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

 

[Belanger]

What you describe is what keeps me in business

Some people won't accept change of any kind, so there's not much to do about them -- their own livelihood will die off because they will be making parts that don't conform or are more expensive.
But if you are dealing with reasonable people who are willing to listen, then stick with your approach. However, in addition to telling them that GD&T lowers cost and makes drawings less ambiguous, be ready with some concrete examples that illustrate the concepts, such as the classic square-vs-circle tolerance zone shape for positioning a hole. Or sketch out some numbers to illustrate why the MMC modifier reduces scrap rates -- presuming that's something they will be in favor of.
What you're dealing with usually occupies the first hour or two of any GD&T class.

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

 

[TheTick]

I've seen vendors mark up cost for GD&T--same part without GD&T costs less than with GD&T. Mostly because they don't understand and think it's more work.

 

[ctopher]

I have seen that too. We had one vendor no-bid because it had GD&T.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[3DDave]

Since the previous 10,000 years of civilization managed without it, it's not entirely necessary. And I've seen companies look at drawings and make parts that function correctly even though they don't meet the print, simply because they know that non-functional parts that do meet the print cause arguments and loss of contracts/work. The customer checks them on the assembly line - if they fit and work they don't bother with inspecting them.

As originally designed in the 1940s, using feature control frames and datum feature references was to allow making hard gauges that piece-workers could use to check if the parts would be useful. Now it has morphed into a giant ball of theoretical considerations with a complex vocabulary to be so specific about sub-sub-sub classes of characteristics the original intent of "make it simple, make it fast" is almost entire buried; I attribute this to the CMM proponents selling part descriptions that only CMMs can possibly validate, such as the dynamic profile tolerances, and it helps sell training classes.

At the same time, precision in somewhat ordinary tools has become very good, reaching levels that in WWII would have been state-secrets for the cost/performance ratio. If the variation from part-to-part is small enough, no one cares. They dial in one part that works and know that all the parts will work. I see things heading back to a Stage 2 mass production where controlling the tools to reliably make parts has better returns than checking to see if every part is made correctly.

Keep in mind that "geometric tolerancing" originated out of desperation to put as many parts as possible into weapons - if that same desperation doesn't exist in your company, then it isn't going to have traction to put up with the training and inspection tooling expenses or time on the CMMs. My company only ever paid much attention when $10,000 assemblies weren't working, and even then the QA guys were willing to half-ass the response (It was a failure caused by the unwillingness to control the process, so instead they were checking every part, which they did wrong. Sigh.)

 

[ctopher]

3DDave,
Thank you. Great response. You are right, very true.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[Burunduk]

cthoper,
Have you ever encountered someone who argues against using GD&T while understanding it on a high level? I doubt it. Most people who argue against GD&T and claim it adds costs have little to no idea how to apply or interpret it. At most, they sat through a training course because they had to just like everyone else in their position, and had very little involvement with it since. Once you understand the method, you understand the benefits, and understand that it should be an integral part of the engineering design process.

Engineering of a new product is a process, and if the engineer did all the other steps properly, like choosing the best concept, selecting the right materials, analyzing costs, and finally optimizing the design to minimize stresses and deformations... the same engineer can as well make that little extra effort to prepare a proper drawing or data set for product definition that reflects function and all the geometric design considerations that were made along the way. This cannot be achieved without GD&T.

 

[Burunduk]

I attribute this to the CMM proponents selling part descriptions that only CMMs can possibly validate, such as the dynamic profile tolerances, and it helps sell training classes

The "Geometric tolerances require a CMM to evaluate" is an old excuse for amateurism that dates back long before the dynamic profile modifier and had been used to complain about regular profile tolerance specifications.
Many of the same people who oppose to Y14.5 supported dimensioning and tolerancing from the inspection side claiming that surface profile always require a CMM, actually inspect surface profile thinking they check +/- tolerances, when they use a height gage and base the measurement on a surface plate reference.
Telling the difference between size and profile takes a bit of learning, and many people are too lazy or careless to learn anything but the most basic things that allow them to keep the job, and the fact that career advancement is often not related to competence doesn't help either.

 

[CWB1]

GD&T has been in common use for ~40 years and taught in trade schools most of that. If a supplier doesn't use it or claims it adds cost then my response would be to start a sourcing cost-reduction project to investigate alternative suppliers. If a supplier's print-reading is antiquated then odds are pretty good their manufacturing processes are antiquated as well. While its possible that antiquated methods are the cheapest, its not likely.

 

[ctopher]

Burunduk,
I have argued GD&T with engineers that do claim their understanding at a high level. Then get them all in a room together and bring up a dwg, they all have a different understanding/interpretation of the subject. When the meeting is done, most GD&T is stripped away because they can't all come to the conclusion of certain features.
So, how do you know who is considered "high level" on GD&T? I have worked with many engineers that claim to know it well, but don't. I'm not claiming I know it all, but know enough to know who is BS.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[Burunduk]

cthoper,
Had all the people in the room been competent at tolerancing, they would be arguing over the best way to tolerance the part on the dwg, not regarding what the dwg specification means.
I believe that you do know enough to know who is BS.
Have you ever discussed some specific GD&T callouts on a dwg with someone who argues that GD&T is a waste of money and had the impression he actually knows what he is talking about? I hadn't had that experience yet. They are usually only good at bringing up some anecdotes at which things didn't go well from one reason or another, blaming it on GD&T because they never learned it.

 

[3DDave]

This is a fundamental problem - I've thought that by now tolerance analysis software would be in widespread use and automatically check CAD assemblies, but it appears that tolerance related problems don't crop up often enough. Now, if that was because lots of tolerance analysis was being done, then the software would be everywhere to cut the cost of doing that analysis. Instead it's a case that sufficient manufacturing precision is cheap enough that it generally exceeds the requirements and there is, for most cases, no need to analyze it.

With little analysis going on there is no good way to develop expertise - worse is that the tolerances applied can be very wrong and it won't matter as the resulting products are so close to the nominal, which is sufficiently ideal, to function correctly. This gives people the impression they are doing the right thing, reinforcing their misunderstanding what the symbols were supposed to control.

I did a search for "GD&T cuts costs" and found dozens of training organizations with that claim. I found zero studies or other examples, but the concept was "reduces scrap" without relating just how that was to happen. I frequently had to deal with parts that failed inspection in spite of ample FCF and datum feature symbol decorations. I did find one company that mentioned poorly done "GD&T" massively increased costs - adding controls to non-essential features, and they would charge more for that.

Where it really has a place is when production precision is very close to the useful allowable variation, which was the case at the torpedo factory where the concepts originated.

 

[ctopher]

A lot of the arguments I hear are because most parts are created in 3D CAD, then machined with CNC, why use GD&T. I get this a lot in previous and current job.
I think mostly from lack of experience and knowledge of CAD and machining processes.

3DDave,
I agree, poorly don't GD&T increases costs.
But, I have had high level engineers ask simple questions, for example: They see a concentricity callout, then ask me what is that symbol and what does it mean.
When I ask that we have training classes, I'm turned down.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[3DDave]

That is frustrating - fortunately, concentricity has been eliminated in 2018, so no worries.

I would not stress about it, just because it will become clear when some assembly disaster occurs and everyone is looking for the reason and perhaps you can point out how a couple of geometric tolerance controls would better regulate the procurement of that part.

It's also the case that you can just put notes on drawings and create a simple document to contain language to cover whatever condition you have. Those notes is what led to creating the standard - or should have been. (I have asked for examples of some conditions that are in shipped products and no one found a single one, even though they should have been presented as useful in order for the committee to consider them. Now it's much theory, not so much practical.)

'Y14.5 is just a shortcut to doing that effort on your own, but if the rest of the procurement chain doesn't want to use it, you might as well be an expert in "C" programming in a room full of Excel users. Doesn't matter how more capable "C" is, if the Excel "git-er-done" is working no one will want to change a thing.

 

[geesaman.d]

There are two kinds of vendors: those who can read and use GD&T, and those who cannot.

To get the cost savings and quality improvement of GD&T it is reasonable to expect to change over some of your vendors. Other vendors might simply need to be abandoned in favor of better ones.

I strongly encourage you to prioritize which GD&T offer the most to you, and see if you can train your existing vendors. For example, teach your turning vendors how to use a compound datum line A-B and how the perpendicularity and runout tolerances work. To this day I do not send certain GD&T callouts to certain vendors because I know they're not set up to work with it and it's not worth figuring it all out for a single part.

 

[Burunduk]

An unambiguously defined drawing of an average complexity part that doesn't use GD&T would be loaded with notes to the level of being much more text than dimensions, and then the level of interpretation would be dependent on the willingness and ability of the user to read and follow a bunch of written instructions.
I have never seen a drawing like that though.

Can accurate tools and machines justify ambiguous tolerance specifications? Not more than the "GD&T=CMM" excuse.
As for the latter, there are enough companies that "saved costs" by avoiding GD&T just to end up with inspection of parts made to ambiguous drawings cluttered by a huge amount of toleranced dimensions and some incorrectly applied "GD&T" - on CMMs.

 

[ctopher]

We do have one engineer that has been working here a very long time, way before my time. He clutters the dwgs with notes. Most notes either don't make sense or can be interpreted various ways.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion

 

[3DDave]

" the level of interpretation would be dependent on the willingness and ability of the user to read and follow a bunch of written instructions."

Congratulations. This is exactly why some companies refuse to read a complicated standard that has generalized instructions that may or may not explain what is meant by symbols on a drawing. It's even more text in the standard to read through to find the portion that applies and it won't be specific so they also have to learn all the secondary theory that might be summed up with the procedure the inspectors typically write up anyway.

 

[3DDave]

ctopher - if they cannot write up why the tolerancing scheme is chosen and how the features depend on each other, then they haven't figured out why there is a tolerancing scheme. If they can write it up, it's a simple step-by-step note.

 

[ctopher]

I agree.

Chris, CSWP
SolidWorks '20
ctophers home
SolidWorks Legion