GD&T discussion topic for all of you nerds out there

Salutations my Nerds, inspectors, engineers, and Ladies,

It has occurred to me that ATS has no where that Geometric Dimensioning and Tolerancing is discussed much (if at all) here on our humble virtual rock called ATS. I find this to be odd and not forth coming seeing as much content I have seen over the years (and years) I have been a member of this digital space especially in regards to folks that often speak of and talk about alien technology and ancient technology, etc. I would love to open the floor to talk about GD&T especially the Y14.5 2018 and just geek out with you all involved with utilization of such.

As always my friends, the first beers on me. Stay thirsty my friends.


B-man

a reply to: Brotherman
Hey Brotherman, nice to see you around

Having no idea about the Y14.5 but knowing what this is about..I start by punning that a zero tolerance policy can lead to bad things happening or not happening at all. Schrödingers Tolerance.

With a little tolerance life is so much easier. Oh the daily grind, too much tolerance and you are becoming sloppy, too. We could get into heated discussions and expanding on them, or we keep it cool and contract the topic.

Coming back tomorrow to see how this thread developed.

a reply to: ThatDamnDuckAgain

I don't mind zero point tolerancing because they usually come with bonuses at least in MMC and LMC which is better than nothing. I was recently working on some parts that had zero point tolerancing with an mmc and projection call out in tenths (for you other folks that don't use freedom units) +/-.0003" on diameter and 0.0 and with the projection they wanted some bull * held at 1.032" cylinder with +/-.0005". Sometimes it is very apparent when an engineer that isn't a machinist so called engineers something.

Besides tolerance is important, its how much you can be good or bad one way or the other till your considered junk =D

a reply to: Brotherman

Hey, Brotherman.

I didn't know there were engineering certifications involved. It's neat stuff and totally nerd-worthy geekout-material, but...

Fer what? Like, 3-D printing projects and stuff? I get where you are going with alien/ancient tech and surface tolerances and stuff, but I dunno? What are you thinking?

In terms of my personal encounters with tite tolerances and 3-D printing and consumer products: I think it mostly sucks. You have to now break in to remote controls to change the batts.

originally posted by: Brotherman
I would love to open the floor to talk about GD&T especially the Y14.5 2018 and just geek out with you all involved with utilization of such.

That seems sort of like starting a topic called "spiders" and saying "Spiders, the floor is now open for discussion". It seems customary for the person initiating the discussion to get the ball rolling by talking about it first and your OP doesn't do that, it just mentions the subject. But at least you finally mentioned something about it in your second post, so that's good.

Anyway, dimensioning and tolerancing has been a significant part of my work. I won't name a specific part or a client but just for an example that many people might be able to relate to even if not nerds or geeks, is a part like this, a male electrical connector:



I won't name the customer but it was a huge automotive company. That male electrical connector is supposed to fit snugly in a female connector and make a good connection. If the fit is too loose it won't make a good connection, and if it's too tight there are problems inserting the tab.

So the automotive design engineers decided how much nominal interference they wanted for a good fit, and how much tolerance could be allowed before it got too loose or too tight. They then split the tolerance between those two components, which sounds perfectly logical to someone who knows nothing at all about manufacturing. But to people who understand how things are manufactured, it's not logical at all.

Why? It turns out the thickness of that male part can be very tightly controlled to reasonably tight tolerances.
The female part is also formed from rolled metal so the thickness of the metal itself can be similarly controlled, but that's where the similarity ends. You then stamp out what will become the female component, and have to form it by bending it into the right shape. This is a much more difficult process to control precisely than just controlling the thickness of the male tab.

So to make the dimensioning more logical, what we suggested to the automotive design engineers was that instead of giving half the tolerance to the male part and half to the female part, it would make more sense to put tighter tolerances on the male part and looser tolerances on the female part which would result in the same interference as their original design which gave each component half the overall tolerance. We finally convinced them of the logic of this after explaining how the manufacturing processes work, but do you know how hard it is to get specifications changed for cars that are already in production, even if the changes make perfect sense? Well it wasn't easy because apparently the changes are supposed to be tested and verified.

a reply to: TheDiscoKing

I have been a welder for a good many years, a machinist for most of this decade in both manual and cnc, and now I am an inspector. I work in a job shop and work with a lot of sub projects from anything to infrastructure like energy to trains all the way to government work in the defense realm, all the way back down to working on porting farmer Johns 16 cyl engine and sleeving it, it doesn't matter. I also work alot using different tech in CMM and LLP style equipment.

This conversation is all about form, function, and location. Think about the chick with the nice ass that doesn't cook real well but looks good doing it? In a lot of ways GD&T is like that unless used and drafted by folks with a great understanding of engineering and manufacturing principles.

My dumb ass intro is out of the way, what can you tell me about what you make or bitches gripes and complaints about drawings and how # works on an assembly phase? Id love to know (I secretly hope you work in a beer factory and can tell me how they engineered tabs to open beer cans without snapping)

a reply to: Brotherman

...what can you tell me about what you make or bitches gripes and complaints about drawings and how # works on an assembly phase? Id love to know (I secretly hope you work in a beer factory and can tell me how they engineered tabs to open beer cans without snapping)

I do not work in a beer factory. Somehow that one passed me by.

Ugh, maybe something will come to me, but I have only ever been on the receiving end of very well manufactured equipment, and I have not been part of a testing phase where something just flat out didn't fit together. I worked in the medical field for my first career. The only thing that jumps to mind immediately is general complaints about some devices made from silastic. One interesting example may be the eternal fallibility of the Kerrison Rongeur; no amount of subtle precision can match the imprecision of human anatomy.

a reply to: Arbitrageur

Its good to see you man. You were my first friend I made here on the old system of ats too which is really cool.

The part your showing there some folks would like you said think just one thing over the other and split the difference just makes sense. With all things considered its function not one side to the other is equal, Im not an engineer or tester but that was a really cool thing to read about.

These days I am having issues when hyper critical parts come with prints that require cmm and they only use # like 2 datums, bull# engineering when you get like .125" in basic dimension but profile to surface you get like +/- .002 and on the other hand I get to work with some really well designed and thought out parts and systems where gd&t is almost like a painting of a brilliant mind.

Talking here when so many folks talk about so many things speculative about what is what I thought I would start this conversation.

good to see you my friend.

a reply to: TheDiscoKing

Im going to have to look up some of the stuff you said to know more about what your referring to. I mainly work in a machining environment but sometimes work with other materials than metallic stuff. I wish to one day understand the beer tab conundrum so forgive me for trying to goad that out of you from the jump, but i really wanna know.

a reply to: Brotherman

I think the probability of addition instead of subtraction is the likely scenario used.

What I mean is that instead of machining a chunk of material, essentially you print or extrude the desired work piece. Tolerances become less and less problematic as the machinery being used does exactly as is programmed.


IMO anyway.

a reply to: Brotherman

The silastic is a type of silicon rubber that is inert and works well for all sorts of medical and surgical applications. The rongeur is a surgical instrument that works well in one direction but always needs help making it through step 2.

This is super cool stuff, GD&T. I went and took in a short vid on YT to get oriented. I think that the Datum Simulator and runout are the two things that interest me the most.

Don't EVEN tease me with a thread to geek out on GD&T in, I live and breathe it. I could go on and on. I have stories...

One of GD&T's many advantages over the old school plus/minus tolerancing system is diametral tolerance zones.

Plus/minus tolerances have square tolerance zones that fail parts that would actually function in the assembly.

A torpedo factory technician named Stanley Parker was the first to realize what was going on and come up with the concept of True Position. The cool kids call it 'position' nowadays due to changes in the standard over the years.

Anyone out there specifying "Zero at MMC" for anything? I LOVE LOVE LOVE having to explain to suppliers that zero at MMC does not mean they have to make impossible "perfect" parts. /sarc

What's the first rule of GD&T?? Anyone? limits of BLANK controls the form...anyone? Bueller?

a reply to: JinMI

GD&T has a lot more to do with ease of manufacturing with engineering intent. One of the best examples I can use is the idea of coupling surfaces with pistons using maximum material conditions meaning the hole sizes are of the smallest diameter giving you an M callout on a feature control frame. These parts of the engineers intent to me are interesting because what there are more or less saying is that, the dimensional diameter doesn't mean as much as the function or a feature location so they let you make a hole bigger or machine smaller (if LMC is invoked) to allow form to fit. Using this also requires measure of form versus 2d features as well which I also find to be very interesting from part to part.

a reply to: tjack

I agree with you. Some folks also think surface to profile tolerancing means if its .010 they have over all of .010 but forget about location in xyz that it needs split or factored. True position callouts are always the bane of my life when it comes to threaded features with a projection and I think alot of that is because not all of us use the same gauges and torque for whatever shoulder bolt needed for whatever projected lengths, either way I hate those they can kiss my shiny metal ass.

originally posted by: TheDiscoKing
a reply to: Brotherman

The silastic is a type of silicon rubber that is inert and works well for all sorts of medical and surgical applications. The rongeur is a surgical instrument that works well in one direction but always needs help making it through step 2.

This is super cool stuff, GD&T. I went and took in a short vid on YT to get oriented. I think that the Datum Simulator and runout are the two things that interest me the most.

remind me to talk to you more about some of the things you said here. I got a lot to say as both an inspector and a machinist that you may really find interesting. Maybe one day we can make a real ufo.

a reply to: Brotherman

My dumb ass intro is out of the way, what can you tell me about what you make or bitches gripes and complaints about drawings and how # works on an assembly phase? Id love to know (I secretly hope you work in a beer factory and can tell me how they engineered tabs to open beer cans without snapping)

Look up Daniel F. Cudzik ... inventor of the Sta-Tab.

A very good book on that subject and others is Invention by Design; How Engineers Get from Thought to Thing.

a reply to: Brotherman

Lol as a technical designer coming from high-end watch making, I love to make your life hard with unnecessarily over the top tolerances.

As for the meta talk, intended or not. It does not compute in an infinite setup with infinite possibilities, it only applies on a timely limited experience where you have a limited amounts of tries...

SnF for the effort though...

a reply to: Brotherman

Such an exciting subject?

Sorry but it won’t impress the ladies. Nor was it ever intended to.

I was trained in it by The General Motors Corporation and passed their in house examination. I am also Six Sigma certified as Black Belt.

I know it but WTF. Based on his thread I will assume that you are toasting marshmallows and Twinkies on the fire at home tonight. However, I hope not and you’re chasing your significant other around your home.

Its "in brief" basically 3D design such as to account for the +/- stack up tolerances in the X Y Z axis so are all accounted for so as to eliminate any and all errors in manufacturing straight off the design board. My training book is over one inch thick and Y14.5 2018 is necessary. Otherwise CNC machines or other AI devices wont we able to function in manufacturing, assembly and operation. Thus we have CADCAM. Same goes for CMM machines.

I was in Quality Assurance for decades as an executive in manufacturing. I worked for large publically traded corporations such as General Motors, HARSCO and Air Products & Chemicals. I know all the specifications including military ISO-9000 and QS 9000, NASA, ASTM, ASME and others as I worked in the manufacture of nuclear electro-mechanical devices on stealth navy warships and submarines

I also held a Top Secret Security Clearance and have served as an Expert for the US DOJ in criminal fraud trials in DOD matters along with the US Commercial Nuclear Reactors.

I was forced to become a Mechanical Engineer as that what Grampa was. I find Geometric Tolerancing and Design work in general incredibly boring but I did make coin so it paid the bills.

I can but I hate design work

"Engineers are the dumbest smart people you'll ever know"- The smartest engineer I've known circa 1992 when I started in the machining trade.


And F@ zero point tolerancing.

a reply to: Lumenari

I responded to Geometric Tolerancing in another post but my hay was made in Statistical Analysis. Specifically Predictive Failure Analysis.

Back in the 2000's I inherited a Company issue that was losing over 20% of their annual gross revenue to warranty claims in the "trucking" industry.

In brief I called Harvard University’s Department of Statistics and they gave me the contact information for Dr. Robert Abernathy. He’s formerly of the SR 71 Black Bird Project at the Skunk Works.

Dr. Robert Abernathy

He had just left industry and was quasi retired. He agreed to work with my IT programmer and we linked our global warranty database with over 20,000 dealer and customer locations. That also meant merging his Weibull Analysis Software with our Microsoft Access, Excel and Oracle financial systems such that they all talk to each other.

We also backflushed prior years data to tune up our assumptions as they are many to determine especially as to how long the product will stay on the shelf prior to sale amd installation. We also found many hot spots on certain OEM's engines and they ended up moving oir product to accomodate as the warranty in the Heavy Truck industry is at times is THREE MILLION miles!

In two years based on the data, we reduced the warranty loss in dollars from 20% to less than 0.5%. That was the result of a Corrective Action Committee that I chaired. Most warranty payouts were determined to be all design or component related failures. We also cleaned up our supply base.

That software could have been modified to predict Covid-19 hot spots in 2020. I contacted my original IT programmer but he wasn’t interested.

I am sure that Fauci and the others knew exactly what was going to happen