Showing results for 
Search instead for 
Did you mean: 

Re: Best practices for PMI MBD


For Best Practices, one of the biggest question I have is, are companies using basic dimensions?  Per the Standard all dimensions that are queried off of the 3D model are basic.  For Hole locations and such are you using basic dimensions?

Re: Best practices for PMI MBD


Dave, I'm also working for a team trying to assimilate all our drawings from 2D drafting to 3D MBD / PMI and PDF's. Is there an email list to be added to or a webpage I can visit to stay in the loop with any updates?

Re: Best practices for PMI MBD


Hi, I typically observe this discussion and infrequently post a reply, then only when I feel I have something worth contributing.


Regarding the "All dimensions are basic in the CAD model" thought process that is being touted by some in the industry. According to “the” standard this is stated in, this might be true. However, does this constitute good communications to the folks who need to determine what exactly to produce? If we use 3D Printing exclusively, this might be okay. Showing dimensions (or other requirements) or not within the product definition is a decision each company must make for themselves.


Not showing all requirements in “man-readable form” is well and good if computer systems are the only recipients of the technical data. And therefore, it is those computer systems who operate on the data (making CNC, CMM programs, creating every dimensional shape within the model and then verifying QC requirements of the resulting products and documenting those results (for legal purposes).


I have been working on the MBE/MBD processes since the beginning of US Army MANTECH involvement and I have resigned myself to not arguing (too much) about what the right way is and isn’t. I believe in the end the right processes will float to the top and all will be well.


 I would just request that you consider asking the people who need to use this data what they need to make the products being designed.

Ask the Manufacturing and Quality Engineers what they need to fulfil shop floor production requirements.


While working with a “Lean Six-sigma” expert (my manager) eons ago. He said to me; “Treat every shop floor tradesman like a doctor. Keep everything needed to complete the job within arm’s reach. Just like a doctor holds out his hand and a tool is placed in it, we need to insure that every operator, fabricator, assembler, ect. has the opportunity to complete his or her job with the greatest efficiency possible.”


I also believe that this thought goes for the “Product Definition” documentation as well. It doesn’t matter if it is 2D drawings or 3D MBD, the documentation must be correct and complete. If you allow a milling machine operator to measure dimensions in a CAD model before an operation is setup, how lean is that process? All requirements should be readily visible when a human is making a dimension within a product.


Rich Eckenrode

Re: Best practices for PMI MBD


Hi @sdeters,

Yes, we are using basic dimensions.

Re: Best practices for PMI MBD


Crikey folks, this is an interesting post, and one I have been following with interest.


It highlights some of the pitfalls and issues with using MBD, what it is intended for, how it should be used and how it is actually being used.


I was tasked about 18 months ago to investigate the implementation of PMI at my company, and very quickly came across loads of issues with trying to use PMI to replicate a 2D drawing. This is not the intention of PMI.


We have got the ability to run a model from design, through manufacture, to inspection. Without ever producing a drawing. However we have only done this as a trial, my advise was not to proceed for full manufacture as there are too many problems with application, too many workarounds and not enough understanding of what the benefit is.


Let me try to explain why I recommended this:-


  • What is the perceived benefit?

Well, the idea of using PMI for ‘man readable’ purposes is common, but incorrect. It is purely to allow for Tolerance Analysis and CMM to understand the intended application of a component, anything other than this is very hard to achieve and causes massive issues with downstream electronic consumption.


  • The mind-set for using PMI is inaccurate in most understanding

With PMI, it must be made clear what the mind-set needs to be. In order to do this it must be realised that PMI is not used for human consumption, but for electronic consumption.

With this in mind, the software knows what size features are in a 3D model, it has to as it has already been specified! The only thing required to make this usable is the Datums, Geometric tolerances and any specific linear tolerances.

Any TEDs, Overall dimensions etc. would be over constraining.


  • This seems sensible, what are the problems?

This is harder to explain, the application of PMI in the model with the correct mind-set is fairly simple, but there are limitations within NX, such as a tolerance radius can’t be specified without first creating a section view. A theoretical centre datum through a block can’t be identified. Tapped holes are also difficult to specify.


  • What about geometry that can’t be inspected by CMM?

This has been bought up with Siemens, if there is geometry that can’t be CMMd, then there must be a way of specifying this on the model. The problem comes when the software tries to read this, and fails because it is ‘over constrained’ even though it must be a manual inspection, such as internal ‘o’ring grooves.


  • Can a drawing be produced using PMI

In theory…Yes!! In reality, it takes significantly longer to do this and you end up in the same place, in order to produce a drawing, over specified PMI must be applied, meaning it can’t be read electronically without causing errors. If a mix of PMI and traditional dimensions are used, it gets really messy on the drawing, as the PMI dimensions will ‘move about’ when models are updated and traditional dimensions wont.

Advice currently is ‘if a drawing is required, it is better to produce it in the specific application’

  • Site implications

Before PMI can be adopted, businesses must be prepared for it. There will be no drawings to check, microfilm, release or print.


Having also spoken to both BSI and ASME experts, I am also aware that there is potential for massive re-work when the fundamental understanding of 'real world' model based definition has been applied.

In answer to the original question, There is no 'relevent' best practice for the application of PMI because what it is intended for vs the real world application are quite different.


Re: Best practices for PMI MBD

Thank you very much for sharing your experience.

Re: Best practices for PMI MBD


Currently we are not using dimensions queried from the model. Until now some of our manufacturing guys build their own models with mid tolerance values, which is clearly not the preffered way of wrking, but until now it was the only way for them to use CAD models for CAM.

When it comes to queried dimensions, you have to be careful. We still have segnificantly less profile tolerances compared to other tolerances. A lot of tolerances have developed over time with internal manufacturing, QM, procurement and suppliers. It takes some effort and time to make a shift for at least some parts and then use dimension querying.


In addition it is very usefull to have 3D scans at QM when using dimension querying. Otherwise QM might get a little angry with you for checking profile tolerances on complex geometry.


Benefits do not seem to outweigh the effort for us at the moment.
Always make sure you take the whole process chain into consideration and speak with all steakholders to find a solutions.

Product Designer
NX 11

Re: Best practices for PMI MBD

Gears Esteemed Contributor Gears Esteemed Contributor
Gears Esteemed Contributor

@sholz wrote:

Currently we are not using dimensions queried from the model. Until now some of our manufacturing guys build their own models with mid tolerance values, which is clearly not the preffered way of wrking, but until now it was the only way for them to use CAD models for CAM.

You can use assemblies & WAVE linking (or promotions) and synchronous modeling tools to modify the CAD model to "fix" them to mid tolerance for the CAM model.

Good points:

- CAM model associative to the original CAD model (note this works best if the CAD model is also in NX, and the designers modify features, rather than delete & re-create, geometry)

- The list of features in the CAM model show you EXACLY what has (or hasn't) been modified


Downside is now you need to control the information flow to make sure the CAM model points at the CORRECT CAD model (PDM systems, e.g. Teamcenter, can help with this).

Ken Akerboom Sr CAx Systems Engr, Moog, Inc.
Production: NX10.0.3.5 MP16/TC11.2
I'd rather be e-steamed than e-diseaseled


This group is dedicated to the topics of PMI (product and manufacturing information) and MBD (model-based definition) in NX.

If you have a question, please post it in the NX Design Forum.

We also recommend you be a PLM World member, as this group supplements an annual roundtable at PLM Connection.

Members (278)