Dezignstuff SolidWorks Blog

In the last episode, we joined the fender and the cowl with a Boundary surface. In this episode we will attempt to make it all the way through without using a word that rhymes with “cowed”.

Why is it that this and most car tutorials that you see start at the front and build to the back? I don’t know.

In the immortal words of Primus;

Is there Heaven?

Is there Hell?

Is that Tuna Melt

That I smell? C’mon!

So what are we gonna model today?

After the Boundary surface joining the fender and the cowl, I split the freaking thing in half. Why? Because that’s what I did. The cowl was easy to build about the center plane, and less easy to build only half of it.

Here’s a great tip that Ed Eaton turned us all on to years ago: If you split something (symmetrical) in half (along the plane of symmetry) and then later mirror it, you don’t get the split down the middle. When you say it out loud, it sounds kinda stupid, but when you do it, it seems cooler. Just remember that next time you work on something symmetrical, and it is more work to model half of something than all of it. You can use a plane to do a Trim (on a surface) or a Cut With Surface (on a solid).

Ok, next we start working with bodies. Most of the car body turns out to be a single surface body, but some of the moving parts will be separate surface bodies. Like the jet rotor. The central rotor is just a revolved surface. Then the fan blades are simple lofts. When making a “looks like” model that doesn’t have to be very accurate, as long as you have something that looks complex, sometimes you can get away with a lack of accuracy. That’s the situation we have here. You can call it “fake” if you like. I call it not spending an entire weekend on detail for a project you’re not getting paid for.

Offset plane. Copy and rotate sketch slightly. Loft. Pattern Body. Apply materials. Enjoy your weekend.

Backpedalling a little, here’s a tip for that nose in the center of the rotor. To get really smooth looking shapes that aren’t all herky-jerky, use splines, and pay attention to the details.

This is a two point spline. Notice I changed the color of the curvature comb (Tools, Options, Colors, Temporary Graphics Shaded) because the default yellow on white is nearly invisible, and added the cap to the curvature comb (Tools, Options, Sketch, Show Curvature Comb Bounding Curve) because it just makes it easier to look at.

The two points of the spline start at the center line of the rotor, and end at the OD. At the CL, the handle is given a vertical relation, and at the OD, the handle is made horizontal. To get the handles on a virgin spline, just select the spline, then select the diamond on the spline handle to change the angle. The curvature comb can be shown via the RMB. When the curvature comb tapers to nothing on the right, it means that the spline is nearly a straight line. Ideally, I would have liked the curvature comb on the left end to end vertically, but it didn’t. That kind of thing can be hard to control. The slope of the curvature comb signifies the rate of change of curvature. If the rate of change of curvature is continuous but changes abruptly across the center of revolution, well, frankly, that’s something that you’ve got to have a really good imagination to see. I don’t think I could tell. If you have a really really sensitive eye that’s offended by abrupt changes in the rate of change of curvature, SolidWorks doesn’t have great tools to accommodate you. You could make the argument that they really really need something like that, but frankly, I wish they would work on other things like flipping connectors or stuff like that. With practice you can tell when the curvature changes abruptly, but the rate of change of curvature is more subtle.

Anyway.

The ground effects around the front of the fender were tough. This is mainly because of quirks in the Ruled surface tool. Don’t get me wrong, I love the Ruled surface. Where would we be without it?!? But it is quirky, and sensitive, and often gives results you wouldn’t expect or want.

The first ruled surface creates a flat surface out from the bottom edges of the fender. Selecting the type can be confusing. “Perpendicular to Vector” means perpendicular to the normal vector from a plane. Wha??? Yeah, it does. So it asks for a vector, and you select a plane. The secret knowledge required to make this work is that each plane has a “normal vector”, which is really just a perpendicular direction. Nowhere in the SW documentation that I am aware of do they make a discussion about plane or face normals or vectors in general. The people who do the programming are geometrical mathematicians, and to a mathematician, all planes have normals, and normals are vectors, so vectors and planes are interchangeable, right? Whatever. You might be better off to just click through the choices and see which one gives you what you want. Just like the new Plane interface.

Next, I used the edge of the first ruled surface to make a second ruled surface, but this one at an angle.

Ruled surfaces have curvature in only one direction. Remember the U-V directions you read about in the SolidWorks Surfacing Bible? On a ruled surface, one of those directions always has straight lines, and the other direction can be any curve. Ruled surfaces can save you a lot of time, or they can be very frustrating. Getting the angle to go the right direction and the surface to go the right direction involve using the Alternate Side button (edge in selection box must be selected) in combination with the arrows button (next to Top Plane above). Each angled ruled surface can go in one of 4 directions, so you have to get used to using these options.

Next is the wheel flare. These can be difficult on cars like the 911 or 350Z, where the flare is really a continuous part of the fender, but on this car, there was no flare aside from a small, very controlled one. So the first thing is to trim it out of the fender. Easy enough. Notice how it has also trimmed out the ground effects flare. We’ll patch that up later.

Next is a small flare that comes out parallel to the axis of the wheel. I used Perpendicular To Vector, with the Top plane as the vector. This is the same settings as the first ruled surface shown above.

Then a small return that looks like a flat face on the outside, giving the flare some depth. This one uses the Normal To Surface option.

Finally the angled return

And after a series of trims and fillets, you get the final product:

Alright. So now you know more about ruled surfaces. Maybe some day SolidWorks will fix them up real nice so you can do stuff like this without swearing.

Next time we will finish the main body and the cockpit.

So. This is a bit of a walk-through talking about how I modeled the Batmobile. Is it more entertaining than a $70 video tutorial of how to model an Audi R8? No idea. I do know it’s about $70 cheaper, though.

I wasn’t trying to be exact on the model. There were things I wanted to make a little differently, and some things that just turned out differently for reasons I can’t exactly identify. I just wanted to do it, so I just did it.

You have to decide early on just how much detail you can afford to put in. The cost is mainly in time. The detail comes later in the process, but having a time budget even for a just-for-fun project is important.

I started, like most of these things I do, from a set of images to place on planes. These particular images it might have been better to make them black and white, then invert them so you would have a white background with black text. Less junk on the screen.

To get these images, I scoured the internet for a while. Any time you are working with artist interpretations rather than actual photographs, you run the risk of various views not matching up. This is part of the reason why I had to do some interpretation: I didn’t have perfect reference material. Even when you are doing product design and the artist providing you with orthogonal views is supposedly on your side, you rarely get stuff that exactly matches up in all views perfectly.

Placing the origin is always a question. For me, I always have trouble lining up the bottoms of the tires for renderings, so the point of contact for the front tire is where I assign the origin. That means that if the rear tire is a different diameter, it is still easy to line up the rear tire so it doesn’t sink into the ground on the rendering.

You’ve got to use a size reference, which in this case was just a straight line representing the wheelbase. Then I applied a dimension according to the image, and scaled the images to fit. Remember, 1 pixel = 1 mm by default for these sketch pictures. Name these something obvious, and don’t put any sketch geometry in the sketches that you are going to use in other sketches. Otherwise, they will go somewhere other than the top of the FeatureManager.

I think that conceptually, the most difficult part of models like this is trying to lay out which features you will use to create which shapes. A very respected SW user once said that you shouldn’t think about features, or something to that effect. But I disagree. I think you have to think about features. You have to know which features can achieve which shapes. You also have to understand that the Trim tool can help you separate the useful part of a feature from the not-so-useful.

My next step was to decide on a face to start with. I like to start with the hood. Since this car really doesn’t have a hood, I started with the tops of the front fenders. To make this shape, I’m going to make a surface that is bigger than I need, but the right contour. Then I will trim it to the right shape.

Remember that the Boundary surface is magical. You can make surfaces from sketches in the shape of a T. Usually, a T would only make a sweep, but it makes Boundary surfaces well. You can also do an F, and whatever the letter would be called if F were mirrored about the vertical line in the F. Or an H where the center line is extended. That’s what I did here because it was the information that I thought I had. I didn’t need to give it more info.

Try to name planes and sketches, especially if you think you might edit this later or someone else will look at it.

Don’t work in perspective. Even if not working in perspective offends your finer artistic senses, SolidWorks sketching and dimensioning do not work very well in perspective. Plus, if you are such an advanced being, you should be able to compensate for the screen’s lack of perspective in your mind. Renderings need to be done in perspective, sketching and dimensioning don’t. Remember cameras use perspective by default.

Next I made the cowling for the jet engine. If you have all of your sketch pictures on, this can start to get a little overwhelming. Remember that in 2010, the Display Pane (flyout shown to the right of the FeatureManager) can help you control the display of bodies. This can be very helpful in complex models. You can also use it to turn off sketches. I think this is a great bit of functionality.

So now the jet cowl and top of the fender are complete. These two shapes are going to have to blend at some point. How to do that may be a tough question. This was something I had to settle through trial and error. I didn’t just make up my mind to do it one way and it magically worked the first time. I went back and forth on this. Don’t fool yourself. This is not an exact science.

A little bit on the trimming of the top fender shape. Remember that there are two ways to handle blends or fillets. You might have to model to a sharp and then fillet it, or trim back from the sharp and use other features to blend. The smoother the blend, the more you will have to cut back.

Next I added an interior to the inside of the jet cowl, and a side to the fender. I just modeled right over the wheel well, knowing that the wheel opening and flare will be added later. Some wheel flares are so extreme that they have to be modeled as part of the panel they are on, but that is difficult.

I’m looking at doing a Nissan 370Z, and the front fender flare is gonna be tough. It is all a single flow along with the front bumper and the side of the car, right up to the hood.

Next I added the inside of the fender. This is mostly a throw away surface used to make the transition between the fenders and the jet cowl. I have two vertical sketches and the single edge of the top fender to create a Boundary surface.

According to Mark Biasotti, you need to get in the habit of using Boundary. You have heard that before, but maybe you haven’t hear why. The thing is that Boundary is more accurate than Loft, and as we see later in this model, when you try to Knit together lofts, you might run into tolerance stackup problems. I know. You are used to thinking of your CAD models as being exact, and only worry about tolerances in manufacturing and assembly. But now we learn that the computer is not always as exact as we imagine it to be.

When using Knit, if you start getting strange errors, such as “there are no edges to knit”, try messing with the Knit tolerance settings. To make it work, generally a setting of .001″ will do it. At least it did for me. The physical size of the batmobile was about 20 feet.

Anyway, the last feature to deal with for this post is the front corner of the fender. This is one of those situations you should avoid, but in a case like this where we are just making a fun model, the degeneracy might be something we can get away with. In this case, I used two edges and a sketch. The sketch contains a spline with c2 to edges on both sides. I just try to allow the software do what is natural, which will be the thing most likely to work and look good. If you try to force a shape on it, that force fit may backfire on you.

Notice that each edge selection is actually a Selection Manager selection, where two edges are picked for each of the Boundary’s Direction 1 curves.

Ordinarily I would do something like this with a Fill, but for some reason, a Fill must not have given me the results I wanted. I spent a lot of time on this feature.

Come back next time when we will step through modeling the transition from the fender to the jet cowl, and you’ll hear Bat Lombard say “Holy 3D, Batman, this thing has handles!”

I know, modeling an automobile is daunting. I almost dread starting a new car modeling project, even if it is just for fun. It took me weeks to get around to finishing the batmobile once I started it. But the actual work only took about 10 hours. I still have a couple of adjustments to do to this, but I’m a little hesitant to adjust much, since models like this tend to be fragile.

So for this challenge, submit models or renderings of models you’ve done. Or if an entire car is too much, just a car part. You could model a mirror mount, a hood scoop, rimz, a grille, steering wheel, a truck. I don’t know, just submit something. You don’t have to model something that actually exists, you could do your own concept car, or a cartoon car or whatever. The last challenge (mannequin) got pretty poor results, so we need to compensate.

In the coming week or two I will put up some posts about my experience modeling the batmobile, and try to go through each major component of the model. I’ve already started documenting some of the difficult areas, and areas in which SW performance was less than stellar. Overall, it’s a cool model, and I was able to do most of what I set out to do, so the software is capable.

I promised Matt Sederberg over at T-splines I was going to feature an automotive competition he held on his site. It took me some time to get around to it, but this is it. You can get some ideas for stuff to model from the t-splines automotive modeling challenge. There are some nice models over there. My favorite combined an old Firebird with the new Camaro. Nice…

I don’t expect the stuff here to look as nice as all that, but then I might be surprised. Tsplines and Rhino just make that kind of modeling easy. SW doesn’t. Drawing cars in SW is like doing chainsaw sculpture.

Looking for more SolidWorks automotive inspiration? Check out Paul Salvador’s stuff. Paul has done a lot of cool models – the very first nice Audi  models I saw, and this was at least 5 years ago. Paul does nice stuff. Check out the link.

Just as a side note, in making some of the images for this blog post, I used PhotoWorks and PhotoView360. I have to say that for all of the hoopla gone into PhotoView, I don’t care for it. It is slow to revolve your part, and severely limited in the kinds of backgrounds available. I know you can add your own, but I thought this was a dumb-it-down kind of product. I definitely need a dumbed-down rendering software. I think the main problem was that I could just never see the model very well in the PV, where the PW model stood out clearly. Anyway, I’m not a real renderer, so what I think about that stuff may not count. Just for reference, the one at the top of this post is PhotoWorks, the one at the bottom is PhotoView. PhotoView images all appear to be dunked in oatmeal.

Also, I want to put you on notice that the next modeling challenge will be an aircraft, so you might want to get started on your favorite WWII or Vietnam jet.

======================================================

Entry #1 is from Robert Bilbrey.

These renderings were done in Photoworks 2010.  I also find most of PV360 difficult and “fuzzy” although some of the textures and manipulation tools are useful.    The car is a 1953 Jaguar XK120 ( one of which I years ago sold for pennies, dammit ).  It was one of my earlier surfacing attempts and I can see flaws that could be remedied.

Nicely done! I love the old Jags!

============================================

Entry #2 is from Fernando Mota. Renders done in Hypershot, PV360 and Vray:

=======================================================

More models from Fernando Mota, modeled in Rhino, rendered in Hypershot:

==============================================

Entry #4 is from Mark Biasotti:

Couldn’t resist.

Modeled on SW2003 Beta 1  Rendering using Modo 401.

You know, I’ve always kind of admired people who have the patience to model a car’s interior. I just don’t have that kind of patience. I mean where do you stop? Detail is the most expensive thing you can add to a model. Some times it pays off, and sometimes it’s just useless detail. A car with an interior is more believable.

Anyway, Mark’s Vanquish has always been one of my favorite SW models. very cool, and thanks for sharing.

==================================================

Entry #5 is from Ben Paprocki:

Here are a few things I’ve worked on.  The first two images are of a car I’ve been tweaking over the past couple of years.  As with most projects, it’s never quite done.

Second is a rendering that I submitted to the CSWP SWW2010 design competition.  It took second place.

All of the images were rendered with PV360.

Nice stuff! Thanks for sending it in!

==============================================

Entry #6 is from Steve Ostrovsky, a reseller from the south east.

This is a PV360 rendering of a truck wheel and tire I modeled. Thought it would fit in with your current challenge.

I also attached 2 more that are a little off the beaten path – both done in PV360. They’re still vehicles, but … I was just having fun.

These are fun. I love anything with Marvin the Martian. And that toy tractor in that surreal barn photo?!? Fantastic. You gotta have fun with this stuff. Thanks, Steve, this is great stuff.

============================================

Entry #7 is from Matt Lombard. Some of my old stuff I just wanted to include here.

The cobra was done for the Surfacing Bible as an example of surfacing over a scanned point cloud. It was rendered in Hypershot.

The F1 car was done for the Happy Valley user group presentation, and rendered in Photoview360.

This Porsche 911 model only had a brief life on display back in 2004. It was never quite finished, but somewhere along the way, someone made fun of it. When I look at this today, it really isn’t a very good model, and doesn’t look much like a Porsche. I love the 911, though, so I might take another crack at making a decent model. Rendered in Photoworks. I show it here mainly as proof that a user can really improve if they put their mind to it.

This model was also done for the Happy Valley user group, and rendered in Photoview 360. This is probably the best PV rendering I have made, and it was also one of the first.

Same model as above from Photoworks with a cool paint material that has a two tone effect (works like shining a light of a different color on a part of the model).

===================================================

Entry #8 is from Rob Wolkers from The Netherlands with a model that is probably pretty familiar to most SW users by now:

Great topic, some pics of my Burton Elementz.

Modeled in SolidWorks 2007 and rendered in Photoworks.

=====================================================

Entry #9 is from Pawel Keska. Several great looking models!

I attached couple of my renderings. Each model was done in SolidWorks, rendered in PW or PV360.

There is nothing that will make you forget the problems of the future better than building a difficult model and dealing with today’s problems. Talking about modeling is so much more fun and productive than talking about potential future screw-ups. The present has enough screw-ups for me. I’m trying to break out of this crazy c___d spiral that we’re stuck in, but I just can’t complete this model. When finished, we will have another modeling challenge and another series of posts on surface modeling like when I did the Model A. We (I) need a change in direction.

So as a distraction, can you identify the item in the image to the left, and who made it famous?

 

This is a project Jeff Mirisola sent to me from a company called Status Racing. They have an existing car seat, but didn’t have CAD data for the mold, so they wanted a parametric model. The process was interesting, so I thought I’d pass it along.

First of all, it took a big box to get this thing across the country from Washington to Virginia. It’s a racing seat shell, made of fiberglass. The back side of the seat is nicely finished with the weave showing through clear epoxy. The inside was a little rough. I planned to scan the inside with my Next Engine scanner, and the scanner likes things a certain way, so the first thing I did was to paint the inside of the seat with gray primer. I just needed a flat paint with a fine texture that was light-ish in color. 

This part is nearly 36 inches tall, and the Next Engine scanner has a relatively small field of view, so the scan had to be done in smaller pieces. In order to line up small scans, I made some markings on the part with a Sharpie pen. The scanner takes a photo, and maps the photo onto the 3D points. Then I had to line up the scans, matching 3 points from 2 different scans. The result was surprisingly accurate for what it did. You can see the patchwork of multiple scans that makes up the data. 

The scanner only gets decent results when it is at a certain angle from the face being scanned, so scanning the lip around the outside was a real pain, and had to be done separately from the rest of the seat. 

When working with data sets this physically large, I discovered the hard way that you need to relax the space between scan points. My machine with 4 GB of RAM was choking and paging. My 32 bit machine kept crashing trying to work with the data. 64 bit worked much better, but I still needed about double the RAM for this project. Actually, I took it back tothe Scan Studio and reduced the mesh somewhat to get it in to SolidWorks. Scanning only half of the seat (assuming symmetry) also helped me keep the file size down a little. 

I was able to bring this scanned data into SolidWorks. One of the big problems with scanned data is that it is very difficult to get it aligned to the origin. In this case the data came in completely cock-eyed, and I decided to just model it where it sits, and then clean up the model later by copying sketches.

The next thing you’re faced with is how to break all of this data into features. I decided that this was basically a two feature part. The first feature was just the entire seat. The second feature was the rim that goes around it. Of course it would have to be mirrored, and the cut outs and all, but there are only two functional features here. 

In this next picture you can see the mesh super imposed on the 3D model, with splines showing. The main feature of the seat was a Boundary surface with 15 sketches. 2 in Dir 1 and 13 in Dir 2. I didn’t get this right on one shot, I had to go back and add and remove sections to get the surface as close as possible. Matching the splines to the scan data was the big trick in this job.

I had thought of doing the back in one feature and the sides in another, but they would have to blend together anyway, so I just made it all one. I think it gave a better shape anyway.

The 3D spline that forms the outer edge was difficult to get right. It was basically done backwards, using the end of the section splines for reference. I adjusted some handles on the spline to get it to match the scan data. 

The one section of this part that took the most work was the top side of the wide shoulder section. Notice that one of the section splines intersects the 3D edge spline at a very shallow angle there. This makes for a surface that is tough to control. Boundary surface works much better if the directions are perpendicular to one another. 

The most difficult part of the whole project was the lip that goes around the part. This is a single Boundary surface with 13 profiles in one direction and just using the edge of the part in the other direction. The Boundary surface is quite remarkable, and on this part it really showed its stuff.

Here is a detail of how the lip was done. This is all splines. But splines do have some control. It was very difficult to thicken this part once the surface was done. The reason it wouldn’t thicken was because the splines had sharp kinks in them. To control that. I applied curvature controls, and a min radius marker. With the min radius marker, I would adjust the spline to get an acceptable number on the marker.

To get this sketch spline tangent to the surface, I used the Intersection Curveand just made the section spline C2 to the intersection curve.

In the end, this worked like a champ, but it took a long time to get there. The final rim feature was one of the craziest features I’ve ever created, but you can see in the image with the edges showing that the rim to the seat connection was tangent (shown in phantom line).

Overall, I was feeling better about the software after this project. Sorry, I can’t provide downloads for this part, it is proprietary, but I did get permission to show the part. 

One of the little secrets about creating complex shapes in SolidWorks is that they aren’t all as complicated as they look. The grille is one example of this. Once you know what you’re doing, something like this is actually very easy to create.

First, we have to select a feature type to use. This is going to have to have curves in both directions, so it could be a boundary. The profile needs to change, and doing guide curves to change the profile would be too difficult. So it will be either a loft or boundary. In this case it really doesn’t matter, so I selected boundary, just because it’s the future direction of development, and clinging to the past is oh, so gauche (he says as he models an antique car).

The first thing I’d do to model the grille is realize that the thing is symmetrical, and that there is a sharp edge at the plane of symmetry. This makes things much easier.

Of course something else you want to do is to get a picture of one you like and work from that. Realize that this is a cartoon Model A, so I’m not really trying to recreate this 100%. It’s gonna be a little exaggerated in some ways. This image of the 34 coupe captures the flowing shape between the front fenders and the grille together. This grille is taller and more narrow than what I’m looking to create, but it has the overall form that I want.

You can start to model this part in one of two ways: draw the curve that becomes the “direction 2″ 3D curve or path for the boundary, or draw the 2D sections for “direction 1″. Direction 1 and 2 are completely arbitrary in real application, but I’m pretty entrenched with the old way of doing things, so I think of Dir1 as “loft/sweep profiles” and Dir2 as “paths/guide curves”. It doesn’t matter how you use the two directions, which is part of the point of the boundary feature to begin with – both directions are treated equally.

Ok, so I select to make the 3d curve path first. The 3d curve is essentially a C shape, angled back and curved in both directions. You can do 3d curves in a couple of ways, but the most important are the 3d spline, and the projected curve. In this example, I’ll use the projected curve. A commenter suggested having a look at 3D sketches, and I will get to that in my next post on this Model A where I look at creating the exhaust pipes and the sparkplug wires.

Projected curves tend to be confusing for people. Not sure why. Some people get it immediately, and some just don’t. Anyway, there are 2 ways to “project” the projected curve. The curve can be projected onto model faces, or onto another sketch. Projecting one sketch onto another is where most people who are going to get lost usually get lost. The sketches that you project are usually on planes that are perpendicular to one another. Sometimes you have options to pick different combinations of planes, such as front and top or front and side. In this case, front and top will be the easiest. Sketch onto Sketch projected curves essentially extrude both sketches through space, and where they intersect, it creates a curve. You can think of this like extruding one sketch as a surface and making a sketch onto face projection with the other sketch. You could also think of the projected curve function as being like a reverse drawing. A drawing takes a 3D something and creates 2D views of it. Projected curve takes 2D views and creates a 3D shape.

Actually, the grille in the image of the 34 coupe is slightly angled back in addition to being angled to the side, but I didn’t model it this way. I just took the easy route, but I’ll also explain the more complex operation as well.

So here are the two sketches. The arc with the dimensions was sketched on the top plane, and the spline was sketched on the front. Notice that the spline has 2 endpoints and a midpoint. The shape is controlled by the handles. 

When you make a projected sketch from these, the grille will be essentially straight up and down because the arc is placed on the top plane. To get the grille to angle 

the way it is on the image of the 34 coupe, you would need to put the arc on a slightly angled plane. Anyway, this results in a 3D curve.

3D curves are not as flexible as 3D sketch splines, but they are far easier to control. There is a lot of overlap between curves and 3d sketches, and a future post on 3D sketches will talk about a lot of this. SolidWorks has not put much development into curve features in the last many releases (variable pitch helix was I think the last change of significance). But they have poured massive development into 3D sketches in general. Even the new Equation driven Curve is actually a sketch element. I keep wishing they could dig themselves out of the terminology quagmire they’re in but it just gets deeper and deeper with each new feature.

Anyway. So with a 3D curve in place, now it is time to draw the sections. On the side plane, I drew the section of the grille at the top, and then in the same sketch I drew the section of the grille at the bottom. You couldn’t get away with this in days gone by in SW, but you can now. I still don’t think it’s a great idea, but this was just a model to create an image. Too many things can go wrong when you combine sketches with separate functions into a single sketch.

 

 

This image shows the top and bottom in a closed sketch, and an intermediate profile in an open sketch referencing the curve created earlier. 

When making models for renders, remember that sharp edges don’t render well. Usually you want to have a small fillet on a sharp edge. In reality most edges that we characterize as dead sharp actually have a visible round. 

The next step is to create the boundary feature. Remember that boundary uses some of the same rules as loft. The one I’m thinking of here is that if you want a loft to not twist, select the profiles from the ends that you want to connect. Same applies here. It is nice that with the boundary, you can RMB on a profile in the propertymanager and select Flip Connectors. 

In this case we don’t need any end conditions unless you want to add some shape around the plane of symmetry. 

In the end, the result looks like this. Remember to apply a shiny appearance. Just doesn’t look the same without it.

The next part to worry about is the mesh inside the grille. No, I’m not going to actually model it. This is a “looks like” model, not a production model, so shortcuts like this are permissible.

The easy way to do this is to make a surface of the shape desired and then apply a texture, ahem, I mean an appearance to make it look like slats.

This surface is easier to create than you might think. On the side plane, I drew an arc with a suitable radius to add some curvature to the grille. The arc went between the open corners of the boundary feature on the side plane. Then I used a Fill feature to fill between the sketch and the inside edge of the boundary. And that’s it. It will look better if you turn off  ”optimize”.

Finally, add an appearance. The vertical slats probably look better than the horizontal ones.

Big headlights, baby. Don’t be shy. Mirror and you’re done.

By the way, if you are impatient with how little information is here on this surfacing stuff, just go and buy the Surfacing Bible. I’ve re-read it recently, and, well, I’m biased, but I think it’s a really nice book.

Are those sharks with lasers on their heads?

No, Dr. Evil, but it is almost as cool: a wacked paint job on your new Model A lair. Sure beats ill tempered sea bass.

Wow, that’s a wild effect on the paint job. This is just RealView. No rendering involved. 

Here’s how to get this two-tone effect:

     First, select a Painted Car appearance from the list.

 

 

 

 

 

 

 

  Then select one of the appearances from the list, and Alt-drag it onto the part. The Alt-drag tip was in the Alternate Reality post a couple of days ago.

Alt-dragging the appearance activates the edit of the appearance immediately.

Notice that these car paint colors come with two color options. Assigning a contrasting color allows you to get some really wild color effects. My favorite color is royal purple. No, I’m not gay. Anyway, the red and purple combination looks really cool, but you can get a lot of cool color combinations like red and white, or orange and yellow. It’s interesting to see how they established some of the default colors.  This is something you might have to play around with a little bit. 

Unfortunately, the color funk doesn’t translate into Photoview 360.

Anybody have any cool screen captures you wanna share?

Ok, we’re all back from Thanksgiving. I promised a while ago that I’d go through how I modeled the Model A model I created for the Happy Valley user group. So this is it. I’ll do this in a series of posts that cover the most interesting elements of the work.

Keep in mind that this model was not made for production data, it is primarily just a show model, and so I took some shortcuts with it that I wouldn’t take with a real production model.

Layout Sketches

Many people who ask about models like this start by asking “where do you start?” I can see where it would be difficult to envision where to get started until you’ve been through the process once or twice. Other people who do this kind of work might answer this question differently, but this is where I started:

In the end, the wheels were made as separate parts, so they were not driven by this sketch, but the fenders were. This sketch doesn’t include the cab, or the grille, or the engine, or the windows, or the width of the car, but it offered an overall scale for the model, and a place to start.

It may be dificult to see from the image, but the fender was created as a single spline, and then later split into three sections using the Split Entities tool. Notice that the spline has more points on it than you might consider ideal. This is partially to control wiggle around tight corners, and to try to keep the running board section of the spline as flat as possible. Flatness in splines with more than two points is one o the most difficult things to achieve. Flatness was measured using a curvature comb.

Very often, a complex model will start from more than one sketch. The other starting part for this was a footprint for the top view.

For those who weren’t around for the earlier post, this shape was created with a construction rectangle, then arcs were created at all four sides of the rectangle, and made into construction arcs, and then a Fit Spline was placed over all of the construction arcs, and the fit tolerance allowed for some blending on the corners.

Why didn’t I make the fender sketch offset from the middle plane? Is laziness a sufficient answer? To me, it didn’t matter. Since it didn’t matter and it was extra work, I just drew it in the middle. I didn’t really need a plane on the outside of the car, so I built one fender in middle of the car, and then moved it to one side (Move/Copy Bodies) and mirrored it (Mirror, using the Bodies to Mirror selection box).

Parametric moves like what would be required to move the fender with its plane and all the associated sketches are just very problematic in SolidWorks. To me, it is better to make what you want where it is easy to create it, and then put it where you want it. Moving the finished fender body doesn’t risk anything blowing up. Moving a plane and a set of associated planes and sketches risks blowing up each of the associated sketches. It’s just that simple. Entropy always wins, so start at a low state of entropy, and it does not increase naturally.

Reference Surfaces

When surface functionality was first added to SolidWorks, they were listed in the menus under Reference Geometry, much like planes and axes. Sometimes surfaces really are just reference geometry. One example of this can  be shown in building the cab of the Model A.

From the Fit Spline created in the second image above, I created a surface extruded with draft. The thing was that I wanted to create the cab with a bit of an angle at the bottom, and I wanted to control the angle. I wasn’t able to make this happen through any of the loft options. Sometimes you just have to make your own luck, especially with SolidWorks surfacing.

Here is the cab surface being created from the extruded surface to a sketch point. Yes, things like this happen in real world modeling, this is not just an academic curiosity.

You have several things to look at in this image. The first is that I loft from the edge of a reference surface to a point. The second is that by lofting to a point, I don’t get a pointy loft. The third is the use of the Normal To Profile option with the Point sketch. Forth is the use of the profile weighting. This is a really simple loft, but there is so much going on.

Tangency to a point makes the surface radiate out from the point, which is kind of opposite of what making a loft tangent to an existing surface at an edge does. It is usually used to cap off an open surface, or make a shape like the end of a toothbrush handle, in contrast to the end of a cigarette.

One portion of the truth is that you should try to keep your lofts simple. In the same way that splines with lots of points are difficult to control, lofts with lots of sections are also difficult to control. This loft is the equivalent of a two point spline with tangency and weighting at both ends making a transition around a corner. What splines do in 2 dimensions, lofts do in 3.

Using the weighting enables me to make the top “flattish”, and make the sides follow the drafted reference surface rather than bulging out in a big “pumpkiny” shape. It also makes for a sharper corner. There is a point where pushing the weighting numbers too far can cause an S shape instead of a U shape. The numbers run from 0 to 10, and don’t really mean anything exact. They are just relative weights.

Anyway, after the loft feature, there are a number of things you can do with the reference surface. Some people use Delete Body on it. This doesn’t really delete anything, but only hides it, and makes it unavailable after that point in the tree. Some people will simply hide it. You never know when you might want to reuse it.

I hope this was useful for you.

Splines make some people squeamishly uncomfortable. You can’t put dimensions on them, so how do you manufacture it? How do you make parts fit together if you use splines? How do you run QC to check your parts?

There are a lot of misconceptions about splines. The first is that you can’t put dimensions on them. And the second is that if you put dimensions on them that they become fully defined. The Third turns out to be that if you fully define all the points on a spline, that you can’t drag the points around anymore. Each busted misconception in its own time…

I have good answers for each of these misgivings and misconceptions, but I’ll get to those later. I want to start by addressing why splines are so important, infact indispensible.

Great curvature, baby!

Ok, this is going to sound sexist. That’s because it is. So sue me. If you’re male, there is no way to pretend you don’t notice this kind of stuff. Some types of curvature just draw your eye. That’s all there is to it. Just to keep it kind of cartoony and not too explicit, let’s use Jessica Rabbit from the movie Who Framed Roger Rabbit as an example, because she’s a great example.

Splines are sexy. Jessica just wouldn’t be the same if she were just straight and round, drawn with lines and arcs. Jessica, and just about everything else that we think of as “products”, needs curvature, and in particular, changes in curvature.

The reason why splines are so necessary is that they are the one way we have to make shapes that look organic. Organic shapes are never perfectly straight or perfectly round, they are constantly changing curvature. We as humans are programmed to react to organic shape. Jessica Rabbit’s organic shape provokes a reaction, and product design tries to evoke the same kind of reaction. It’s not a coincidence that they both cry out “empty your wallet”, and we dutifully obey.

If Jessica has an equivalent in the automobile world, it is probably something like a 1960 Corvette. Yow. They both look great in red. Nice headlights! Love those rear fenders. How does it look with the top down? It’s no coincidence that product design is so sexual. It’s no coincidence either that they are both beautifully curvaceous.

Splines or lines and arcs?

Cool shapes in product design is all about curvature that changes smoothly from one radius to another rather than simply a series of tangent arcs that change from one discreet radius immediately to another

discreet radius. Look at these two surfaces below. Which one looks better to you? The one on the right is less lumpy. I think most people would say the one on the right looks better. Of course the one on the right is made with splines. You can almost see the lines and arcs in the one on the left.

Of course it is possible to make splines look very bad if you are trying to do that or are very inexperienced. I think a lot of people make the mistake of thinking that a spline with more spline points is more controllable, and thus a better spline. The truth is that you DON’T want to try to force a spline. The thing about organic curvature that makes it so appealing is that it looks natural. The natural materials are bending or being formed in a natural way. Splines have a very natural way of bending. I’ve been told that the math governing how splines interpolate shapes between points is very much like the natural bending equations for elastic materials.

Interpolated shapes

I classifify shapes into two camps: analytical and interpolated. Analytical shapes have a specific equation that can be written, for example a line or circle each has a familiar equation that you may have learned in high school geometry. A spline I believe is a series of 2nd or 3rd order polynomials. With an analytical shape, you know exactly where every point along that shape is going to be even before you draw the line or the arc. With a spline, you just put down the control points, and SolidWorks interpolates the curve between the points. This concept follows through to features such as the boundary, loft, fill and sweep. You put down the profile curves, and SolidWorks uses the Boundary feature to interpolate the surface between the curves. The difference between analytical and interpolated curves and surfaces is very important. They are not interchangeable.

Arguments against splines

People who don’t trust splines typically talk about dimensions or repeatability. When I create splines in SolidWorks, I never fully dimension them. Its useless to dimension splines. Different CAD packages evaluate splines differently, even if the spline points are all in the same place. Even SolidWorks between releases evaluates splines differently. Even if you were to use a spline on a drawing and fully dimension all of the spline points, the spline could never be exactly recreated based on those dimensions.

The only way to transfer spline based manufacturing data is to send a completed 3D model, and for the manufacturer to use a computer controlled method. Paper drawings may be useful for notes, and reference dimensions but typically not as a way to specify a complex surface or curve. This is most of my business. I don’t remember the last real drawing I gave to a molder, I just hand off 3D data.

Practice with splines

See the spline here with a lot of points looks lumpy. When I first started working with splines, I thought this was the best way to get great control on a shape, but it turns out that the more you try to control the shape, the lumpier it gets. Notice that the second spline is much smoother. This is because you just let the spline math do the work for you. My general rule of thumb is that you add a spline point for every change in convexity (convex up changes to convex down). So you have the two endpoints, which would make a straight spline, a middle spline point, which would make a convex or concave spline, and then a third point which gives the undulating S shape.

In the second set of images, I have applied a curvature comb to the splines. This is a way to visually evaluate the spline. The curvature comb represents the instantaneous curvature at that point along the spline. If the curvature comb changes sides of the spline, that means the spline has changed convexity. Notice that the comb on the top spline confirms that it is lumpy and the changes in curvature are too abrupt. The changes in curvature in the lower spline are much more gradual. Changes in curvature are represented by the change in the height of the comb.

Just for reference, curvature is the inverse of radius, or c=1/r. If you have a large radius, it means small curvature. Small radius (tight corner) has large curvature. Notice that the curvature comb is taller in kinks of the top spline. Notice that in flat areas of the spline (almost linear) the curvature comb is almost zero height.

More to come…

We’ll talk more about this topic later. There is a lot to know about splines.

I’ve been wanting to do an antique car for a while. A visit to a friend with a Model A street rod kind of inspired me. It took a while to get around to it. I promised the leader of the Happy Valley user group that I would create a new model as an example for the next meeting. I’ve got probably 9 hours into this model. It could have used some additional detail, particularly the engine. I got a simple engine off of 3D Content Central, and juiced it up a little. Work like this is usually worth the time because the model might get reused for other things. It’s nice having some models that don’t belong to someone else.

Some really cool effects on the paint for this model. There is a lot to show here. Some Photoworks, some modeling, lots of bugs, actually. Some quick and simple surfacing too, and several examples of different ways to use the Boundary feature.

I started this model on my tablet PC while on a road trip (Kim was driving). At that point the cockpit looked like a pumpkin and the whole thing gave the effect of Cinderella’s pumkin coach. I decided to back off the cartoon a little. This is still a bit cartoonish, and for all you purists, yes, I know the front end does not belong to a Model A, but I saw it on another car and decided I needed to mix and match. Cars from the 30′s are so cool to model.

The images above are all from PhotoWorks. This one below is from PhotoView 360. Some day I’ll get around to doing a review of PV360. Everybody else seems to have done one. Just as a hint, don’t try to run PW and PV at the same time. PhotoWorks always loses.

PhotoView360 has a distinctive look. That’s good and bad. It’s good because the look is gorgeous lighting, reflections, and shadows. It’s bad because it’s the only look it’s got, and there aren’t many options. Anyway, I’ll be back later with a surfacing tutorial on this model, and maybe a rendering tutorial too. Happy Valley gets first dibs, though. If you’re gonna be there, you’ll see it first.

Yes, this was modeled and rendered in a 3D modeler. Click on it to see the full size and full detail. Amazing. This was one of the entries that wasn’t chosen as a winner in a Luxology contest to model and render a Hellfire Widow spider. This stuff amazes me. The people who do this use multiple tools to accomplish a single image.

Keep in mind, this Luxology is the same company that brings us PhotoView360 (Rob R, Ricky, Smack). I’m anxious to see where this is going. Frankly, I’m not excited about the fact that SW has put so much development resource (money, time, people, etc) into an area that is so completely half done, if you will pardon the contradiction. PhotoWorks gets gutted every other year, it seems, so I can never get any traction on learning it. But you can’t replace it with this new software because there are too many things it can’t do. Still, the new PV360 is as easy to use as Hypershot, but with a much better interface, and gets good results fast. Most importantly, making small changes enables you to iterate quickly.

I liked the comment so much before I’m going to use it again. Getting images out of SolidWorks is like going to visit my brother. At any one time his house has 2-3 half finished construction projects, and he’s always saying “just wait til this is done, it will be great!”  Or you could say it’s like driving the interstates in Pennsylvania, where the orange cone is the state bird.

One of the great things about software development that is different from my brother’s house, or I-81 in PA, is that you can (and should) do the dirty construction work behind closed doors, and present a finished product when it is ready for real users. SolidWorks is so eager to share the new stuff that as users we are perpetually burdened with determining exactly what percent baked this shiny new stuff is. We need a plan. OpenGL/DirectX, RealView, PhotoWorks, and now PV360. This looks like an appalachain automotive lawn managerie, not a plan.

Anyway.

We as SolidWorks users tend to be a little parochial when it comes to trying to do everything with a single tool, regardless of how ill-suited it is to the task. modo is only $895. Upgrades are $395. Plain and simple. You can get a 30 day trial, and for $25 you can get a 30 day trial with full tutorial documentation. 

Here’s a great review of the software that gives you an idea of what it’s all about. This is all from a CG point of view, not considering CAD applications. After reading the review, it sounds like SolidWorks Corp could learn something from these corporate rejects about how to deliver software that people really love and can really excel with. It sounds like modo uses a somewhat limited function set, but what’s there is really there.

There are some real masters out there with this stuff. SolidWorks is capable of creating all of the geometry on this spider. Why do we not see people doing this kind of work in SolidWorks?

I just can’t let this one go. Check out a time-lapse of a stylized car being created in modo, and a little movie of a stylized car being created in SolidThinking. You’ve got to click the “car sketching” button to see the SolidThinking video. This is the difference between mesh and nurbs modelers. The SolidThinking method is very similar to the way I created the Cobra model.

modo is a mesh modeler, and the workflow in a mesh modeler is completely different from the workflow in a parametric nurbs modeler. Of course the utility of the mesh model for manufacturing is limited, but the shape creation and concept development is fantastic. When I get out from under current obligations, I’m going to play around with this stuff a little and see what it’s all about. From time to time a mesh manipulation tool might be just the ticket.

The one thing I am wishing is that the whole PV360 thing is just the beginning and that the real product in the wings is some sort of modeling collaboration between SW and modo. It would have to offer nurbs on top of the mesh to be of any more interest than modo by itself. This could be the concept modeler I’ve been keeping my eye out for.