stupid question ... but which GD&T can I reasonably ask or GD&T guidelines 1

[chedalb]

Hello Everybody,

the question is stupid, the answer is not.

I am working for an market where basically it never happens to have the same part twice. Depending on the projects, it happens that we act as procurement office. When dealing with a new part to be manufactured and designed by a third party, we are often confronted to the workshops saying "this is not doable" or "not attainable".
Then I have to go back to the partner company that designed and start a trio discussion "workshop"-"designer company"-"procurement company".

So, if we would be able to go back to the designer company before the workshop, we could shorten the preparation of the offer preparation and order preparation.
So, the matter is, the part that we have to manufacture today is different from that for which we asked yesterday, and what learned yesterday is possibly of little use today.
I have been wandering around on the web, but apparently no one is stating "which tolerance is attainable when... by working this area extension ... with this equipment ".
It is very understood that this is the typical "trade secret" and experience that people close to retirement are full of and newbies need to learn from scratch..

I also have been looking for books, but i am running under the impression that they always provide the indication on how to use a tol, but not what what is reasonable to ask or what are the limits...

So this is the question..
Are you aware of any GD&T guidelines, providing indication of what is attainable under which conditions?
Thanks in advance,
chedalb

 

[KENAT]

So you're looking for some guide on process capability, what processes & 'special circumstances' etc.?

Variations on this question come up a lot, searching the site might get you something of use for example:

https://en.wikipedia.org/wiki/Process_capability

thread281-290732
thread404-284781
thread404-278305
thread404-197410

However, experience is probably the real answer.

Posting guidelines faq731-376

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

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

 

[chedalb]

Thanks KENAT,

maybe also due to my non native english, I wandered for some time on the forum without solution..
thanks for spotting out!
chedalb
.

 

[JNieman]

If you are asking what tolerances are reasonable for certain processes (CNC Milling, manual turning, cylindrical grinding, or sawing, for example) then there is no guidebook. CNC Mill machinists won't even agree. You ask one machinist and he'll tell you a 12" part shouldn't have tighter than +/- .005" on lengths and sizes. Another will tell you +/- .001 is just fine all day long.

The knowledge of such things really comes from experience, I hate to say. I don't know another way to "teach" what process is required for what tolerance (within reason).

I also feel like I should make the point that the process does not dictate the tolerance. Your tolerance dictates the process. Just because you want "bandsaw prices" doesn't mean that previously CNC Milled part should have the tolerances loosened. The designer shouldn't assign tolerances tighter than necessary, either.

Without more details, examples, or context, I find it difficult to give more specific advice.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

 

[KENAT]

You could look at something like ISO 2768 for guidance as it was presumably based on real world data.

However, that real world data is now probably several decades old so may not be as applicable as once was. Also you're still stuck trying to work out what the different grades imply on cost etc.

Posting guidelines faq731-376

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

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

 

[TWJR]

I also feel like I should make the point that the process does not dictate the tolerance. Your tolerance dictates the process.

Couldn't have said it better myself. Tolerance the part by function. This will dictate how it's to be done. As far as CNC mill tolerances, for example, there are too many variables for a one-size-fits-all answer. What kind of machine? How old? How well maintained? Skill of machinist? Quality of tools available? Type of material? Etc, etc.

 

[KENAT]

I 100% agree about function is the primary driver but...

In a cost driven world having some idea of typical process capability and ensuring your design falls in that isn't such a bad idea. I've redesigned parts/assemblies that required very tight tolerances that pushed process capability for the sake of reducing cost. After all, that's part of DFM/A after all.

Posting guidelines faq731-376

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

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

 

[CheckerHater]

There is nothing wrong about actually knowing if your ambitious requirement can be met by real people on this planet

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[drawoh]

chedalb,

My understanding is that someone is designing stuff and preparing drawings, and you are sending these out to the shop. You need to find shops that can work with your drawings, or you need your designers to understand fabrication processes and clear documentation. Welcome to the wonderful world of office politics!

It sounds like you need your designers to specify realistic tolerances, and come up with designs that work with those tolerances. How much authority do you have over them?



--
JHG

 

[3DDave]

There's no good guide because that would be anti-competitive. Imagine if every company had their cost vs capability exposed. No chance for them to pad bids, plus the liars would make stuff up to get the business and everyone would quickly be back to the current situation.

A great guide would tell what range of variation is possible for particular processes and what the tooling, processing, inspection, and assembly cost is for each sort of feature one might think up.

So, it might be cheap skip tooling, then slap in a crappy hole somewhere near the desired location, verify it's a crappy hole, but then cost a lot because the assembler needs to juggle a bunch of parts or shims to get the machine to work right. It's how costs average out.

The suggestion I can make is to try to find items that are built in similar quantities to those you expect, of similar materials, and look at the processes used to make those parts. Then you can talk with the particular suppliers of those processes.

Sadly, there's no useful index.

 

[CheckerHater]

ANSI B4.2 has some basic information you may find interesting.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[JNieman]

For reference, look at this chart, which seems to get spread around very much, at the bottom of this post:

Now, when I was still in school and learning, I thought this was neat. Finally, some of these questions I had that got wishy-washy answers of "it depends" had a strict answer! This chart was a great example of how intelligent people can document black-and-white expectations of manufacturing!

Then I got experience and learned that this chart is mostly worthless. Yes, it is mostly good but there is a whole lot of "it depends". Also notice how VERY VERY BROADLY these rows span. Boring/Turning, for instance, spans from 1000 u-in to 1 u-in. That basically tells you nothing. You can narrow down that process "anything can happen" basically.

Anyways - my point is: when you try to take a big question and reduce it down to a simple "go-by" - it becomes generic, bland, and nearly useless.

https://upload.wikimedia.org/wikipe...Surface_Finish_Tolerances_In_Manfacturing.png

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5