manufacturing capabilities by feature type? 1

[danondabass]

Does anyone know of any resources anywhere showing the manufacturing capability* of (let's say) a sheet metal emboss, an injection-molded rib, etc.?

*In other words, how tight can tolerances be held for a particular type of feature in a diecast part, sheet metal parts, etc.

Thanks in advance.
Dan (on da bass!)

 

[KENAT]

Good one, and one I'd been planning on posting at some point.

I've looked on the Internet and you can sometimes find information for certain processes but I've not found it all in one document.

I'd love something like the standard surface finish chart (

http://icrank.com/cgi-bin/pageman/pageout.cgi?path=/data/surface_finish.htm&t=2)

for example, but for typical tolerances.

I assume you've already looked at Ctophers post from a few days ago.

KENAT, probably the least qualified checker you'll ever meet...

 

[wes616]

This is very highly dependent from shop to shop and machine to machine. For instance, I have 2 bridgeport mills in the tool room. I can often achieve tolerances of .0005 with one, whereas the other I'm lucky if I can hit .01. (both are manual).

Also, it is going to depend on what you define as tight tolerances. A .0005 tol may be tight for an engine valve, but may be loose as can be for a heart valve.

You really need to consider tol on a case by case basis and a shop to shop basis; and it's highly dependent on your product's requirements.

IMHO there is too much generalization that is attempted regarding tolerancing.

Wes C.
------------------------------
No trees were killed in the sending of this message, but a large number of electrons were terribly inconvenienced.

 

[wes616]

oh... and budget

Wes C.
------------------------------
No trees were killed in the sending of this message, but a large number of electrons were terribly inconvenienced.

 

[KENAT]

Wes, you are correct.

Ideally I design a part based on my experience/function and then the relevant process expert looks at it to suggest changes, including tolerance, that could reduce costs/ease manufacture etc.

Back when I worked in a place that had its own machine shop and reasonable quantities of most parts I did this. In fact for high production parts it was part of the design review.

This is more difficult when you only use external machine shops and are pressed for time. For some parts it's still appropriate to take the time but but in my situation I have a lot of different mostly relatively simple parts but with low production rates.

So the initial guess is gonna be what you go with a lot of the time and it would be nice to get some information to support the guess.

Sometimes the machine shops come back with suggestions, sometimes they just charge to meet your requirements.

We've also had machine shops take it too far, make multiple phone calls etc on how eliminating or changing a certain feature could save money only to find out the saving is only say $20 each on a total run of only 5, the amount of time spent on the phone & investigating the change etc probably cost more than this.



KENAT, probably the least qualified checker you'll ever meet...

 

[KENAT]

ISO 2768 kind of has this information.

http://www.ramo.se/iso_2768_english.htm

I'd caution against directly invoking this standard as frankly it's kind of lame in its entirety, especially some peoples interpretation of it. (thread1103-196260 thread1103-197786) What joebk points out on the first thread is a major concern.

However using the tables to guide you in tolerancing a part per ASME Y14.5M-1994 or equivalent may be useful.

KENAT, probably the least qualified checker you'll ever meet...

 

[KENAT]

http://www.gatewaycoalition.org/files/kinzel\Tolerance/NVtolerance.ppt

slide 6

Not by feature as such but may be useful. I've seen similar charts and I'm not sure how accurate they really are but may at least encourage people to try and relax tolerances where possible.

KENAT, probably the least qualified checker you'll ever meet...

 

[MadMango]

If you have the chance, ask your vendors what they can typically hold, it's a good first step. If you have internal capabilities, ask the supervisors. Everyone wants to do a good job, some just charge more for it.

(Off topic Dan, I'm learning the bass, week 5 and counting)

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

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

 

[MikeHalloran]

A former employer paid for all its designers and engineers to spend several days attending an onsite course, "Boothroyd- Dewhurst Design for Assembly/ Design for Manufacture" or something like that, including a room full of computers and access to the software package that actually contained process capability information for a huge assortment of machines and processes. The price included an ongoing license or two installed on our network.

The software helped you generate what amounted to a process sheet for each part and then the assembly. It generated a score for relative cost and manufacturability, and encouraged you to explore variations in methods, materials, etc.

It would have been a wonderful tool for someone just out of school, if you could convince said young person to pay attention to the tool.

The experienced guys mostly scoffed that it was all common sense, except of course where it disagreed with what they wanted to do, in which case it was obviously wrong. They didn't use it much, either.

It did effect one change. The software was very heavily biased toward snap fits. All of a sudden, there were snap fits everywhere, most in inappropriate places.

Of course it couldn't deal with individual machine variations like Wes pointed out, but if you were considering a process that you hadn't used (and it was embedded in the database), it could help you avoid a few beginner's mistakes.

I have no idea what the training cost. It might be cheaper to hire or rent someone with actual experience, and it would certainly be cheaper to buy a book.



Mike Halloran
Pembroke Pines, FL, USA