ORDINATE DIMENSIONING PER ANSI Y14.5 1998 6

[greggula]

ORDINATE DIMENSIONING PER ANSI Y14.5 1998

REFERENCE:thread1103-169768
REFERENCE USERS: snowman64 and drawoh and Tunalover and MechNorth and TheTick and CheckerRon and KENAT and kbro15

In the original thread the question was asked:

snowman64 (Aerospace)
(OP)
3 Nov 06 16:55
In a drawing, with the origin of ordinate dimensions located in the middle of the part geometry, is it proper to have positive dimensional values going in each direction from the origin?


From me:
It is my opinion that nobody answered snowman64s question correctly, however CheckerRon and drawoh were close and an additional post from drawoh was indeed elegant, however I will try to clear this up.

drawohs close to correct answer, is my favorite because he/she invoked ANSI Y14.5 1998, and followed with another elegant post in which I quote a part...

"Your drawing is your primary communication with the outside world. A good drawing is like a good written paragraph. You follow the rules of grammar, and you organize for clarity."

Then CheckerRons close to correct answer retorts to drawoh giving hint to the missing link of drawohs answer. I quote CheckerRon...

"Referring to drawoh's comment, I hope that origin in the center is a relating to a real, measurable dimension, such as the center of the part width defined as the datum."

NOW TO CLEAR THIS UP

Per ANSI Y14.5...
In Section 1.6 "DIMENSIONING FEATURES" subsection 1.8.7 calls out figure 1-31 that has an example of ordinate dimensioning, and as drawoh points out there is not an example of negative number dimensions, however there is a reason for this. The standard specifies where the origin needs to be.

Quoting from ANSIY15.5 1.8.7 .......... "See Fig. 1-31. Coordinates are dimensioned from base lines."

So what is a base line?


1.7.5.3 Baseline Dimensioning. Baseline dimensions
are shown aligned to their extension lines
and read from the bottom or right side of the drawing.
See Fig. l-49.

1.9.1 Rectangular Coordinate Dimensioning.
Where rectangular coordinate dimensioning is used
to locate features, linear dimensions specify distances
in coordinate directions from two or three mutually
perpendicular planes. See Fig. l-48. Coordinate dimensioning
must clearly indicate which features of
the part establish these planes. For methods to accomplish
this, see Section 4.

1.9.2 Rectangular Coordinate Dimensioning Without Dimension Lines. Dimensions may be shown on extension lines without the use

of dimension lines or arrowheads. The base lines are indicated
as zero coordinates, or they may be labeled as X, Y,
and Z. See Figs. l-49 and l-50.

Figure 1-49 clearly shows the 0 0 origin in the bottom left in accordance with ANSIY15.5 1.7.5.3,

If following the guideline as written. you would never need a negative dimension because your 0 0 origin will not permit it. The standard also would seem to
disallow origins that are not baselines. So no 0 0 origins in your corner mounting holes, or a center feature. etc. However a guideline is only a guideline, I therefore suggest that you use a corner baseline origin whenever possible. Whenever possible means, ...use the corner baseline origin always unless the dimensioning otherwise better describes your features.

 

[pylfrm]

greggula,

I don't think ASME Y14.5M-1994 para. 1.7.5.3 means what you imply it does. The term "reading direction" just refers to the orientation of the text of individual dimensions or notes with respect to the drawing sheet. It does not say anything about where baselines should be located within a drawing view.


pylfrm

 

[greggula]

pylfrm,
Thank you for your response, and also for correctly translating my antiquated reference of ANSI Y15.5 into the correct and modern language of ASME Y15.5. My age has been betrayed.

Your point is well taken, that the standard is often difficult to understand, ambiguous, and subject to lawyering LOL.
However the standard does indeed fulfil the minimum content to support the base line origin requirement.
The purpose of figures in the standard is to give clarity to the written instruction. The figure shows the x,y or 0,0 on the baseline and not on a center feature. The standard additionally references the section 4 for additional clarification. In section 4 there are zero repeat zero examples of a none baseline ordinate dimensioning scheme, and several examples of designs that would be good candidates for ordinate non-baseline origin, that of-course are not ordinate.

ASME Y15.5 has a continuing slant to use actual features throughout its narrative, I again quote drawoh... and CheckerRon ...

"Your drawing is your primary communication with the outside world. A good drawing is like a good written paragraph. You follow the rules of grammar, and you organize for clarity."

"Referring to drawoh's comment, I hope that origin in the center is a relating to a real, measurable dimension, such as the center of the part width defined as the datum."

I will add, that having positive dimensions in opposite directions in an ordinate scheme is begging for misinterpretation, and quite unnecessary. Why do it?

By perpetuation of the written words in the standard and supporting figures for base line origin, and the absence of any contrarian examples, written or in figures of non-baseline origin, the standard exactly guides the engineer to place the origin exactly (on the baseline).

 

[drawoh]

greggula,

You have drafting rules and you have drafting tools. The tools have improved quite a bit since 2006, when we had that discussion. Still, we have to ensure that the tools work. I have a part with a functional datum feature somewhere in the middle. I am using SolidWorks...

[ol]
[li]My datum dimensions are on both sides of dimension zero. It is easily observed which side of zero they are on. [/li]
[li]SolidWorks does not put minus symbols on the dimensions on one side. SolidWorks does do this with its dimension tables, for rather obvious reasons.[/li]
[li]I prepare drawings long before my part is finalized. Holes can move from one side of the ordinate, to the other.[/li]
[li]Critical information entered manually to the dimension, will not update when the model changes. Manually inserting minus symbols to dimensions on one side of the datum feature is very bad CAD practise. [/li]
[/ol]

In summary, there is no problem. If we try to solve the problem, there will be problems.

--
JHG

 

[greggula]

Hi Drawoh,
I was certainly hoping to hear from you. Thank you for your response.
We worker bees don't always get to select the tools we have to use and the standards aren't laws they are guidelines. In the Printed circuit Board world where the designs are rather flat my argument holds a bit more water. Ordinate dimensioning can be very handy. Still I was recently caught by a rigid flex design. Which brought me to the original post.
The design can be described as 2 rigid PWBs connected together with a flex ribbon board. For No good reason the origin for the Fab drawing was placed in the center of a lower left mounting hole. Why do that when you have a fine corner, a real feature, not a point in space like the center of a hole, that can be your origin? Alas when my number didn't add up I discovered that is was the origin location not easily discernible across the page. Had the origin been in the exact corner or negative numbers had been used or ordinate dimensioning had not been used or the origin was placed in the corner, the error would not have occurred.
PS Had I been more observant, perhaps the error would not have occurred. I with 30 years engineering experience made this error, so how more likely is for a junior engineer, or some one in manufacturing to error as well.

In conclusion, my answer to summer64 would be to avoid using positive dimensions in both directions of the origin, even to the point of not using ordinate dimensioning, but best answer would be to place the origin where this would never occur anyway.

 

[drawoh]

greggula,

A point that came up in thread1103-420481 is that PCB manufacturers do not use GD&T. Do they use ordinate dimensioning? If you need to explicitly dimension a PCB (really?), why would you do anything other than a dimension table?

I thought that PCBs were fabricated straight from the model data.


--
JHG

 

[ewh]

While I agree that negative dimensions in an ordinate drawing may add some complexity and should be avoided if reasonably possible, I do find a problem basing practice on figures without supporting text.
"The figures in this Standard are intended only as illustrations to aid the user in understanding the principles and methods of dimensioning and tolerancing described in the text." (ASME Y14.5-1994)
While there may not be any negative values used in the figures, the text does not directly address this. Basing a prohibition only on the figures used to expand on other guidelines is invalid. There may be drafting "errors" in some the figures, but since the figures aren't addressing those areas, they can be ignored instead of followed.

The above figure violates ¶1.7.1.4, and showing it in a figure does not mean that it is good practice to cross dimension lines when it is avoidable.

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

 

[greggula]

hi drawoh, YES REALLY
I'm giving allot you of stars anyway!
Of course PCBs need to be dimensioned not only for MFG but QA. I need the dimensions as well.
I receive a drawing from an ME, and use that as instruction to draw the outline of the PCB and for the hole locations.
A PCB is Part of an assembly with very tight tolerances. Connectors on one pcb mate to connectors on another PCB which are all mounted in an enclosure. In aerospace or telecommunications these assemblies are the ultimate in jigsaw puzzles in order to achieve the smallest packages. I know you new that.

 

[pylfrm]

greggula,

Perhaps my question is a bit off-topic, but I'm curious. Are these PCBs you work with just dimensioned, or are they also toleranced? If they are toleranced, what sort of schemes are used?


pylfrm

 

[SwinnyGG]

Dimensions are scalar, not vector, quantities- ordinates included. They should not be negative, ever.

I use ordinate dimensions on basically every drawing I produce, which is a lot- and double-sided ordinates occasionally. I've yet to have a part come back wrong because someone didn't understand a double sided ordinate.

 

[Tunalover]

I hat to nit pick but there never was an ANSI Y14.5M-1998. ASME took over from ANSI in 1984 by revising ANSI Y14.5M-1982 and issuing as ASME Y14.5M-1984.

ElectroMechanical Product Development
(Electronics Packaging)
UMD 1984
UCF 1993

 

[greggula]

Printed circuit boards are dimensioned and tolerance as any other mechanical design, complete with holes and champers, etc

 

[drawoh]

greggula,

PCB manufacturers do MBD. They don't read the drawings. I have modelled PCBs in SolidWorks and then dimensioned the outlines and the holes, however, this was an instruction to the PCB designer. They used PCAD or Altium to do the actual PCB design.

I did not apply tolerances on my drawings. I expected the PCBs to be made to PCB tolerances, which are pretty good. If I do apply tight tolerances somewhere, what will the PCB designer do about it?

--
JHG

 

[greggula]

drawoh sir!
I have been designing PCB for 30 years. Currently using Mentor Expedition software at Raytheon.
For what we call the drill and trim drawing, all the mechanical features are drafted to a drawing using traditional drafting standards and dimensioning, just as any other mechanical part, like a door hinge for example.
What some non-pcb designers may be getting confused about, may be what we call artwork. The artwork is supplied to PCB manufactures in the form of what is called Gerber files. These files control the etching of copper off of sheets of copper, leaving defined conductor path suitable for the flow of electrons. These etched sheets of copper stacked or sandwiched inside a fiberglass like dialectic material, forming an assembly of layers, of conductor paths. (Multi layered PCB)
The conductor paths are usually not dimensioned except on very rare occasions for reference. So lets not confuse the mechanical features of the form fit and function of a PCB, with the etch process of the conductor paths.
reguards,
greggula
PS when do I get a star!

 

[greggula]

drawoh,
Although it is true that much of the machining, if it be by drill bit, laser, saw or hammer is do done using electronic files. Still we in engineering are to follow that old rule which is, Engineering shall not tell manufacturing how to build it, just what to build. Thus is true for PCB documentation.

 

[Tunalover]

Guys,
I've looked through the IPC drawing standards and IPC has fully adopted D&T per ASME Y14.5-2009. But I have yet to meet a PCB designer who cared one iota about using it. It's the old saw: it's not the design/drawing that makes parts fit, it's the "dead nuts" manufacturing methods. Since drawings are legal documents, when parts are poorly dimensioned and toleranced, meet their drawings and don't fit, then the liability falls solely on the design activity. The manufacturers are off the hook because their parts met their customer's drawings.

BTW, ewh, I've been drafting for very long time and I've never heard anyone say that crossing dimension lines with leader lines is poor practice. Why would it be? What's your source of information?

ElectroMechanical Product Development
(Electronics Packaging)
UMD 1984
UCF 1993

 

[SwinnyGG]

Since drawings are legal documents, when parts are poorly dimensioned and toleranced, meet their drawings and don't fit, then the liability falls solely on the design activity. The manufacturers are off the hook because their parts met their customer's drawings.

This is precisely why GD&T exists.

Interpreting the fact that drawings are legal documents as a reason to not use them is so backwards I don't even know where to begin.

Model based manufacturing and manufacturing from drawings both have the same problem with regard to what happens when parts meet specification but don't function.

 

[ewh]

Tunalover
Besides very strict checkers when I was a jr detailer, Y14.5-1994 (we're behind here) ¶1.7.1.4 states "Crossing
dimension lines should be avoided. Where unavoidable, the dimension lines are unbroken."

edit: On the board, the luxury of moving dimensions around easily wasn't available. Some thought as to dimension placement had to take place, such as considering which features would require a leader to define and allowing space for those leaders, crossing extension lines if necessary but avoiding dimension lines wherever possible.
It is not "shall" but a "should" and there are many situations where it just isn't worth the effort. When it can easily be done though, it should be.

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

 

[BUGGAR]

What do the Machine operators want to see on Ordinate Dimensioning: All dimensions positive from the origin regardless of direction, or with minus signs when appropriate? This would apply to where we put 0,0 on the dxf drawing.
I'm finding my dxf files go straight to the machine (in this case waterjet cutter) and out comes the part. I know the operator has to set 0,0 on the machine but I don't know how he does it from my dxf's. I set 0,0 on my dxf at a definable location on the part. I use positive ordinate dimensioning with no minuses (AutoCad) and it works in my case but is this universal for machine files? My operator has never questioned me and I never thought to ask him.

 

[SwinnyGG]

The parts I typically deal with are machined, with *most* parts done on a manual machine or on a 4-axis CNC.

The agreement between our design staff and machine operators (which took some time to work out) is that zero points for ordinate dimensions are set based on 1) the two primary datums on the final dimensioned drawing or 2) One datum face and the perpendicular feature which is easiest to locate or 3) One datum face and the feature which is cut first.

Those three options are evaluated in order and the decision is made based on what the part tolerances are and where various features are located, with an eye toward making life as easy for the machinist as it can be without compromising part function.

In short, there's not a simple answer that will apply 100% of the time and keep everyone happy; so you either have to make a decision and stick to it regardless of grumbling (which has potential value because of consistency), or work things out with your production staff as we have.