Dimension origins in the basic views 6

[johnfalky]

Our machine shop prefers the dimensions to cascade from the upper left hand corner in the base 2d view of the part. But this does not appear to a basic standard.
In working with a number of defense contractors that have extensive documentation and procedures on drawing standards, I was surprised to see that this basic requirement, namely the basic ORIGIN of the dimensions is not specified. Some research on the net of drafting schools and other standards also show confusion and a wide divergence on this very basic requirement. In an examination of various drawings and samples from technical schools and universities, I have seen the origin of the drawing in a basic view starting at the lower right hand corner, upper right hand corner, upper left hand corner or lower left hand corner. One manual from a technical school proudly stated "read up from the bottom" for the dimensions. Another major Aerospace company crowds the dimensions in starting from the lower right hand corner, creating an incredible mess between the projected views! Others split the dimensions, essentially creating the classic error of chaining the dimensions, thus causing a potential huge tolerancing error buildup. Overall, my impression is that a lot of technical schools and universities appear to be picking up tribal knowledge, ANSI standards notwithstanding, (even tho the standards are quoted over and over again) and are repeating these standards without questioning their actual validity. So, after this lenthy statement, does anybody know of a document out there that specifically states in clear and unambiguous language the following: "The origin for the basic dimensions should be the upper left hand corner in a plan view of the part" or whatever?

 

[johnfalky]

Allow me to give a sample of shop floor brouhahaha and then I will rest my case and disappear. Case: A technician in a Local Aerospace high precision machine shop, using ISO quality procedures for process and specification control proudly submits his work for final inspection. (Its a valuable defense contract part). Each rivet in each countersunk hole is perfectly positioned, perfect bucktail on the other side. the rivet is absolutely flush. The technician glows with pride and arrogance. There's only one problem; Every Countersink is also perfectly oversized. ALL of the rivet heads are floating perfectly in said countersinks. The part is rejected. The angry technician walks off and complains bitterly to the shop manager, who comes over and insists on a buyoff. His argument: "The part is per print, buy it off" I refuse. He shows me the drawing spec for the countersink, then proudly shows the quoted document the shop is using...Which contains a very neat, hand-lettered note "countersink diameter tolerance plus or minus .010 inch." I reject the document as having been altered after release. The shop manager is visibly angry and the generator of the document, an elderly designer, is called and he states that he entered the note "based on experience" I note that he is incorrect and that he had made an unauthorized alteration to a technical document that in turn has generated a small fortune in incorrectly manufactured parts. In this case, the drawing used a spec, but did not directly specify the countersink dimensions. The shop's failure was in an uncontrolled specification. Because NONE of the CAD design staff and management knew what a correctly installed rivet head was supposed to look like they had made scores of these parts and other inspectors had bought off on them, using the incorrectly altered specification. The scrappage was costly, the elderly draftsman never spoke to me again and quit about a month later..and I was transferred to the night shift...Based on this experience, I take the "extra step" with my drawings. The end result is a better end product, and actual "thank you" phone calls from the shop floor.

 

[MadMango]

This example is a training issue, on several fronts.

Engineering- all critical specifications should have be referenced in the general notes of the drawing, and ISO procedures should have been followed regarding change control.

Manufacturing- parts should not have been made from a redlined drawing without approval (stamped and signed) from engineering.

Inspection- the "CAD design staff and management" didn't need to know what a proper installed rivet looked like, but your inspection personnel should have been trained to know the difference between accept/reject criteria.

[green]"I think there is a world market for maybe five computers."[/green]
Thomas Watson, chairman of IBM, 1943.
Have you read faq731-376 to make the best use of Eng-Tips Forums?

 

[Heckler]

It's extremely common in the defence industry to call out hole preps IAW a MS or NAS Spec: two examples, Port/Fluid Fittings and Helical Inserts. So this was a manufacturing mistake by not having the correct spec.....experienced a lot of that. It certainly saves time just calling out the hole prep verses detailing all the MS/NAS Port Holes. That's my two cents....anything more I will have to charge somebody.

Best Regards,

Heckler
Sr. Mechanical Engineer
SW2005 SP 2.0 & Pro/E 2001
Dell Precision 370
P4 3.6 GHz, 1GB RAM
XP Pro SP2.0
NIVIDA Quadro FX 1400
o
_`\(,_
(_)/ (_)

"There is no trouble so great or grave that cannot be much diminished by a nice cup of tea" Bernard-Paul Heroux

 

[ctopher]

Follow the drawing and all specs, use machinists experience and training, don't use mark-ups ... no one will get hurt.
If someone wants to list how to machine a part on a drawing (a side from special circumstances, like molding) then that person is not experienced enough to make drawings in the first place. IMO.
Sorry, not trying to knock anyone, just real world.

Chris
Sr. Mechanical Designer, CAD
SolidWorks 05 SP3.1 / PDMWorks 05
ctopher's home site

FAQ559-1100
FAQ559-716

 

[RedPen]

The standards clearly state engineering drawings specify the finished part and not the process(es) used to create it.

Okay, then the real world interferes. What one does when the fabricator(s) are in-house/captive is different than what is required when everything is sub-contracted out.

There is no standard that specifies a corner or ??? for the dimension origin. There are too many variables for that to work in all cases and dimension origins should be chosen by the requirements of the part/design. That being said, there is usually a way to accomodate the desires of the machine shop AND the requirements of the part.

Generally, don't specify process on drawings. First, the engineer/designer/drafter is seldom the fabrication 'expert'. Second, someone will come up with a new, better way of making the part (stronger/faster/cheaper/etc) but can't use that method because the drawing specifies a different process.

A good example of this is "flow-drilling". I needed 36 holes with 1/2" diameter female pipe threads in an extruded 1/8" wall aluminum square tube (fluid manifold). My out-of-date machine shop knowledge said these threads were virtually impossible to achieve. The shop said "no problem" because they could "flow-drill" the holes. "Flow-drilling" extrudes the wall material into an internal boss, pierces it and creates threads thru the combined wall and extruded boss thicknesses. Neat!

Specify performance instead of process and you'll get what you want without "tying someone's hands". It doesn't matter how a part is produced as long as it meets the functional requirements (including delivery and cost).

Best regards,
RedPen

 

[drawoh]

ajack1,

Even if the fabricator does not look at the dimensions, the inspectors do. If the inspector rejects something, you need to clearly tell the fabricator what is wrong. There is also more than one way to move from scale CAD model to toleranced dimension.

I have been looking into having parts fabricated by processes such as Stereo Laser Lithography, where the fabricator would work off my 3D CAD model, rather than my drawings. Since the shapes are likely to be weird, we are going to have to work out an inspection strategy. Probably, I will have to design inspection fixtures of some kind. I am sure I am in for a long chat with someone.

JHG

 

[ctopher]

drawoh,
I have been thinking of that also. But not 'til next year.
Please let me know how it works out.

Chris
Sr. Mechanical Designer, CAD
SolidWorks 05 SP3.1 / PDMWorks 05
ctopher's home site

FAQ559-1100
FAQ559-716

 

[TheTick]

I think you're missing my point. There's a difference between "design for manufacturability" (DFM) and "drafting for setup".

 

[drawoh]

RedPen,

No one here has advocated specifying processes on drawings.

I do not like walking into a shop or even to a colleague's desk and finding my fabrication or dimension drawings covered in pencil marks due to somebody recalculating all my dimensions. Even if I am not responsible for an error, it is still an error.

For example, I was about to produce an outline drawing of a printed circuit board to go into a housing I was designing. I asked the PCB layout person if she could work in metric, and she said "Yes" in a thick eastern European accent. When I cruised past her desk a couple days later, I noticed that she had taken a pencil and converted all my dimensions into inches. Clearly, the CAD software she was using was limited in capability.

The question "Can you work in metric" has a different meaning when I am using SolidWorks. I went back to my desk, changed the PCB drawing and model to inches, I wiggled all the features so that they worked out to round inch values, then I re-issued my drawing.

No more pencil marks. No need to check pencil marks. The mechanical drawings work because SolidWorks makes everything line up.

Both of my drawings met all the requirements for an engineering document. One of them was very much more clear and reliable than the other one. The fabrication drawing of my housing is in metric, and the PCB hole locations are not round numbers, which is unfortunate. It was CNC machined, so it does not matter very much. Everthing lined up, so our communications worked, and the PCB fabricator and machinist did their jobs.

If I had wanted to keep my fabrication dimensions are round numbers, I could have changed my PCB drawing to dual dimensioning, or I could have just changed the units to inches. In mass production, I might have. In limited production, I judged that communicating with the electronics people was more important.

JHG

 

[ajack1]

drawoh, that is not true, you may choose to inspect to dimensions but it is not necessary.

What you say about inspection fixtures is correct, you will need them and certainly in the automotive trade the models will show the location areas and the locations (datums) beyond that nothing is dimensioned other than tolerances. A CMM is perfectly capable, with the correct software to inspect the part.

We have ISO 9000 and also we are audited by amongst others Ford, BMW, GM, Audi and Bentley and supply them all as first and second tier suppliers.

Working purely from models is now the norm in the automotive trade, certainly for panels and fascias.

If you do go down the road you are talking about in a years time you will look back at this and what you are now doing and laugh.

 

[ctopher]

Seems to me ... Automotive and aerospace are different animals. Automotive can go without proper drawings and inspect with different methods. Aerospace/military, you are required to have correct & complete drawings and the parts are checked to the drawings. Material certs are required and approved machine shops are required.

Chris
Sr. Mechanical Designer, CAD
SolidWorks 05 SP3.1 / PDMWorks 05
ctopher's home site

FAQ559-1100
FAQ559-716

 

[Hydroformer]

That's a fact ajack !

Very few dimensions are shown on automotive drawings. Mostly limited to hole diameters. Everything else is spec'd out by GD&T referenced to the CAD model. I'm quite surprised to see so much discussion on (old school) dimensioning and so little on GD&T applications.

My favourite drawing note: "Math data is the master, drawing is for reference"

Ok, actually my real favourite comes from Japanese drawings that state (in English), "In the case of any discrepancy between the English and Japanese text on this drawing the Japanese shall take precedence"

Not being able to read Japanese, I've never found a discrepancy.

regards,

Hydroformer

 

[ewh]

ctopher,
That is not entirely true. We are involved in a military program providing parts for an unmanned aircraft, and we have no requirements for drawings. The files we receive from the customer (of parts we have to produce) have no drawings, and all tolerance callouts and related information is in the file with the solid model. We do not produce drawings of the finished part, as their files cover everything (build to "print").
Another customer of ours is producing a small corporate jet, and while they do required drawings, dimensions are kept to a minimum, mostly hole sizes, overall dimensions and geometric tolerances.
Prceedures, certs and related documentation are provided, but fully dimensioned drawings seem to be going the way of the buggy whip. The future is closer than you think!

 

[drawoh]

ewh,

Cars and airplanes are full of complex curved surfaces. Conventional definition and inspection with dimensions on paper is difficult. Working directly from the model in CNC or with SLL is fairly easy. Using the CAD data to generate inspection fixtures is easy too.

Most of my stuff is orthogonal, and fabricated by sheet metal and machine shops. Given our production levels, 2D dimension drawings are still convenient and efficient.

Curved stuff looks neat though, so I am getting interested in SLL and in castings.

JHG

 

[ewh]

drawoh,
I was suprised when we received some of these part packages. Not only are the formed and molded parts without drawings, but the orthogonal hardware is too, which made my job a little harder because I had to get some of that information on paper for our vendors. Sometimes I miss the days of square parts with round holes!