD companies providing software to do the analysis to the general user population and coordinating with CMM makers to feed back the production capability to allow designers to better adapt their designs to the least expensive tools knowing that they would hit reliability targets.

OTOH, the reality is CAD companies taking years, sometimes a decade, to even add the newest symbols to their software. Almost no one uses full-stack analysis, and effectively no one feeds back data from CMM to CAD designers.

 

[Burunduk]

drawoh,
Of course there is a normal distribution and hopefully 57% more tolerance doesn't mean 57% more parts within tolerance, otherwise it would indicate the process is out of control. Admittedly I did not read the entire article, but does your analysis take into consideration that a realistic process is not perfectly centered to the nominals? I've seen cases where a slight increase in the tolerance (nowhere near 57%) eliminated much more than 2% of scrap.

3DDave,
Do we need a case study for everything?
Why don't you require FEA software developers to present case studies to show that performing an FEA on newly developed mechanical components reduces the probability of their failure during operation?

I conclude that you never converted a drawing cluttered by dimensions with ambiguous +/- tolerances to a concise data set with functional datum reference frames, never seen how this reduces inspection time and the frequency of MRBs.

Maybe you also didn't solve any customer complaint issues by redefining geometric limits from vaguely specified requirements that lead to poor control of functional features to rigorous geometric tolerances. In such cases, the process is readjusted to produce parts that conform to the redefined requirements, which do not necessarily require any capabilities that it lacks. What is the cost of an unsatisfied customer to a company?

I can only attribute your reluctance to acknowledge the importance of implementing the principles of dimensioning and tolerancing covered by ASME Y14.5 and ISO GPS standards to the lack of the relevant type of experience. However, it seemingly doesn't prevent you from sharing what may sound like far-reaching observations and insights:

"Beginners analyze the parts.
Intermediates analyze the assemblies.
Journeymen analyze the inspection.
Masters analyze the manufacturing process.
Top masters look very carefully at the requirements."

That sounds kind of poetic.
But frankly, it feels like some random stuff written for the sake of saying something.
Beginner or not, how does one analyze the part without analyzing the requirements and the assembly? That's like analyzing a leg without knowing that it is supposed to be attached to the rest of someone's body and allow walking. Who does this?
If you're a beginner you may do a lousy analysis, but you still analyze component tolerances to answer to requirements and assembly needs, preferably while having an idea about the process and what it is able to do as well as the applicable inspection methods. Otherwise, there is nothing to analyze.
Journeymen analyze the inspection?
You analyze the inspection if inspection is your job, regardless if you're a journeyman, a beginner, or a top-master.
Likewise for the process engineer.
If you are on the design side, being generally familiar with the process and inspection should be enough.
Solutions require collaboration, not a single know-it-all that knows better than everyone what to put the focus on.
"Management is expecting the engineers to understand the entire process so that the engineering design decisions are the lowest cost for the required performance, not to act as experts at every technical aspect" - your own words.