The Politics Of GD&T 2

[fsincox]

I am starting this thread to remove this discussion from kasabis's "TP & FCF Basics".
dingy2: I do not know if you are or are not the person I am referencing. I really did not mean to single you out, I disagree with the whole politics of “make it easier for them and they will come”.
The old system does not work. The new will not be followed until that fact is generally understood. The 2009 standard finally says it in the forward: ” …stronger admonition than in the past that the fully defined drawing should be dimensioned using GD&T with limit dimensioning reserved primarily for the size dimensions for features of size”.
Anyone who understands zero tolerancing at MMC or LMC should know we are not making the part harder to make. We are trying to provide manufacturing options. The fact the educators have not gotten this across and seem to pussy foot around management is what keeps us in the current state. It may not be popular, it may not be what management wants to here, sorry. I had hoped here I was speaking mostly to true believers and people who want to learn about GD&T not how to avoid it.
"Do not use perpendicularity on small land pilots", (because we can’t check it and it probabally made it anyway?) why shouldn’t I say it then, I don't care if they check it or not, we do not reall check every point on flatness either guys (and girls, I hope?). "Don’t specify finishes on 125 on greater" (because we will get it anyway?) How about in the future when almost everything is rapid manufactured except for fine machining. I think it is our job to state known requirements on the drawing so people don’t need to guess.
I believe, other than the validity of the design itself, it’s proper documentation is engineering’s second most important function, I mean no offence to anyone in particular but I do want to battle this whole culture of lets avoid GD&T.

 

[axym]

I'm not sure what to say about this. The thing that bothers me is the implication that tolerances should be based on the process that is going to be used to make the part. This goes against the "tolerance for function" philosophy.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[KENAT]

Agreed Axym. There is not point producing parts that are cheap to make but don't work. There is sometimes a point to making parts that are almost impossible (very expensive) to make but work.

Usually the optimum is somewhere between the 2 extremes. Sometimes parts get redesigned to better suit manufacturing capabilities etc. I'm all for making sure your required tolerances are realistic and within process capability etc.

However, if you don't worry about function first isn't there a worry you'll make useless parts?

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Standing]

drawoh
I will vote for you as dictator.
Well said, a star for you.

Standing
SolidWorks Pro 2009 x64, SP3.0, PDMWorks Workgroup, SolidWorks BOM,
HP xw8600, 64-bit Windows Vista Business, Service Pack 1
Intel Xeon CPU, 3.00 GHz, 16 GB RAM, Virtual memory 166682 MB, nVidia Quadro FX 4600

 

[drawoh]

I'm not sure what to say about this. The thing that bothers me is the implication that tolerances should be based on the process that is going to be used to make the part. This goes against the "tolerance for function" philosophy.

Theoretically, production should be able to take my drawings and send them out to any shop that suits their fancy, be it machine shop, sheet metal shop, foundry, blacksmith or anything else you can think of. The reality is that when I complete my drawings, I know how the part is going to be fabricated. I do not know how this works for you, but I would be very surprised if there was much flexibility.

If my requirement is to place a bend in my part to ±.002", a sheet metal shop cannot do it. The part will have to be machined. The bend radii will be irrelevant. The material will have to change from Aluminium_5052-H32 to Aluminium_6061-T6. Now, I can change some of the wall thicknesses.

The overwhelming probability is that you know what the fabrication procedure is. Your tolerances have to be feasible for this.

There is a very minor issue of you telling a machine shop to drill a 10mm hole and get the diameter to, say, within ±2mm, somehow. You can seriously clutter up a machining drawing in attempt to provide maximum allowance. It will be a complete waste of time.

JHG

 

[KENAT]

The thing is Drawoh, I've seen too many cases where the designer thought they knew how the part was to me made and created the drawing accordingly, when in fact there were other ways of making it some other discrepancy. Now you are likely better at knowing manufacturing processes than I or at least some of the folks that I've worked with. However, fundamentally as a Designer you care about the end item, not how you get there. Now that doesn't mean you give no thought to the how, but that as much as possible you don't tie yourself into a specific how unless functionally required. ASME Y14.5M-1994 1.4e lays this principle if I recall correctly.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[marshell]

"fundamentally as a Designer you care about the end item, not how you get there."

This is a statement that I have heard, used, and been perplexed about for years.

On one hand you dont really care if they carve the metal from a billet, using a toothpick and butterknife as long as it meets the drawing. When we spec a thread, we dont care if its roll tapped or tap and drilled.

On the other hand, you do know how they are going to build it. If we have a 4'x4'x 1/2" thick plate with a 1/2" high boss in the middle, do we show it that way or do we show a sheared plate, with the boss welded on? We knew how it was to be made. If later its decided machining would be better, we would make a product or cost improvement change. If we went a step further, why even spec a specific material, why not just the desired range of properties it has to meet.

When we open our cadd program, we have to decided... hmmm, assembly, part, weldment or sheet metal. We have an idea of what production will do.

Though ASME Y14.5 does not want you to specify specific processes, such as tap drill size, etc. They do supply tools that apply more to one process than another. If you are machining you would specify primary, secondary and tertiary planes to help the machinist. The same part, if molded or cast may contain datum points. So in a sense the process has been defined.

Thank you

Marshall
Marshell Design

http://sites.google.com/site/marshelldesign/

 

[fcsuper]

Marshall,

I'm not sure you are actually contradicting the sentiment of the statement you question.

Matt Lorono
Lorono's SolidWorks Resources & SolidWorks Legion

 

[fsincox]

I think the standard is pretty clear on intent: the ideal is no process, but, sometimes it may be necessary. As is usual, the application of this principal is the unstated/unclear part. I am not sure what more they can do.

 

[axym]

I once heard this idea summed up in the following way:

"The drawing should be a contract for inspection, not a recipe for manufacture"

Unfortunately, in many companies, the operation would grind to a halt if designers actually followed this.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[KENAT]

Marshell, take that statement in context and I don't think it's that perplexing at all.

There are shades of gray/grey. For very high volume stuff or difficult to manufacture parts or some other situations, it can really pay dividends to go through the manufacturing process and taking great care to be sure to make the drawing suit it. However even then you want to avoid tying down anything to a specific process unnecessarily incase someone else has a bright idea on how to improve run rates etc. You don’t want to be changing drawings every time someone on the shop floor or in manufacturing has a bright idea, do you?

For a relatively low volume, not too complex part, you can probably waste more money spending time fully optimizing the part for production etc. than you save.

It’s not that you pay no attention to how a part is made as a designer, it’s that you want to give the manufacturers the maximum freedom to manufacture it however best suits them while meeting your functional requirements.

Axyms quote reflects the prevalent attitude at least in the US & UK from what I’ve seen. Then again, maybe there’s a link to the perceived decline of manufacturing in those countries;-).

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[drawoh]

I once heard this idea summed up in the following way:

"The drawing should be a contract for inspection, not a recipe for manufacture"

I agree with this. Manufacturing procedures should not be described on the drawing. The fact remains that at completion of design, you know how your part is going to be fabricated. Your tolerances should reflect this.

Example:

You have just designed the new super-duper handy-dandy widget. You have designed the chassis as a part CNC machined from billet.

You make one and then a couple more, and everyone is happy with it. Your customer says "Cool! Now make another 10,000."

Now, you want to send the chassis out for investment casting. You have options.

[ol]
[li]Send your current drawings out as is. The foundry figures out the tooling and subsequent machining procedures needed to reproduce the machined part.[/li]
[li]Go back and review your design in context of a casting. Casting is less accurate than machining, and produces an inferior surface finish. Casting quickly and easily produces weird, complex shapes, allowing you to integrate components (DFMA), and make the thing look cool. If you can eliminate all the accurate tolerances, you eliminate the machining step and you save serious money.[/li]
[/ol]

None of this affects the need for GD&T or standards compliance. It does affect the organization of your drawing package, and the configuration and tolerances on your drawings.

JHG

 

[axym]

drawoh,

Your story involves designing a part with a certain fabrication process in mind, and basing the tolerances on what that process is capable of.

The intent of the statements in Y14.5, and (what I believe to be) the intent of the quote I gave, is the opposite of that.

The idea is that at the completion of design, you know what tolerances are required for the function of the part. The process chosen to fabricate the part should reflect this.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[drawoh]

axym,

It is probably not possible replace a well designed, functioning machined part with an economical casting. You will have to redesign your product to work around the limitations of the technology, and to take advantage of it.

Example:

In a message above, someone mentioned a boss that might be machined in place, or added by welding. Machining the boss in place means paying for a large, expensive billet, and machining away most of the material. Welding the boss in place is not completely simple with mild steel, and quite a bit more complicated with aluminium. Casting the boss in place is dead easy, as is hollowing it out from the opposite side. You no longer pay for the extra part, or for the fasteners, or for the assembly procedure.

There is no problem drilling and tapping holes in a machine shop so that something can be assembled using small machine screws. On a casting, I would look for a way to lose this process. A solution would be fewer, larger fasteners, with some sort of retained nut.

If we go back and redesign the product around a welded sheet metal chassis, again we have a series of challenges and opportunities that result in re-design.

JHG

 

[MechNorth]

Design work shouldn't be perfomed in a vacuum. Hopefully the designer has some growing knowledge of manufacturing or supporting infrastructure (mfg engineering, etc.) to call upon. Function should always be the primary consideration, but not the exclusive factor. Reality is that manufacturing and inspection processes ARE kept in mind at least in a broad categorical sense (i.e. this part is thin sheet metal, therefore likely to be stamped, so I'm likely to need loose tolerances on most features ... though I can tighten up as needed, vs this part is a moderately complex form of varying thicknesses and therefore less appropriate for sheet metal stamping; whether it is machined from block, from forged or cast material is irrelevant as long as it works as needed).
The engineering print should be a complete part specification; material, finish, feature sizes and tolerances, without having to reference another document to determine a nominal size or tolerance (my soapbox item) such as referencing ISO 2768.
The argument that GD&T isn't needed because a manufacturing process is in place to make sure you get good (functional) parts regardless of the drawing is biting a lot of nay-sayers in the butt these days as their companies move further towards outsourcing (domestically or offshore) and the quality "miraculously" drops. If they had a product specification which was complete and accurate for the part functionality then tribal knowledge would not be a crippling factor.
That a startling number of companies pay only lip service to GD&T is not the fault of trainers, it is due to what I call "management by magazine" mentality. Ever come across a manager who reads an article & decides to implement the changes he just read about ... of course his people are brighter & faster & more professional and dedicated, so they can do it in 1/10th the time and 1/100th the cost ... ? That's management by magazine. I enjoy having engineering managers sitting in the class, and putting them on the spot just like everyone else in class. If they're good leaders, their people already know that they are not gods and they relax and enjoy the class. If they're bureaucratic managers, they at least get a good view of the efforts being made by their people and how much there is to absorb and process.
On the item of issuing only what a shop can handle as far as drawing content is concerned, I offer this thought ... who here would like to write using Egyptian hyroglyphics instead of the modern English language? Things evolve as a fact of nature, or else they become extinct. Manufacturing businesses, shops, engineers (designers & draftspeople too), machinists and inspectors that can't or won't evolve will drive their own extinction.

Btw, interesting debate. Tks for restarting it!

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com