Tolerance Chains & Tolerance Stack-up 4

[steedj]

Does anyone know where I can find information on Tolerance Chains, Tolerance Stack-up, etc. I've searched the web and found very little.
I'd love to find a website link but any information whatsoever is greatley appreciated.
Thank You very much and Merry Christmas!

 

[zzy]

check out

http://www.ci.com

for a software maker named Cognition Corp., which is specialized in a tolerance stackup software -- MECHANICAL ADVANTAGE.

 

[Viktor]

This program, unfortunately, has nothing to do with the tolerance chains.
Unless you know Russian, I do not know any book and website related to the topic.
Viktor

http://viktorastakhov.tripod.com

 

[swertel]

What about tolerance chains and stack-ups are you trying to find out?

 

[steedj]

I have never done a Tolerance Stack up before; all I have is a very basic idea as to the concept and I am looking for any information that can help. Thank You.

 

[swertel]

Tolerance stacks are common sense issues to determine if all the parts of your mechanism fit together, allowing a functional mechanism without being overly costly (too tight of tolerance on parts). If you can do simple arithmatic, then you can do tolerance stacks.

I have found, though, that many engineers and designers can not do this simple arithmatic. Someone who can do 3rd grade math is portraid as the department guru.

All tolerance analysis needs is time, a trusty calculator, and a focus on what the final (functional) design requirements are.

As zzy stated above, there are computer programs that do tolerance stack-up, but for the most part this technology in relation to CAD is in its infancy and rather cumbersome to use. You are better of doing it long hand.

If you have a specific example that can be described here, I'd be happy to walk you through one.

--Scott

 

[steedj]

Thanks Scott.
I appreciate your input.

I would love to send you a copy of the assembly I am required to do a tolerance Analysis on but I am not permitted to do so -- my company has confidentiality concerns.

However, I can attempt to explain my situation a little. I am a designer for a German owned company located here in the states. We are nearing the completion of a major electro-mechanical project; however, a lot of Assembly interface problems have cropped up. As a result, we are now required to do Tolerance studies on all our Assemblies. I have only worked here a couple of years and I was not part of the original development team. I am told that some sorts of Tolerance studies were done in the past but not in a method that would satisfy the German owners. Our German friends seem to be more adept at these Tolerance Stack-ups than we are.

We have very small and intricate parts that make up some of these assemblies. Some of the assemblies have parts with multiple interface points and some of the parts rotate as well I would like to find any info on how to do these Stack-ups correctly. Do you know if there are any standards, and/or rules governing these Tolerance Stack-ups? If I had something to use as reference or some examples to go by then maybe I could see exacly how to do these; anything would be helpful and welcome.

Thanks again.

 

[boo1]

I assume you are trying to refer to is tolerance accumulation. Pull out ANSI Y14.5 it will provide you with the information you need. But then you will need to "do the math" as Swertel indicated.

Cheers

 

[Viktor]

A calculator and a pen are good tools but of little help in real analyses of dimensional staking.

Even for a simple calculations one should know about tolerance fields, basic dimensions, nominal size, MMC, unilateral and bilateral tolerances etc. Moreover, one should be well aware about a number of statistical distributions occurred in manufacturing. On the top of that, it is desirable to know the assembly and working temperatures of the unit, properties of materials of parts included into calculations.

When one wants to do it for real, he/she should complete the analysis by the optimization of tolerances accounting for many real-life factors. The result is truly amazing and rewarding for those who know what this analysis and optimization is all about.
Viktor

http://viktorastakhov.tripod.com

 

[borjame]

I would also suggest getting your hands on the dimensional standards for your tolerances, if you interpert a location tolerance different than your machine operator, all your hard work will be useless. Some companies also use root sum square (RSS) tolerances which is a statistical method that assumes the odds of all the tolerances going "one way" are slim.

 

[Viktor]

Try to look at StackSoft

http://www.uts.com/products/ssdoc.html

Looks like a nice toy.
Viktor

http://viktorastakhov.tripod.com

 

[GregLocock]

For what it is worth there are companies in the USA that specialise in tolerance studies of assemblies. For a complete suspension module the bill was around $30k, as I remember.

Nasty questions you need to ask to do it properly: is each tolerance normally distributed? are any tolernaces cross-coupled during the manufacturing process? what is/are my measure/s that define a good assembly?

A good example of tolerance control is the BMW X5s engine/gearbox/transfer case/diff assembly.

The driveshaft needs to be in line with the output gear of the transfer case, and the pinion nose of the diff. The diff is bolted to the sump which is bolted to the block which is bolted to the bellhousing which is bolted to the transmission which is bolted to the transfer case. How do you maintain angular and positional alignment across all those interfaces? I wouldn't mind betting the maximum permissible angular error is a couple of degrees, and maybe 2 mm in positional error as well.



Cheers

Greg Locock

 

[Viktor]

Greg is completely right! Viktor

http://viktorastakhov.tripod.com

 

[roseda]

borjame, could you (or someone else) give a simple example of how to use RSS.

 

[gdtdave]

There are several programs that can do a good job.

VSA, currently from EDS.
Valisys, which was developed at the same time as VSA, but is separate.
Technomatix comes to mind as the supplier.)
3DCS from Dimensional Control Systems
CE-TOL from Sigmetrix (via ADCATS) (Pro/Engineer only)

The trouble with 'hand' calcs is that without a lot of practice and some accurate feedback, the analyst can convince himself of most anything. This is especially true when dealing with datum reference frames. Just as few people can write a computer program that runs right the first time, few people really understand the results that will be allowed for all but the most trivial dimensioning schemes. What they -think- will happen is what is analyzed, but not what is -allowed- to happen by the drawing requirements.

VSA was not a trivial program to use, but the interface has likely improved. It was more lifelike than CE-TOL (formerly TI-TOL) seemed to be, though arguments can be made on either side for particular facets of dimensional and tolerance. investigation. If I had only one? VSA.

 

[Tunalover]

ASME (

http://www.asme.org)

has at least one good book on tolerance stacking. Seeing how they are also the keepers of ASME Y15.5M-1994, I think it would be worthwhile taking a look. As it goes with their typically formal publications, there may be other books on the market that are shorter on the formalities and longer on practical problems.

As for tolerance chains, assuming you are referring to chain dimensioning with bilateral tolerancing, the problem is completely eliminated with the use of positional tolerancing per Y14.5M.

As others have indicated, a full tolerance analysis, even with a computer program, can be quite involved and expensive.

Bruce

 

[fastech]

As many have advised you, a true complete tolerance study is neither easy nor cheap. A recent one I was involved in was over $50k

VSA from EDS was used. They sent their own people in. I recommend it highly.

Merry Christmas & Happy New Year.

fastech
csft, inc. Bill P.
CSFT, Inc.

 

[gearguru]

VSA is good and as said before, expensive. It is also not easy to use. It works with models created in UG, or they can be made in VSA.
If there is only one assembly which needs to be evaluated, a good spreadsheet and a thorough knowledge of what you calculate are sufficient.
The second part - the knowledge - is the most important. It is always necessary, also when working with any commercial software.
Merry Christmas !
gearguru

 

[diamondjim]

It seems to me that the manufacturer of the
chains should be responsible enough to provide
what appears to be a simple problem at least
to the tolerance ranges. As to stack up of
tolerances, I would assume that you would be
interested only on how is affects the sprockets
and not what is happening on the free chain.
The chain manufacturer must also know the tolerances
that allowed on the the sprocket for this to
work. I would work with them as they have more
knowledge about their product than others.

 

[steedj]

Diamondjim,
I aplogize for not being more specific. In my original question ("where I can find information on Tolerance Chains, Tolerance Stack-up, etc..&quot, when I refered to Tolerance Chains, I meant "Dimensional Chains" -- not "Chains" in the literal sense. Sorry for the confusion.

Thanks for your input,
Steedj