Stop whining about MBD and accurately model your products

Why is it so important for 3D data to be exact? After all, drawings on a drafting board weren't a mathematically accurate representation of designs and they got the job done just fine. Right? Well sort of. We’ve all heard the stories of someone faking in a dimension to show the right value, causing the 2D drawing to become a less accurate representation of the final product and out of sync with the 3D model (if there was one). The results of these out of sync documents usually amount to an unhappy boss.


We have to remember that when 2D drawings are used as manufacturing instructions, the human reading the drawing is able to extract the information and intent that is being communicated by the CAD designer because the human brain has incredible inference capabilities. Even though the drawing might show details that aren’t at exact scale, the reader is able to figure it out. The reader may even be able to see the remnants of the original penciled in dimensions, which might clue them in that something had been fudged.

Today drawings are simply less accurate presentations of the 3D model and it’s associated data elements.

When you intend your product designs to be reused or consumed by another software application, it is critically important that your 3D data be exact. If for no other reason, consider all of the downstream software that may consume this critical product data: analysis software (FEM and FMEA), Computer-Aided Manufacturing (CAM) applications, automated checking applications, automated tolerance analysis, or even other departments or suppliers that rely on the accuracy of the information.

Unless you’ve been living under a CAD rock the past few years, you’ve probably heard of Model Based Definition (MBD)

The key advantage of MBD is the ability to reuse data already entered in 3D, and to consume that data in an automated fashion. Essentially, this documentation strategy cuts out “human error” due to manual reentry. Having a truly model-only product documentation strategy is the “holy grail” of documentation; having one place where all of the information needed to design, manufacture, and service a product is stored.

MBD used to be a major issue for the engineering workforce, but is now starting to become more readily embraced as it finds its way with a younger, more digitally inclined workforce.

The initial perception of MBD implementation can be that this is more work for the designer who is no longer allowed to “fudge” a dimension on a drawing to make a quick change. It is perceived as having to jump through a bunch of hoops for no good reason, but that’s a short-sighted view. MBD can benefit your company, but doing it the “old way” will not work. As adoption of MBD increases and standards evolve, the realization is hitting home that the benefits do indeed far outweigh the investment.

Isn’t this article about checking your CAD data for accuracy?

Well, the two really go hand-in-hand. In order to survive in the landscape of just in time (JIT) manufacturing and in order to avoid the time and cost required to create a physical prototype, your data absolutely needs to be accurate. Hands down.

It brings to mind one of my favorite quotes from business management: “If you think it’s too expensive to hire a professional to do the job,wait until you hire an amateur.”

Of course, that can be reworded to suit this discussion and will still hold true:

If it’s too expensive to accurately model your product, wait until you come across a discrepancy in the data.

 Maybe you’ve been the person on the receiving end who came across that discrepancy and can really appreciate the point here. We’ve all got a horror story or two to tell of our experience in the trenches.

So if accuracy is expensive, and inaccuracy is potentially even more expensive, you really need to find a way to determine if the 3D data is “good enough.” I say "good enough" because “good” and “a complete model” mean different things to different industries, or possibly even different companies within the same industry.

There are several tools available today that you can employ to check aspects of your MBD data

Here are a couple of examples:

  • Geometry Validation Tools are available to compare native CAD files against other native CAD files or data that has been output to another format. These tools serve as a way to ensure that the translation or passing of the data hasn’t changed the data in the process. A quick web search returns a handful of companies that specialize in this area.
  • Neutral formats and standards committees

Let's face it, drawings are boring (maybe they earn you some old school bragging rights). But add in a 3D rotatable model, and see your users light up. It's a fact that humans can conceptualize better in 3D and when humans are excited about something, they will perform at a higher level and dig deeper. Give it a try – strut your 3D model down the virtual runway to see what sort of attention it will capture – just be sure to check its accuracy first.



  • Bob McGill

    A major goal of Model Based Definition (MBD) is to extend the geometric description to include manufacturing related information such as tolerances, material and surface finish which historically has been placed on the drawing. The ultimate goal is to eliminate the need for a human in the loop to interpret the drawing.

    Sometimes referred to as Product and Manufacturing Information (PMI), MBD faces a Chicken and Egg conundrum: why should a CAD user add PMI to the model if downstream applications like Computer Aided Manufacturing (CAM) don’t take advantage of it, and why should a CAM vendor bother investing in the support of PMI if so few CAD users are adding PMI to their models.

    The US DOD is pushing ahead with MBD, but most CAD-based MBD tools still lack the elegance and clarity of a good drawing. The good news is that for many parts, +/- .005” is good enough and toolpath can be generated directly from the 3D model without the need for drawings or embedded PMI.

    • That last paragraph is very interesting, the thought that tolerances and by definition quality, doesn’t need to be fixed but we should always try to find ways to make parts to the tightest tolerance possible. It’s true this only seems feasible with an all-digital approach.

  • Jimbo

    The cheapo CAD application my employer bought me to use, can’t make an accurate representation of a screw thread, without bogging down my machine. So what has the CAD vendor suggested? Create a simplified representation of the thread (lacking a helix, and with a simplified thread profile). Now what happens when a job shop uses my 3D model to cut the part, without referring to the 2D drawing? Are their systems going to realize the hole is threaded, or are they going to think they’re dealing with a hole with ridges? (i.e. no helix)

    I’ve already had cases where a sheet metal fabricator completely ignored the 2D drawing that accompanied the 3D STEP model. Thinking the fabricator was misunderstanding the intricate details in my 2D drawing, I kept revising it to make it more understandable. After three production orders that didn’t come out right (we allowed exceptions), I realized they weren’t even referring to the 2D drawing to make the part.

    How are we going to specify tolerances, surface finishes, and heat treating on a 3D model? This whole concept, (MBD) looks like something a freshly minted “engineer” would come up with. New grads think that just because something is built once, they don’t have to bother with tolerances and specifications. That’s not engineering. It’s lazy inventing.

  • MBD is founded on keeping the GD&T (Geometric Dimensioning and Tolerancing); as it is the math behind defining parts so that they will always fit together. Drawing-only product definition is limited because the GD&T is not machine readable (semantic or representative). When MBD is used with efficiency, then all the design intent data is digitally captured and can be consumed and re-used downstream, or over and over again with exact repeatability.