Design Engineering

CAD Report: Crossing the Rubicon

By Ralph Grabowski   

CAD/CAM/CAE Autodesk

Since its inception, Autodesk Inventor has laid down the gauntlet in the mid-range 3D CAD space. Here’s how Inventor 2018 stacks up. 

Autodesk began work on Inventor as a reaction to the run-away success of Solidworks. Self-funded by its programmers, this brand-new, mid-priced (roughly $5,000) parametric MCAD program seemed to “come out of nowhere.” In a couple of years, Solidworks had become the best selling Windows-based MCAD program, undercutting the high-end ($10,000+) market of Pro/Engineer, UGS and Catia. To this day, it remains such.

At the time, Autodesk offered two AutoCAD-based packages to MCAD designers – Mechanical Desktop (MDT) for 3D modeling and AutoCAD Mechanical for 2D detailing. The AutoCAD engine lacked MCAD necessities like constraints and assemblies and, in any case, it made MDT a hard sell to non-AutoCAD users.

To compete with the Solidworks juggernaut, Autodesk first tried buying them, but lost to Dassault Systemes. So, instead, it launched “Project Rubicon” – a new MCAD system written from scratch.

The word Rubicon refers the shallow river that, in Ancient Rome, marked the northern border of Italy proper. It was made famous when Julius Caesar, and the army he led, crossed it in 49BC,  and thereby broke Roman law and irrevocably began a civil war that led to him becoming dictator of the empire.


Figure 1: Defining the length of partial fillets in Inventor 2018

Since then, “crossing the Rubicon” has become an idiom for committing to a course of action past the point of no return. In the same way, Autodesk committed to challenging Solidworks’ dominance of the mid-priced MCAD market by launching Project Rubicon (now known as Inventor) in 1999. The launch was followed by an aggressive twice-yearly update schedule designed to bring Inventor’s feature set in line with Solidworks’.

At the same time, Autodesk undertook a similarly aggressive anti-Solidworks campaign of in-their-face marketing and lawsuits. For those first few years, Inventor’s marketing staff maintained a presence outside Solidworks World user events – displaying ads on nearby billboards, handing out boxing-motif leaflets and inviting Solidworks users to Inventor social events. Autodesk next targeted Solidworks Corp (and parent Dassault Systemes) with lawsuits contesting its use of similar trademark design (i.e. an orange rectangle) and the use of DWG in its free 2D utility, DWGEditor. (Autodesk lost the orange rectangle lawsuit, and Dassault negotiated a settlement over DWGEditor.)

Ultimately, the three-pronged Rubicon tactic failed. Today, analysts believe Inventor might be in second or third place after Solidworks and Solid Edge, maybe. We don’t really know, as CAD vendors no longer report exact seat numbers. Analysts estimate that Inventor sold 24,000 seats last year and Solidworks 70,000 seats.

Flash forward nearly a decade and now Inventor needs to be killed off, just as MDT was. Ultimately, Autodesk wants all its software running on the cloud, but Inventor is tied so tightly to Microsoft’s proprietary codebase that porting it to another computing environment is neigh impossible. In other words, we’ll never see an Inventor 360.

The designated replacement for Inventor is Fusion 360…eventually. Even though Autodesk touts it as a “cloud-based platform,” Fusion runs on MacOS and Windows desktop computers with some links to Autodesk’s proprietary cloud. Fusion lacks some of the basics that mechanical designers expect from MCAD systems, although it is admired for its HSM CAM (Computer Aided Manufacturing) component. Fusion files are not compatible with Inventor, or Vault even, with users reduced to exporting non-associative STEP files from Fusion.

Sales of Fusion to existing Inventor customers are poor; Autodesk admits 90% of Fusion users are new to the company. As a result, Autodesk slashed the price of the base version to $300 a year and made it entirely free to students, startups of under $1 million in revenues, “enthusiasts.”

To be fair, Dassault Systemes has faced an even more difficult path bringing Solidworks customers to its 3DExperience cloud platform. After ten years of programming effort, it has no cloud-based replacement that Solidworks users want.

Figure 2: Adding GD&T leaders to a part in Inventor 2018

At last winter’s Autodesk University, Autodesk affirmed that it expects Inventor to keep going for another five to ten years and plans to keep enhancing its functions. Inventor always was strong at industrial machinery design, and that emphasis will continue, the company says.

What’s New in Inventor 2018

There were four mid-stream releases of Inventor 2017, and so 2018 naturally rolls up all those new functions. New to 2018 are a number of impressive enhancements:

  • Partial fillets defined by offsets (see Figure 1)
  • Extrudes are relative to other geometry
  • Holes extend in both directions and can have zero depth
  • Sheet metal bodies have thicknesses defined by styles
  • The Measure command automatically determines the geometry being measured

Other new functions fall into the “what took them so long” category:

  • Rectangles placed around text
  • Meshes displayed in drawing layouts
  • BOMs sorted by user-defined strings

Let me touch on some of the significant features new to Inventor 2018.

Big Models: There is a space race going on among MCAD vendors. For each, the issue centers around how to make their software work ever faster with really large, complex 3D models. As it turns out, there are only a few ways to do this:

  1. Avoid loading and displaying unneeded parts of assemblies (easy to do)
  2. Offload display calculations to the graphics board (moderate difficulty)
  3. Access additional cores in the CPU to perform actions in parallel (difficult to implement)

Inventor 2018 also adds tweaks to reduce delays. Some of these include the following: When parts are not loaded, we still need to see them as 3D bounding boxes; in 2018, we can force the display of bounding boxes even of parts that are loaded. There’s faster hidden-line removal and view changes. Two-dimensional elements like sketches, drawing views and raster underlays are displayed faster.

MBD: Model-based design (and its superior cousin MBE, Model-Based Enterprise) is a way to integrate manufacturing information into 3D model files. This lets CNC tools read the 3D model, then produce the part. The idea is to avoid generating 2D plans, which nevertheless remains the common practice.

MBD isn’t terribly complex; it’s just a 3D model attached with leaders tagged with information the CNC machine needs, like finishes and tolerances. Inventor 2018 integrates this capability, along with a Tolerance Advisor palette that warns which aspects still need MBD data (see Figure 2).

Once we’ve attached GD&T and other manufacturing data, we export the model as a 3D PDF file or as a STEP file using AP217 standards. If we still work with 2D drawings, then we ask Inventor to add MBD leaders to the layouts automatically.

AutoCAD: In most design firms, a few seats of Inventor are supplemented by dozens of seats of lower cost CAD software for making detail drawings. So it’s important that Inventor have access to DWG files, and here 2018 makes things better when we place DWG files as underlays, like raster images.

In Inventor 2018, we can open DWG underlays in AutoCAD for further editing. Changes are updated manually in Inventor. When we create sketches, we can project geometry from DWG underlays.

In the past, we could insert DWG files only in part files; with 2018, one or more DWG files can be placed in assemblies. Layers and other properties are prefixed by the DWG file name so one can be distinguished from another. (Think xrefs.)

When it comes to non-Autodesk 3D models, Inventor’s AnyCAD function now lets us edit the foreign parts and then update the native assembly. The same AnyCAD system allows Inventor 2017.4 users to natively edit 2018 files.

Autodesk hopes to sacrifice Inventor in five to ten years to the greater good of subscription-only, cloud-based utopia. This may, however, be difficult to execute, given Fusion and Inventor’s inability to work inside each other’s workspaces. Autodesk suggests using their online A360 project collaboration system, but this isn’t an effective workaround. The transition to the future needs to be made clearer to Inventor users; perhaps it will take Autodesk those 5-10 years to accomplish a transition.

Then there is the problem of Autodesk’s buy-or-die subscription plan, which all new customers are required to adopt. In theory, Autodesk could turn off subscription versions of Inventor to force customers onto Fusion. There is some evidence Autodesk’s subscription plan is giving potential customers pause, as Dassault Systemes credits Inventor’s switch to subscriptions for last year’s 12% increase in Solidworks revenues.

There are design firms that engage in projects that take years to complete; additionally, access to archival data is required for decades to come. They expect long-range certainty from Autodesk. Even as Inventor users continue to enthusiastically design with “their” software, its future in their livelihoods is precarious. Inventor is being squeezed between a Solidworks of the past and a Fusion of the future.

Ralph Grabowski writes on the business of CAD on his WorldCAD Access blog ( and weekly upFront.eZine newsletter. He has authored many articles and books on AutoCAD, BricsCAD, Visio and other design software.


Stories continue below

Print this page

Related Stories