What's new
What's new

GD&T, Drafting Question

dodgin

Hot Rolled
Joined
Mar 16, 2015
Location
MI, USA
My shop does a pretty large variety of work for various companies around the globe, and I feel like for the most part the bulk of the models and prints that we see are somewhere within a ballpark of what most in manufacturing might consider "normal". I mean, a good chunk of min and max dimensions, some companies model at one end of the tolerance or the other, dimensions corresponding, etc. Those types of things that you get used to seeing and working with based on who the customer is.

There's one company that isn't in the ballpark though - prints are full of -/-, +/+ (think -.03/-.06, not -0./-.03), with the model not being to the high or the low, but actually made out of that tolerance band. This is every part that comes across my desk, and I'd say it represents about 80% of the dimensioned features on the parts...so it's not insignificant by any stretch. I can't say with any certainty where the customer is headquartered, but based on hints I've gathered over time I'd led to believe it's somewhere in Asia.

I'm just curious if anyone has any insight or experience with this? Is it geographical? I have a hard time wrapping my head around the idea that it's somehow easier or makes things more functional to do it this way. Is there a reason for it that I'm missing or is it done strictly to be a pain in the ass?
 
My shop does a pretty large variety of work for various companies around the globe, and I feel like for the most part the bulk of the models and prints that we see are somewhere within a ballpark of what most in manufacturing might consider "normal". I mean, a good chunk of min and max dimensions, some companies model at one end of the tolerance or the other, dimensions corresponding, etc. Those types of things that you get used to seeing and working with based on who the customer is.

There's one company that isn't in the ballpark though - prints are full of -/-, +/+ (think -.03/-.06, not -0./-.03), with the model not being to the high or the low, but actually made out of that tolerance band. This is every part that comes across my desk, and I'd say it represents about 80% of the dimensioned features on the parts...so it's not insignificant by any stretch. I can't say with any certainty where the customer is headquartered, but based on hints I've gathered over time I'd led to believe it's somewhere in Asia.

I'm just curious if anyone has any insight or experience with this? Is it geographical? I have a hard time wrapping my head around the idea that it's somehow easier or makes things more functional to do it this way. Is there a reason for it that I'm missing or is it done strictly to be a pain in the ass?

You see alot of -/- +/+ in metric tolerances for shafts and holes/bores. I used to get prints like this all the time -.0002/-.0006". Can't explain why the previous wouldn't be dim'd at nominal with +/-.0002" ??
 
I've seen it from one place in Wisconsin. Come to think of it, they are international and have a few locations in China. I don't know the reason and can't think of a good reason why.
 
I guess it could be used to describe the functionality of the part..

6.000 -.030/-.060 could mean that the part has to fit somewhere that is
6" wide.. And the mating part would probably be 6" +.03/+.060...

There have been many times I've questioned why things were dimensioned
the way they were.. And then when you get your hands on the mating
parts and sometimes you even get to do the assembly.. And then you
can start seeing WHY some things were dimensioned in a strange or
complicated way..

On the flip side, sometimes you see the mating parts and assembly and
you realize the engineer is lucky drafting is done on a computer
now, otherwise the guy would have probably poked his own eye out
with his drafting pencil.
 
Don't know the answer to this question but it does remind me of something I was told back in high school nearly 50 years ago. I had a drafting teacher who posed the question: "What is the best way to draw and dimension a part or assembly?" The answer being: "Do it the way the man who hands you your paycheck wants it done."

After having been a professional draftsman at only one business for many, many years, a lot of OP's drawings managed to cross my desk. A lot of these left us scratching our heads. We had our own particular ways we expressed ourselves on our own drawings to be used in our own shops. Later on, when we started farming out some parts to be machined, we decided to dip our toes in modern GD&T in order to express our wants and needs to outside vendors in a consistent manner. We hired consultants and took courses to learn GD&T. We came to our own understandings about how to use it and thought we were on the right track.

What we found out was that GD&T was not the universal language we thought it was. Outside vendors often interpreted our intent differently than we did. We would check back our work and it was ok; it was just that different people thought it meant something different than we did. When we started dimensioning with GD&T, we started getting a lot of no quotes because vendors were not sure what we were aiming for. So we found out we had to explain what we wanted to a lot of people; basically explaining a baseline of institutional knowledge for them. It wasn't our work that was bad, it was that a lot of vendors just didn't know what to do with it.

This has changed and most vendors got up to speed so they could compete. But when I retired a few years ago, I found there were still a lot of shops who would justifiably call to get a good understanding of what we wanted just to make sure of things before they started machining. Just like before we started using GD&T. So I was left wondering? Why did we even go down that rabbit hole? It was supposed to clarify everything, but instead it was apparently just as open to individual interpretation as anything else we ever did. I am pretty sure that we missed the mark on how we used it and the people we were farming our work out to were just as clueless as we were.

So I am ambivalent about the stuff. It worked very well in house as everybody in our company was always on the same page. But I guess if you are working off of somebody else's prints, it's still a good practice to ask the customer what the fuck he wants in plain language before you end up with a pile of scrap. A telephone is still useful when interpreting drawings.
 
That's why I like drawings that are LOADED with GD&T call outs...

I love that little 'm' with a circle around it.

Yabutt, even then it is sometimes mis-interpreted! OR called out in such a way that you say fuc* it, just make it "perfect" LoL.

I have shared the stories about the QC asshat I worked with... ya, maybe he was "doing his job" but with the quality of work we bought, compared to what we made, he was a dick.
 
Yabutt, even then it is sometimes mis-interpreted! OR called out in such a way that you say fuc* it, just make it "perfect" LoL.

I have shared the stories about the QC asshat I worked with... ya, maybe he was "doing his job" but with the quality of work we bought, compared to what we made, he was a dick.

Yep, it's only as good as the guy inspecting it and how he inspects it.
 
Don't know the answer to this question but it does remind me of something I was told back in high school nearly 50 years ago. I had a drafting teacher who posed the question: "What is the best way to draw and dimension a part or assembly?" The answer being: "Do it the way the man who hands you your paycheck wants it done."

The alternate answer to that question is, Do it the way the man who will be making the part, wants it done. When I draft a print, I pretend that I will be
making the part myself, and dimension and tolerance accordingly. This includes giving wide tolerance for features that don't mattter. No solidworks
8 decimal place dimensions on ever feature!

Any print that does not have the dimensions toleranced according to the laws of gaussian distribution will have that done one way or the other, while
it is being made. If the print has goofy tolerances then I will state with absolute certainty the man making that part will rectify it with a calculator before
starting the job.

Why the HELL would you draw the print so it makes his job harder than it has to be?
 
My little pet peeve is when our engineering dept draws something that was obviously designed using fractions. and the starts rounding down to two places. On a bolt circle that can really screw up real positions when it is rounded down .005 and then automatically given +-.01. Sometime I think they forget to keep trailing zeros, other times given the datums used they don't know how to mfg a part.

Side note: why do most prints seem to call out a bare minimum chamfer around threaded holes. There is something wrong with the design if you can't leave off that last thread. On disassembly I can always see a shiny spot where the last thread is starting to pull up.

Dave
 
Don't know the answer to this question but it does remind me of something I was told back in high school nearly 50 years ago. I had a drafting teacher who posed the question: "What is the best way to draw and dimension a part or assembly?" The answer being: "Do it the way the man who hands you your paycheck wants it done."

The alternate answer to that question is, Do it the way the man who will be making the part, wants it done. When I draft a print, I pretend that I will be
making the part myself, and dimension and tolerance accordingly. This includes giving wide tolerance for features that don't mattter. No solidworks
8 decimal place dimensions on ever feature!

Any print that does not have the dimensions toleranced according to the laws of gaussian distribution will have that done one way or the other, while
it is being made. If the print has goofy tolerances then I will state with absolute certainty the man making that part will rectify it with a calculator before
starting the job.

Why the HELL would you draw the print so it makes his job harder than it has to be?

That's what I'm wondering. If I get a print to process work or a model, and these features aren't actually designed in a tolerance band I've got to take time to go through and make sure everything is programmed or processed to the mean.

When most of the features on the work are designed that way it takes a sizeable amount of time to go through and make sure everything is kosher.
 
My little pet peeve is when our engineering dept draws something that was obviously designed using fractions. and the starts rounding down to two places. On a bolt circle that can really screw up real positions when it is rounded down .005 and then automatically given +-.01. Sometime I think they forget to keep trailing zeros, other times given the datums used they don't know how to mfg a part.

Side note: why do most prints seem to call out a bare minimum chamfer around threaded holes. There is something wrong with the design if you can't leave off that last thread. On disassembly I can always see a shiny spot where the last thread is starting to pull up.

Dave

Well I can't speak for everyone, but we do alot of small holes on small parts. 0-80 2-56 and 4-40 being most common. On these we typically tap .30 deep or less. We want to keep the chamfer to the min to maximize threads. I actually program (using a 3/16 spot drill with a roughly .005 flat on tip) directly to size, which generates almost the perfect chamfer diam, assuming tools are touched correctly and such....

4-40 = spot depth .056"
2-56 = spot depth .043"
0-80 = spot depth .031"

Those ^ numbers (our machines and tooling I should add) give us the correct chamfer to not have the first thread rev "pull up"

Now to a more general approach, yes chamfer that bugger so you have a nice lead, but not so much it looks like the great blue hole LoL!! That was a pet peeve of mine, chamfer was WAAYY big, or too small so threads pull out... It's not rocket science folks :crazy:
 
Nothing I work with needs that last thread. I work in a maintenance shop so I know what most of the parts are and where they are critical. I've had to grind back the chamfer on some of the hardened tooling due to half of a thread cracked out.
#6 is usually the smallest I work with and then with plenty of length of thread.

Dave
 
Don't know the answer to this question but it does remind me of something I was told back in high school nearly 50 years ago. I had a drafting teacher who posed the question: "What is the best way to draw and dimension a part or assembly?" The answer being: "Do it the way the man who hands you your paycheck wants it done."

The alternate answer to that question is, Do it the way the man who will be making the part, wants it done. When I draft a print, I pretend that I will be
making the part myself, and dimension and tolerance accordingly. This includes giving wide tolerance for features that don't mattter. No solidworks
8 decimal place dimensions on ever feature!

Any print that does not have the dimensions toleranced according to the laws of gaussian distribution will have that done one way or the other, while
it is being made. If the print has goofy tolerances then I will state with absolute certainty the man making that part will rectify it with a calculator before
starting the job.

Why the HELL would you draw the print so it makes his job harder than it has to be?

A couple of things:

I came to the drafting table after many years on various shop floors doing a variety of fabrication. So I was well aware of the problem shop people have trying to make something off a print that some dumbass drew who had no clue as to how it would be made. I liked to consult with the shop guys who were going to machine something I designed. My boss did not like that at all - he insisted that he was the contact point with production, not me. I just did it anyway and kept my mouth shut. I would go to the shop for a smoke break a couple times a day and shoot the breeze with the guys who were working on the machines. I learned a lot that way. They liked it and my boss was usually none the wiser. Keeping a draftsman off the shop floor is just plain stupid IMO.

We also eventually went the way of 2 place decimals with a block tolerance. It wasn't good to update old prints to this method so we usually left that alone. It was ok when one was doing an original model and drawing as long as it was designed that way to begin with. It only works if everything adds up to a whole of 2 decimal places without any conflicting numbers.
 
OK, here is one thought on this. You (they) want a 1" pin to fit in a 1" hole. With some clearance.

CAD programs work with a GRID and things can be drawn to snap to that grid. Or, you must take extra time to draw them to an actual size that is slightly different than the grid. And when you dimension a CAD drawing, the dimensions are normally snapped to the edges or corners of the part or the end points of some line. They are derived from the actual drawn sizes. So the nominal part of the dimension on the CAD drawing IS going to be the actual size that it was drawn. This can be changed, but it is extra, time consuming work.

The designer figures that the stock will be 1" or perhaps a bit smaller so he figures that the pin can be a little smaller than 1". Let us say a nominal dimension for the pin would be 0996". And drills will make holes that are somewhere from a thousandth or two small to several thousandths oversize. So the nominal dimension for the hole may be 1.003". Sounds like it should be a winner.

But how do you draw them? Do you set your CAD program's grid to 0.001". I certainly wouldn't: it would be almost impossible to use. Do you draw to the grid and then use a move of stretch command to change the size by a few thousandths? That is a lot of extra work. I use 1/8" for mechanical and 1/10" or 1/20" for electronic work. For metric work I use 1mm and 10mm. Those are my "go to" grid settings. I may change them while working, but that's where I try to start from and where I go back to after using some other value for a specific purpose. And I DO USE THE GRID and the "Snap To Grid" feature for most of my work. Why? Well, it is just plain FASTER that way. So, if I draw a line, it is going to end in x.000" or x.125" or x.250" etc. And yes, 1.000". So, I would draw a pin that is 1.0000" in diameter and dimension it with a -/- range and a nominal value of 1.000". And I am going to draw a hole that is 1.0000" in diameter and dimension it with a +/+ range. The two ranges would be mutually exclusive to provide the needed clearance. Note: I said the pin and hole would be drawn as 1.0000" but in reality there would be a lot more zeros in those nominal dimensions. My CAD program can show a proton in proper scale in a properly scaled drawing of the entire solar system.

My 2D CAD program, FastCAD, allows me to create different specs for dimensions and record them with Style names. So switching back and forth between different settings is fast and easy. If I did so often I could even write macros to do this and add a button to activate those macros. One click is all that it would take to go from -0.006"/-0.002" to +0.002"/+0.005". I do have such macros and buttons for setting line widths, with and without fill. I am sure other CAD programs have the same features.

I am not saying that this is the best way or the preferred way of doing this. I am just saying that the desire to use the CAD program efficiently is probably one of the reasons for using this kind of dimensions.

BTW, before I get a ton of comments about the Offset command, yes I do use it for most of the odd dimensions that I draw, when they are necessary. But that takes more time than simply drawing a line that snaps to the grid. After all, you must input the offset value. And at times, I even go completely "off grid". But the grid is THE fastest and easiest way to draw to an accurate dimension.
 
I have always done without the grid. If the grid is small enough to be of use for a dimension, then it is hard to use, it is too fine, and requires a lot of zooming in and out.. So I just do the feature to "the" dimension, and geal with it in the dimensioning tolerances. The process of putting in the dimension is the easiest way to go for me.

Disclaimer here is that I do 3D CAD, of the so-called "parametric" type, so that it is essentially required to actually put in every dimension. Parts which are "snapped to grid" HAVE no dimension until given one, and may move around as other work is done. The sketches done to define a feature are required to be given dimensions, the process is to draw the feature sketch, then dimension it, and finally extrude or loft etc to get the model feature added to the model. The act of adding the dimension then makes that part of the sketch take on that dimension.

For instance, a circular feature is drawn, and given a diameter as the second step in adding it. Then dimensioned relative to other portions of the sketch (if any) A rectangle, with two dimensions plus a location is drawn to get it into the sketch, then the 2 size dimensions and the two location dimensions are added.

The snap to grid is really a carry-over from old-time AutoCad, etc, where you could draw the part, and the as-drawn was the part unless you erased it and drew it again differently.. Some modeling CAD will take the grid poins as "automatic dimensions" if the ;ines are snapped to them, but I find that to be more of a nuisance than a help.

In the end, the model has to be drawn to SOME actual dimension, there is no way to draw it to a +- tolerance band (which could actually be good for structural parts). That size might be maximum material, minimum material, or nominadescribing the part.l, according to what is corporate standard, or best for
 
I have always done without the grid. If the grid is small enough to be of use for a dimension, then it is hard to use, it is too fine, and requires a lot of zooming in and out.. So I just do the feature to "the" dimension, and geal with it in the dimensioning tolerances. The process of putting in the dimension is the easiest way to go for me.

Disclaimer here is that I do 3D CAD, of the so-called "parametric" type, so that it is essentially required to actually put in every dimension. Parts which are "snapped to grid" HAVE no dimension until given one, and may move around as other work is done. The sketches done to define a feature are required to be given dimensions, the process is to draw the feature sketch, then dimension it, and finally extrude or loft etc to get the model feature added to the model. The act of adding the dimension then makes that part of the sketch take on that dimension.

For instance, a circular feature is drawn, and given a diameter as the second step in adding it. Then dimensioned relative to other portions of the sketch (if any) A rectangle, with two dimensions plus a location is drawn to get it into the sketch, then the 2 size dimensions and the two location dimensions are added.

The snap to grid is really a carry-over from old-time AutoCad, etc,
where you could draw the part, and the as-drawn was the part unless you erased it and drew it again differently.. Some modeling CAD will take the grid poins as "automatic dimensions" if the ;ines are snapped to them, but I find that to be more of a nuisance than a help.

In the end, the model has to be drawn to SOME actual dimension, there is no way to draw it to a +- tolerance band (which could actually be good for structural parts). That size might be maximum material, minimum material, or nominadescribing the part.l, according to what is corporate standard, or best for

That is what I am thinking too, not to say it is never useful. I started with Autocad R16 (I think) which is a whole lot different, at least using a digitizer vs a mouse lol, to today's cad programs. I am thinking using Solidowrks, Creo, NX, and Mastercam all of them either

1) need an actual dimension input, for say a circle
2) need a constraint to keep it in place (no grid involved)
3) allow you to easily edit the feature by simply clicking and re-typing a number (well sometimes depending)

Now actually 'drafting' or creating a working fully dimensioned print is a whole different story! That is where I have found IME that cad software is all over the place. That just takes practice with the software and an understanding tolerances, gd&t, and how a shop might machine and maintain those tolerances.
 
As far as GD&T, I have found it hard to do sometimes, and also I have found that if you do it as correctly as you can, you find that others either do not, or that they have "their own version of right".. And that is not just US companies....

It's a good system, and should be as universal as "SI" for measurement. But like many things, it seems to be a "foreign system" to everyone, For most people, one never seems to speak a foreign language as well as one's native language (there are exceptions). So only a few seem to do it "right" as far as GD&T. At least I see various "versions", I am not claiming to do it right myself or to really know "right" when I see it. I may jut not "get it".
 








 
Back
Top