Siemens Tolerance Analysis Tools Reduce Cost & Development Time
Tolerance analysis continues to be put off to the end of the detail design phase within product design lifecycles, and sometimes just prior to production release. Why? I think the simplest answer is most designers simply do not like doing tolerance analysis. And why is that? Because most designers do not believe they are getting right answers. Even designers that have put together some elaborate spreadsheets have admitted to me that they are not really sure if the calculations are correct, and because the spreadsheets are so elaborate, the spreadsheets are difficult to use and especially difficult to re-use.
For years I did not like doing tolerance stack-ups either. I avoided spreadsheets because I know that the variety of circumstances in dimensioning and tolerancing was so complex, that I always took the stack-up on a measurement by measurement basis. Having what I thought was an excellent knowledge of GD&T, I always felt my answers were accurate but documenting the calculations in a consistent and re-usable format was very difficult to say the least. Then I discovered VSA, now known as VisVSA, an option in the Teamcenter Visualization product.
What I quickly discovered was that my answers did not always match VSA answers, especially regarding which features were the highest contributors. But I put some trust in the product and sure enough where the VSA analysis indicated issues, we had issues. Using a well-engineered software package for tolerance analysis suddenly made tolerance analysis an enjoyable challenge. Why? The answers were correct, providing greater benefit. The interface made doing the work “interesting” and ultimately much more re-usable. Compared to the effort of manually doing tolerance stack-ups, spreadsheet or not, with VSA it was like comparing 2D manual drafting to 3D solid modeling, or 2D annotation to 3D PMI.
Lets fast forward now to VisVSA. Coming on the market I think around 9 or 10 years ago, VisVSA is not all that new. But I am amazed of how few people really even heard of VisVSA, including Siemens customers. VisVSA is an outstanding 3D tolerance analysis package that has an interface that is much easier and enjoyable to use than the original VSA Classic product and it is CAD neutral using the JT file format. The 3D interface is much like being a basic CAD modeler except modeling the tolerance features. Going into the details of VisVSA is beyond my intent of this article. I just want potential users to know that the tool exists and that it is a mature and robust product. It has been through many updates and like any software application, bugs fixed and enhancements made. Let me highlight two most notable capabilities of VisVSA.
First, VisVSA evaluates datum shift correctly. Datum shift basically occurs when the datum at virtual condition rule is applied. With that said, there can be more than one virtual condition for a referenced datum feature. The key is to know which virtual condition is applicable to a specific analysis or measurement. Many good GD&T users have trouble with these issues. But I have personally evaluated VisVSA and with the exception of a minor bug that I did report, VisVSA correctly solves for the correct virtual condition. (The bug is a very simple work around and is planned to be fixed in the next release.)
Second, VisVSA uses a mathematical technology known as the Conjoin technology. This allows for best fit evaluation of seemingly over constrained feature assemblies. For example, most other tolerance analysis applications can not assemble a pattern to a mating pattern, at least not correctly. Typically, another system will assemble a single feature of a pattern to a mating feature and then another single feature of the pattern to a mating feature separately. These systems are often limited to the classic “3, 2, 1 move” for assembly. It is no wonder why so many designers establish two separate datum features, typically datum B and Datum C of a pattern where the entire pattern itself physically acts like a single datum feature and therefore should be referenced and analyzed as such. However, VisVSA understands the actual potential movement of the entire pattern simultaneously relative to the datum reference frame(s) and the feature to feature design requirement. In other words, VisVSA analyzes composite and dual single segment feature control frames the way the pattern features may actually work in assembly. This more realistic evaluation typically demonstrates that larger tolerance allowances may be applied to achieve the true functional design objectives and therefore potentially reduce many costs in manufacturing and inspection. In my experience, many designers avoid composite tolerancing simply because the tolerance stack-up analysis is more complex. But, composite tolerancing, along with many of the symbols of GD&T are there to provide all the functional tolerance that the design truly physically allows. But again, many symbols are left off simply because many designers don’t know how to evaluate them. The classic example is leaving the maximum material condition modifier off of datum feature references. Why? Because that allows potential datum shift (see paragraph above) and it is often difficult to evaluate by hand. If the people signing checks and purchase orders knew how much costs increase when modifiers are not used where they should be, they would put a stop to the “simple drawings”.
Now, fast forward to the present and get a picture of the evolution of tolerance analysis. Within the NX CAD system there is a tolerance analysis option called TSV or Tolerance Stack-up Validation. This is a somewhat more simplified lighter version of what VisVSA does, but it is integrated directly in the CAD system. There are major advantages for VisVSA to be CAD neutral using the JT file format, but the TSV product makes it very practical to perform early and consistent tolerance analysis during the earlier stages of design, including conceptual design. Also, quick impact evaluations can be done during detail design as actual data begins to be returned from prototype parts. What amazes me is how many NX CAD users are not at all familiar with the product or the fact that they very often already have it. As for the interface, the TSV menu is really only three new icons. The other icons are really just duplicates of existing assembly icons but packaged in a simple tool bar. TSV comes as part of an optional package to NX, but there is a good chance that if you have PMI, you have the TSV application. One might even think that PMI and TSV are somehow related, and they are.
Tolerance Stack-up Validation is not exactly new either. It works very well with the older embedded GD&T application. Siemens is working on developing the same sort of related functionality for PMI and has already done so with some success.
The icing on the cake is that TSV models from within NX can be directly exported into VisVSA where more sophisticated and detailed analysis can be performed. And, as of about 3 years ago, I found it to be very robust. The transition included the features, feature tolerances, assembly constraints in the correct sequence, and measurements. This effectively reduces the development of a VisVSA model by as much as 50%. Another “soft” benefit is that the early use of TSV forces better use of GD&T.
The reason my evaluation is 3 years old is because of a fundamental career change, I don’t use NX these days, although I am still a Siemens customer using Teamcenter Visualization for VisVSA and collaboration as well as Solid Edge. I am still in the field of dimensional management and still converse with designers and engineers, many of whom use NX.
In summary, my intent of this article is to highlight that Siemens has some great tolerance analysis tools that have been around a while and fit very well into a dimensional management lifecycle. The tools are robust and I know the development of having PMI more completely work with TSV, and TSV work more completely with VisVSA, continues. I can safely say that the investment to start using these tools now is a sound investment in the development of cutting edge product development lifecycles. And I can safely say that the payback in reduced costs, higher quality, and time to volume can easily compete with any “lean” programs I have ever seen. You can lean out a lot of processes if the tolerance requirements are accurately calculated, defined, and understood, and the earlier the better.
Norman Crawford
GD&T Instructor / Consultant
Applied Geometrics, Inc.
GD&T: Theory to Reality
740-505-2079
