UG NX5: first impressions

Posted by adam on Oct 10, 2007 in Blog | 1 comment

I’ve been learning NX, and I have mixed feelings about it. If you know another parametric package, it’s not hard to pick it up, but it’s a pretty frustrating interface in a lot of ways. There are also some interesting rumors floating around about the future of the product. More after the jump.

If you’re not familiar with NX, take a look at wikipedia. It’s an enormously powerful CAD package similar in scope and depth to CATIA V5. SolidWorks–though purchased by Dassault–is actually built on the NX platform. The first time you open NX, the relationship to SolidWorks is obvious. The interface, the tools, and the methodology are all very similar. I’ve done a week or two of training in NX now, and here’s my initial response:

Pros:

• Good Parametrics. Most operations can be done associatively, though there are many tools that can accidentally disassociate your geometry quietly; a danger.
• Large tool set with broad functionality
• Very interesting “selection intent” manager. In V5 we were taught to select everything at the “feature-level” to make sure that the historical tree referenced the feature, rather than a b-rep contained within it. NX5 gets around this problem by using a form of selection filter. For example, say you select a segment of a complex sketch using the “tangent curves” selection filter, and use it to create a flange. V5 would remember the specific b-rep clicked, so if the b-rep breaks, the feature breaks. NX5 is much smarter than this: it remembers the “selection intent,” so the b-rep picked is picked along with any other tangent curves (as specified in the selection filter), so that as long as any one of the tangent curves remains in the history, NX can reconstitute the selection. It’s a difficult concept to explain, but the effect is simple: fewer broken links.
• “Direct Modeling” features for working with non-historical solids. Quite powerful, but then I’m pretty sure V5 has most of the same tools.
• Tools are much more “discoverable” than in V5. Workbenches are much less restrictive and confusing.
• Broader import/export support than V5, but still nothing close to Rhino.

Cons:

• Ordered feature list. If you’ve only ever known an ordered list for your parametrics (like SolidWorks or ProE), you won’t be bothered by this. For those of us who prefer the “tree” approach in V5, it’s a serious disadvantage. I’ve tried to find ways of using this more effectively, but to no avail. It has some interesting “filtering” tools that allow you to shorten the length of the feature list as needed, but nothing as powerful as the V5 tree.
• Many tools are not associative.
• Isolated or non-associative elements do not show up in the feature list, which is very confusing. Working with isolated elements is very frustrating, since the interface for these tools is very old.
• Different sets of tools for each type of geometry. For example, there is a trim tool for curves (non-associative), a different one for surfaces, and a third one for solid bodies.
• Surfacing tools are very lacking. There are a number of very interesting tools, but the massive overlap between tool functionalities is incredibly confusing.
• Continuity control is very limited. Some tools allow you to switch from G0-G3, but do not give curvature tension controls.
• No geometrical sweeps (circle sweeps, conic sweeps, etc). Lots of input geometry is required for all sweep types.
• Spines have very unpredictable results in all sweep functions.

Overall my impression of UG has been mediocre at best. I’d still rather use NX5 over SolidWorks, but I’d chose V5 over NX any day of the week (then again, I have about 1/1000th of the experience in NX that I have in V5!)

But this is where it gets interesting.

The word on the street is that Siemens thinks that buying UG was a bad investment, and that their only real interest is in the assembly line automation software associated with it. That being the case, it is highly likely that they will sell off NX to the highest bidder. Who would buy NX? Likely either Dassault or AutoDesk. Dassault has a bigger interest in NX (since the codebase supports SolidWorks), but AutoDesk has been clawing at the top-end 3D industry for a while now, trying to become a serious player again. I can’t venture a guess as to who could win the deal, but either way, it seems likely that NX would be in for some BIG changes (and/or a phase out!) in the relatively near future.