What assumptions can be implied on a Proper Engineering Drawing? and how to avoid it?

[vonsteimel]

Greetings,
I've been through drafting school (AS in Industrial Drafting) & Mechanical Engineering School (BS). I've always considered myself to be pretty good at drawings/drafting/blueprint reading.etc and have performed reasonable well

I work at a smaller Mfg. Company so I check all of drawings that our mini-engineering dept produces. Just the other day I found a real head-shrinker for me... A simple part that should be simple to dimension, yet I got hung up on it. It knew it was double-dimensioned but couldn't exactly pin down which ones should be eliminated. Mind you, we only use AutoCAD 2004.

So I took it home and plopped it into my old Student-Edition Pro/E, since it will not allow you to double dimension. And what I found did not straighten my brow... Still a little confused on this one.

Anyway, I've loaded up the drawing. You can see my marks from checking it.
Next are 2 screenshots from Pro/E; 1 showing the simple constraints, the other showing the dimensions that are needed to eliminate the constraints. (all circled in red).

Im at the point where I'm thinking of just having it coordinate dimensioned... Still yet, the Tangents have to be implied...? I've seen a way to dimension an actual "tangent", as they are not quantifiable dimensions. I know I've seen right-angles are implied a lot also (i.e. L-bracket)

I've got the feeling I'm over complicating it. Whats your take? How would you dimension it?
(The part gets welded onto a small shaft and functions as a simple bell-crank. It doesn't need extreme accuracy. They are produce by a Subcontractor who Laser cuts them... Yes, I've got created a DXF datafile for the CNC but it still needs a proper dwg)

 

[ewh]

You need to locate the point at the intersection of the .60 vertical and .13 horizontal dimensions to tie down the orientation of the 2.60 radius. You don't need the .35 vertical and .22 horizontal dimensions, as your part would now be fully defined.
While it could be argued that the vertical dimensions .60 and .35 could be interpolated to give you a .25 long vertical edge (ignoring tolerance stackup) which would complete the definition, this would complicate an otherwise simple part.

“Know the rules well, so you can break them effectively.”
-Dalai Lama XIV

 

[JNieman]

If you work at a manufacturer, I'd suggest talking to the person who has to make or program the part.

The most elegant drawing with the minimum dimensions required to fully define the part is a nice thought, but it may be most economical and practical to over-dimension, sometimes.

For some parts, I may model and dimension the part in a general 'part definition' view, and provide either extra dimensions, or an alternate view simply labeled "reference only" or "for programmer reference" to indicate our inspectors not to check these dimensions. Some prefer/need tangent point coordinates, some projected corner intersections.

Other times, it's better to model it correctly, only call out "critical feature" dimensions, leave it less-than-defined, and inspect it "to the model" on a CMM or comparable if you have the capability.

 

[TheTick]

Usually drawings are overdimensioned because those are the dimensions that fall out as the drafter is trying to extricate his cranium from his rectum.

Sometimes drawings are overdimensioned because too many people have input. The engineer knows what is needed to define the part, but manufacturing and inspection want dimensions that are more friendly.

 

[ewh]

As to the title question of your post, it should be safely assumed that a proper engineering drawing will imply a fully and concisely defined product, subject to only one interpretation.
As to how to avoid it, simply put on as many dimensions and as much contradictory information as you can.

“Know the rules well, so you can break them effectively.”
-Dalai Lama XIV

 

[CheckerHater]

Do they follow any actual drafting standard in your company?

The problem is, drafting standards do not allow "tangent", "vertical", or "perpendicular" symbols the way you are using them on your drawing.

Take them away and add dimensions to taste. Then call it "standard-compliant" and "functional". I won't even mention lack od GD&T.

 

[MadMango]

What is the origin of the part? I would dimension this by making the .38dia center the origin, and locating features from that. Don't dim to tangent point, but rather dim the linear lengths.

"Art without engineering is dreaming; Engineering without art is calculating."

Have you read faq731-376 to make the best use of these Forums?

 

[axym]

Hi All,

CH mentioned that he won't mention the lack of GD&T, but I will.

This thread is a very good illustration of the inherent limitations of directly toleranced dimensions and the obstacles they can create. The directly toleranced arc radii are both non-functional and difficult to meaningfully inspect. When the part has form error it can be very difficult to discern where each arc begins and ends, let alone measure its radius and center.

The bell crank part would be a natural candidate for profile tolerancing, with the toleranced radii converted to basic. This creates a well-defined zone in which the surface must lie, and eliminates the problem of double dimensioning.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[vonsteimel]

I am the one who basically wrote the program. I just create a DXF file of the part geometery and send that out with the PO. The subcontractor simply downloads the DXF into the CNC and programs the toolpath. We do this quite a bit.

If there is no DXF then the subcontractor will have to create the datafile themselves using the drawing. I want to make sure all our drawings will allow this to happen, and also clearly define the part so as to cover our arse in the event of a discrepancy... All of our drafters, including myself, are "certified drafters".

I was taught double dimensions are a cardinal sin and I avoid them accordingly. If I am going to double dim for the guys in the factory, they will be reference dims.
We try to do things by the book as much as we can without confusing anyone in the shop -- which is why we try to avoid GD&T. I know the tangent, vertical & Right-angle markings are not legal on drawings, or at least they weren't when I was in school. That is largely what this thread is about.

Some things have to be assumed to be tangent. It would be really odd to dimension the arc/line end points, like the dims that pop up when I delete the tangents in the pro/E file.
This is an old part that has been produced for years, so its more of an academic question than anything. I've implemented a drawing revision system (where none existed) and have they guys go back and update the drawings accordingly. As a result I'm getting stacks of drawings to check. Some of the dwgs were produced long ago by Aussies and other interns.etc, so some of them need to be redrawn, dimensioned to be brought up to standard.

I've attached an image showing some scenarios of dimensioning. Figure A shows what ewh is suggesting from what I gather. But there is still nothing to stop the part from being produced as shown in figure B.
To me it seems that Figure C is the best way to adequately describe the part without using GD&T. Even then, the tangents HAVE to be implied.... What do you think?

 

[CheckerHater]

I think this is the wrong place to ask how to create drawing without using GD&T

But seriously, how do you get away without placing ANY tolerances on your drawings? How do you tell good part from bad part?

 

[JNieman]

@CheckerHater, it's pretty awesome and rare, but I've been told by some, upon asking what tolerance they require, when given a drawing without them, that since we're making it on a VMC, they'll be fine with whatever it makes.

Gotta love that. Those are typically the customers who know us and already have a relationship with us, though. Sometimes they just need a quick part they can't make themselves and it's not worth the time to revise a drawing to add a tolerance block or note regarding profile and periphery.

 

[3DDave]

"I know the tangent, vertical & Right-angle markings are not legal on drawings, or at least they weren't when I was in school. "

It's not that they are not legal, it's that they aren't in a generally available standard. If you create a document that includes interpretation of the symbols and reference that on a drawing, you can create any symbol you like to mean anything you like. It's only unfair when the drawing recipient doesn't get a copy of the decoder-ring with the drawing. 'Y14.5 just happens to be one of many publically available decoding books, but you are not limited to its contents.

 

[Keith1029]

@CheckerHater and JNieman, it's a digression but quite recently I was working with a (large and reputable) new vendor; I have to assume that they were used to using only the "if it is on a CNC it is good enough" tolerance scheme because I was asked to "please take off all the tolerances on the drawing except for the critical ones". Naturally I said OK... and sent back the same drawing with all the tolerances and a couple of pointed questions (such as "how is a +/-.030" tolerance an issue when other features on the same process are tolerance to +/-.0002"", or the ever popular "if it has no tolerance then how do we know if it is good?" We went back an forth a couple of times, getting nowhere, before deciding that this was a bad road we were headed down and we had to find a new vendor. I don't think they ever did understand my desire/need for tolerances. Minus one for my communication skills I guess.