PDA

View Full Version : Geometric Dimensioning Tolerancing



TECHSHOP
01-17-2006, 12:11 PM
This question maybe a little heavy for the HSM. I bought the book "Machine Shop Trade Secrets".

The author has about a half-page on this subject. I remember having "Geo Tolerancing" training/class in late 1980's or early 1990's. And have pretty much "forgot", working in a place that didn't (I think)use it.

Was this a "flash in the pan" of that time, and it lingers on in "old" prints?

OR:

Is this something that has been widely adopted?

My job searching hasn't brought GDT up. But I am wondering if this should be "prioity learning" for 21st century CAD, CAM, or CNC.




------------------
Today I will gladly share my experience and advice, for there no sweeter words than "I told you so."

david_in_ky
01-17-2006, 12:12 PM
GTD is alive and well (used) in automotive
design and part checking. I see it as permanent and will be with us a long time

Evan
01-17-2006, 12:26 PM
GDT is simply a concise way of decribing unambiguously all relevant dimensions, finishes and properties of parts and assemblies using standard terms with clear unambiguous documented meanings understood by all familiar with the practice.

When applied correctly it leaves no question whatsoever what is required when making a part or assembly. It has been around for a long time and will remain so.

tattoomike68
01-17-2006, 03:17 PM
In college drafting and machine shop the books both had a chapters on geometric tolerance and dimensioning, it was fun and makes a hell of a lot of sence once you get all the nomenclature and symbols down.

scroll down this page to see a nice chart.
http://en.wikipedia.org/wiki/Geometric_Dimensioning_and_Tolerancing

other then the fancy looking crap, everyone here uses all of the main tolerances all the time even if they dont know it.

when you get a shaft running strait in a lathe with an indicator you are working on the concentricity. or maybe flatness and parallelism and perpendicularity.

when you put a part in a mill vice with an indicator you are working on flatness,parallelism, location,perpendicularity or angularity

some tolerances like "total indicater runout" over a group of surfaces can be realy tight and make the part a total PITA.

All it does is set limits on how sloppy a part can be made and the relationship of the surfaces.

fun stuff

[This message has been edited by tattoomike68 (edited 01-17-2006).]

JeffKranz
01-17-2006, 08:15 PM
I have been using GD&T for the past 16 years. I work for an automotive supplier and have run a CMM and have also instructed about 50 people on the proper use of GD&T and how to understand it. It is my opinion that most people that design parts don't have a clue on GD&T. My thoughts are they take all the symbols they have and dump them on the print and push them to the nearest dimension.

If done correct, it leaves little guessing on the design and function of the parts.

It seem they are getting better but still have a long way to go.

Just my thoughts..

RAD1
01-17-2006, 09:01 PM
I had a whole semester of it in college. I also worked for awhile at a shop making Tombstone fixtures and I used it there. As near as I know, anyplace that does Government work uses GD&T.

sch
01-17-2006, 09:11 PM
It was a required course for first yr students at the JC college machinist training program I went to last year.
Steve

DR
01-17-2006, 09:18 PM
"It is my opinion that most people that design parts don't have a clue on GD&T. My thoughts are they take all the symbols they have and dump them on the print and push them to the nearest dimension. "

What Jeff says is true.

Very few of the customers I've worked with over the years use this system.

Many who do cause lots of problems. When we really dig into what they were trying to convey it doesn't make sense.

CCWKen
01-17-2006, 09:46 PM
Yep. As Dave says, it's used in automobiles quite heavily. Even in collision repairs.

greywynd
01-17-2006, 11:34 PM
I see it a lot where I work, mainly automotive parts, but sometimes on other stuff too. I would guess that it is used correctly for about the first 75% of the stuff I see it on, but often, when read correctly, I find the last 25% will be wrong, or sometimes even contradict itself with the earlier specifications.

The unfortunate part is that others see it being used like that, then 'copy' it when designing something similar, so the cycle keeps going, and going....


Mark

Millman
01-17-2006, 11:50 PM
Just to satisy my curiosity, how could anyone come up with an idea such as Most allowable material on ANY dimension? Either a particular dimension is Correct, or it is scrap! Simple as that. This BS is starting to turn to there is no right or wrong way to machine a piece of anything. That's the cool thing about machining; it's right or WRONG.

------------------
Dave da Slave

Evan
01-18-2006, 12:36 AM
You are joking, right? Nothing can be made to exact dimension, ever. Even telescope mirrors have tolerances measured in wavelengths of light, but they have tolerances.

Since nothing can ever be made to an exact dimension it is mandatory to specify allowable deviation from nominal. Without this the calculation of tolerance stacking is impossible.

Millman
01-18-2006, 12:41 AM
Evan, maybe we worked in different environments. Geometric tolerancing is TOO vague

Evan
01-18-2006, 12:46 AM
You are free to dimension to an exact value if you wish but nobody can make it. It doesn't matter what machine you use.

There are single point CNC lathes now that can cut "billet" aluminum to an accuracy of 1/4 wavelength of green light. That's good but not good enough for a visible light mirror for exacting use. They use these machines to make infrared laser mirrors.

Even at that level of accuracy there are still tolerances.

Please explain further what you have in mind.

Millman
01-18-2006, 12:57 AM
Yes, there are tolerances, such as when 2 pieces of steel are perfectly lapped , what is between the 2 pieces. The tightest I had to make were 10 millionths in. using a microscope so you got me on the green wavelength. All I mean is tolerance structure, Ely Whitney held tolerances never before heard of. It either is right or wrong. Which is scrap, do you see what I mean?

------------------
Dave da Slave

barts
01-18-2006, 01:01 AM
What GD &T does is specify the dimesions of the part in
terms of the parts that mate with it. When I worked at
Valisys back in the 80's we used this to design software that would make functional gages that would be fit to a part based
on measurements from a CMM... interesting, but a niche market.

GD & T is used heavily in industry today. If you're really good
with it, you can make a nice living consulting (like most things).

It gets more that a quarter of a million hits on google, so
it's not a flash in the pan :-).

- Bart

Millman
01-18-2006, 01:12 AM
barts, that's what I mean. The part that mates with another part. Why is it so hard to mate a part with another using normal tolerances? Use + or- on all dimensions. Then you have to regulate the Temperature as when the parts get mated. Make sure you have the proper type of fit to permit the functionality of the end user product. It just doesn't seem that complicated.

Evan
01-18-2006, 01:13 AM
"What GD &T does is specify the dimesions of the part in terms of the parts that mate with it. "

Which is essential for calculating tolerance stacking.


Millman, 1/4 wavelength of green light is about 5 millionths. A really top quality telescope mirror will be made to a tolerance at least 5 times better than that, +-1/20 wave. The one in my telescope is 1/10 wave.

Evan
01-18-2006, 01:16 AM
"when 2 pieces of steel are perfectly lapped , what is between the 2 pieces."

An indeterminate region ruled by Van der Waals forces, trapped molecules of air and other contaminants. They have size.

Millman
01-18-2006, 01:22 AM
Evan, like I said, you got me! Tolerance stacking sounds like the inability to create the very FIRST dimension in order to compensate for the inability to recreate the following dimensions which leads up to a whole lot of scrap. HA, HA! Sounds like a way to be pollitically correct to cover up bad machining habits. That's my opinion. By the way, you are a good machinist!

[This message has been edited by Millman (edited 01-18-2006).]

sauer38h
01-18-2006, 01:35 AM
<font face="Verdana, Arial" size="2">Originally posted by barts:
It gets more that a quarter of a million hits on google, so it's not a flash in the pan :-).</font>

Britney Spears gets 60 times as many hits, and she's a flash in the pan. I hope.

Millman
01-18-2006, 01:38 AM
See, I have to know what the indeterminate regions measure. Just to say they exist, doesn't really help someone out if they have to include that into their calculations, maybe to compensate for an accumulated error.

[This message has been edited by Millman (edited 01-18-2006).]

sauer38h
01-18-2006, 01:46 AM
<font face="Verdana, Arial" size="2">Originally posted by TECHSHOP:
I remember having "Geo Tolerancing" training/class in late 1980's or early 1990's. And have pretty much "forgot", working in a place that didn't (I think)use it.

Was this a "flash in the pan" of that time, and it lingers on in "old" prints?
</font>

What other way did you use to put tolerances on parts drawings?

This is good ol' ANSI Y14.5 under discussion, right? (In the US, at least.)

Millman
01-18-2006, 01:59 AM
SAUER, Evan and I stirred the pot, watch what happens.

Millman
01-18-2006, 02:08 AM
Evan, what school did Van Der Balls go to anyway? Hell I can make up imaginary numbers and forces, we can all do that.I am a realist, it does exist or it doesn't. Dimensions are the same way, they are correct or they are wrong. Heaven or Hell? Black or White? Mass production was based on tolerances that fit the situation. If your gun is in a sandy area, widen the tolerances.

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

sauer38h
01-18-2006, 02:23 AM
As Evan more-or-less says, every dimension of a real physical object needs a tolerance, simply because of the limits of fabrication and/or measurement which we are capable of doing, or willing to pay for. How the part designer communicates that tolerance to fabricators and inspectors is what geometric dimensioning is about. Or is that not the question?

Millman
01-18-2006, 02:34 AM
Designer; yeah, I wish a few of them could actually get their hands dirty 1 time doing a set-up. Some of those guys live in dreamland. It's funny, but I feel sorry for them.

Evan
01-18-2006, 10:10 AM
Millman,

There is nothing imaginary about van der Waals forces. It's part of what holds gauge blocks wrung together. It's what makes it possible for a gecko to climb a vertical glass wall.

Millman
01-18-2006, 10:24 AM
Evan, It's still a force that cannot be measured, which brings me back to the original point, without an exact reference surface; how can an exact dimension be calculated? Now we can do the theory of Quantum mechanics and String theory!!

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

Evan
01-18-2006, 10:38 AM
"Evan, It's still a force that cannot be measured, "

It most certainly can be measured. The van der Waals forces were first described by Van der Waals in 1870. He developed the equations of state that describe them. They are the short range attractive force between atoms that are independent of the covalent bonding force.

These forces are responsible for many of the properties of materials, especially liquids. Surface tension, boiling point and freezing point all depend on them.

Millman
01-18-2006, 10:49 AM
OK, Exactly how would you or any other be able to show me in writing how these forces can be measured? Van der Walls is still theory, If anything physical as we see it; cannot be given an exact measurement or value, does it exist? Prove the theory! I can create my own theory about anything, and so can everyone else! If you stick to that theory; then tell everyone what exactly is gravity and magnetism? Van Der Balls is still a theory.

[This message has been edited by Millman (edited 01-18-2006).]

Evan
01-18-2006, 11:09 AM
<font face="Verdana, Arial" size="2">
Abstract. Non-retarded van der Waals forces between a pair of macroscopic bodies have been measured as a function of separation in the range 2-30 nm for three systems: (a) mica/mica, (b) calcium stearate/calcium stearate, and (c) silver/mica. The values of the Hamaker constant, which is a measure of force strength, are slightly larger than those predicted by a simple version of Lif****z theory, but the ratios of the Hamaker constants are in agreement with theory. Experimental reasons connected with the details of surface profile measurements, and theoretical reasons for the differences are discussed.
</font>

http://www.iop.org/EJ/abstract/0022-3727/11/5/002


<font face="Verdana, Arial" size="2">
Distributed
A gold covered AFM tip is placed above an
area of interest at the surface of a cut Au
wire. In order to measure van der Waals
forces in situ, the cantilevered probe is
moved a few nm close to the sample and
retracted again using an STM-Holder.
The defl ections of the calibrated cantilever
provide a measure of the force provided
the radius of the tips and the separation
distance is known. These two parameters
are measured directly from the TEM
images.
Van der Waals forces strongly depend on
the tip and sample shape, as well as the
separation distance. In standard STM and
AFM, these are generally unknown. In
addition, the TEM images showed a new kind
of van der Waals force-mediated material
migration, previously unobserved.
Ref. D. Erts et al Appl. Surf. Sci. 188 (2002) 460
</font>

http://www.gatan.com/pdf/nano_vanderwaals.pdf


<font face="Verdana, Arial" size="2">
Physicists Measure Tiny Force That Limits How Far Machines Can Shrink

The measurement tells nanotechnologists how small they can make extremely tiny devices before a microscopic force between atoms and surfaces, called van der Waals interaction, becomes a concern.
...
</font>

http://www.physorg.com/news6724.html

[This message has been edited by Evan (edited 01-18-2006).]

Millman
01-18-2006, 11:16 AM
Like I said, a theoritical measurement. EVAN, you're a good machinist; just not much of a realist. Without an exact reference point or surface; everything else is purely speculation and imagination. I invite you to prove me wrong! Nano, that's good. EXACTLY; how big or small is that, Another matter of opinion? Prove it. If you can't see it, does it exist?

[This message has been edited by Millman (edited 01-18-2006).]

Evan
01-18-2006, 11:20 AM
"Without an exact reference point or surface;..."

Doesn't that statement support what I have been saying? That is of course the case at the atomic level. It is the limiting factor to the specification of accurate dimensions. When you reach that level the uncertainty principle takes over.

Millman
01-18-2006, 11:23 AM
Did you really say; Uncertainty? Proves my point.

Evan
01-18-2006, 11:29 AM
Hey, you don't get to steal my point! You are the one that said GDT is too vague. I explained why tolerances are needed. All dimensions have an error band, nothing is ever exact.

Millman
01-18-2006, 11:48 AM
If nothing is exact; how do you account for any measurements in nanotechnology? Then again, how small can you go? If there is no measurement involved, then it's all speculation.

TECHSHOP
01-18-2006, 03:22 PM
Thanks guys!

The local 800lb gorilla in the late '80 set up the vo-tech and comminity collage to teach this stuff. The we can whip the Japanese industry era.

My feeling was that it wasn't really understood, and prints were usually "overspec" to the point of confusion, not the intended purpose.

And about 8yrs ago, looking to change jobs, I saw it on the prints, they were mostly auto related.

Didn't think that GDT went away, just wasn't sure how widespread.

Being an unemployed Technical Shop Assistant not sure which camp my tent will be in.

Typical workday:

Big Boss, "You do this, and I'll make a lot of $$$$"

Engineering, "Thats what The Calculations indicate is the optimal solution."

CAD, "Sure, I can draw a xxxxx"

Quality, "I don't care, your #@%@% part is no good, can't you read the print."

Machinist, "#@^&**, and your $#%^&* prints."

TECHSHOP, "Ok, I'll take your suggestion back to engineering."



------------------
Today I will gladly share my experience and advice, for there no sweeter words than "I told you so."

Evan
01-18-2006, 03:29 PM
"Then again, how small can you go? "

Currently it seems the Plank Length may be the limit. It is 1.6X10^-33 cm. Quite small.

Somewhat related here is a page on machining with femtosecond laser pulses.

http://www.physics.mcmaster.ca/optics/2a_Fs_Micromachinning.htm

[This message has been edited by Evan (edited 01-18-2006).]

Millman
01-18-2006, 03:36 PM
Kinda' sad but that is the way things are made now. BullFU==InS==T! That is exactly why our grandkids will fill out their applications in CHINESE. Because that is what they are being taught. Evidently no one can see this coming. Anytime you tell a kid "That will be good enough", are the same people who say that there are no limits or tolerances on an actual product being manufactured in the USA. EVAN are we still discussing a reference point that doesn't exist? Now you tell me just how many kids are going to graduate into the world of Femtosecond Micromachining? That's like people would rather raise their children to be a sport STAR! Why waste the money on a kid that turns out to be a JOCK; but doesn't have the slightest idea how to be productive citizen in the workforce? You really think that is GOOD parenting?

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

[This message has been edited by Millman (edited 01-18-2006).]

[This message has been edited by Millman (edited 01-18-2006).]

Evan
01-18-2006, 03:59 PM
I give up. I have no idea what you are talking about.

Evan
01-18-2006, 04:07 PM
Millman,

Exactly what are complaining about? Things that you can't see? Things you can't measure? Things that are hard to measure? Just because you can't see something doesn't mean it doesn't exist or matter.

I can measure a surface to 1/20 wavelenth of light with equipment thats costs a few dollars.

You say you work to 10 millionths of an inch. You can't see that and yet you complain about micro machining. Also, what does that have to do with the Chinese?

jburstein
01-18-2006, 04:29 PM
You know Evan, I was going to post earlier today...why don't you give up. The guy's obviously either a terrible troll, or as dumb as a rock.

I can't decide which.

-Justin

jburstein
01-18-2006, 04:30 PM
Nevermind, he was the one who was complaining about spam. I think I've figured it out.

-Justin

Millman
01-18-2006, 04:43 PM
For a musician, you;re close minded. Hard to believe. Did you lose your instinct? Justin, what would you have posted anyway?

[This message has been edited by Millman (edited 01-18-2006).]

Millman
01-18-2006, 05:01 PM
Evan, sorry to say you did not understand the difference between theory and an actual surface to create a beginning of an actual measurement. Like I said; anyone can quote another theory that can never be proven. Hell, I could make assumptions myself, but you are not teaching these people other than theory.

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

Evan
01-18-2006, 05:02 PM
Define "actual surface".

Millman
01-18-2006, 05:06 PM
Starting point. An actual starting point.

Evan
01-18-2006, 05:08 PM
But where is it?

Millman
01-18-2006, 05:13 PM
According to you, it's in another dimension somewhere in another man's mind. How did you ever have a reference point on your projects? You had to have a finite point of measurement to start with.

Millman
01-18-2006, 05:16 PM
Gotcha!

Evan
01-18-2006, 05:25 PM
This is an interferometer plot of the surface of a very good flat mirror. This mirror is flatter than your highest quality surface plate. Where is the surface? On the peaks? On the "bottom"? Somewhere in between? The difference is only a few nanometers. It doesn't matter if you are making a hammer handle but it matters if you are making a telescope.

Where would you measure from? Would you perhaps use a plus/minus band to allow for the difference? That is what GDT does.

http://vts.bc.ca/pics/ifmmirror.jpg

[This message has been edited by Evan (edited 01-18-2006).]

Millman
01-18-2006, 06:20 PM
I'm impressed with your knowledge of light patterns, Einstein, but where is any starting point on a surface that may not exist, such as your ability to sense the light patterns? Where would the blind machinist be? He who has the knowledge and skills to make parts by FEEEL?Just curious of the starting point?

Evan
01-18-2006, 06:34 PM
When you feel a surface you are feeling exactly the same thing that you see when you look at a surface, the interaction of electromagnetic forces.

It's what keeps you from falling through the floor.

tattoomike68
01-18-2006, 06:53 PM
OK YOU GUYS

machine a 1" cube flat, square,paralel,& "total runout"
&lt;.0005
If a person could make a cube 1" and hold the sum of error +/-.0001 they are a big dog.
think about that.

ok guys make that.........try 304 stainless.

Millman
01-18-2006, 06:54 PM
Yeah, I fell through the floor a couple of times in my younger days, that's what made me understand the basics.

Millman
01-18-2006, 06:56 PM
Mike, thats a piece of cake with the proper tools and set-up.

Evan
01-18-2006, 07:03 PM
You will have to come measure it yourself...

http://bbs.homeshopmachinist.net//biggrin.gif

http://vts.bc.ca/pics/cube.jpg

Millman
01-18-2006, 07:07 PM
Evan, that was funny; damn man that was quick; like I said earlier;Hell of a machinist! Got to give you credit for your humor if nothing else!

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

Millman
01-18-2006, 07:12 PM
Notice I said nothing else!!!!!!

------------------
Dave da Slave If it don't agree; get a bigger hammer, BFH

sauer38h
01-19-2006, 01:53 AM
<font face="Verdana, Arial" size="2">Originally posted by TECHSHOP:
Engineering, "Thats what The Calculations indicate is the optimal solution."
</font>

That's a sore point there. I had a draftsman once (back when we had draftsmen) who wouldn't put a +/-.004 tolerance on a dimension of a rubber pad. The pad was an important functional part of a tuned vibration absorber I'd designed, tested, and documented. This lump of a draftsman thought there was some mystical significance to the common +/-.005, and insisted that I was being needlessly fussy with my tolerance. I briefly considered explaining how I had arrived at that number, but it would have involved an impromptu dissertation on Gaussian statistics and random-walk theory, so I decided to just chew his ass out instead.

Your Old Dog
01-19-2006, 06:51 AM
<font face="Verdana, Arial" size="2">Originally posted by barts:
What GD &T does is specify the dimesions of the part in
terms of the parts that mate with it. When I worked at
Valisys back in the 80's we used this to design software that would make functional gages that would be fit to a part based
on measurements from a CMM... interesting, but a niche market.

GD & T is used heavily in industry today. If you're really good
with it, you can make a nice living consulting (like most things).

It gets more that a quarter of a million hits on google, so
it's not a flash in the pan :-).

- Bart</font>

I have a friend in Youngstown Ohio that got a fat bonus from the steel company he worked at. He went to the customer and asked him what tolerance they could accept in their sheetmetal. Turns out it was cheaper for the steel mill to give them a little more than they paid for than to make sure they didn't get too much. GD&T makes perfect sense to me. Perfection doesn't exist, it's simply a goal.

If I was going to mass produce 100 mechanical units on a milling machine, does it matter to the size how many bit changes I make? Do I put a fresh bit in for each piece I make? Does Detroit do that in building cars? Tolerance ain't like pregnancy, it ain'a a "you are or your ain't" situation. Just this rookies opinion.



[This message has been edited by Your Old Dog (edited 01-19-2006).]

Your Old Dog
01-19-2006, 07:08 AM
<font face="Verdana, Arial" size="2">Originally posted by Millman:
According to you, it's in another dimension somewhere in another man's mind. How did you ever have a reference point on your projects? You had to have a finite point of measurement to start with.</font>


Millman, I barely made it out of high school so I have to assume you have it all over me but even I know that in precise work, a pencil line has three components, a right side, a middle and a left side. If you want to make an accurate cut you're going to have to rely on a hell of a lot more than what you can actually see. The word "precise" was coined so guys like you could feel good about themselves and slap themselves on the back for being great. In a car repair shop maybe in .0005 is precise, but I suspect in a real machine shop you could go well beyond that. I suspect you're just stirring the pot. It's a pretty basic concept to grasp.