Engine Animation

Today I built a V6 engine in SolidWorks. This is a model for a writing project I’m doing. It’s not done yet, but I wanted to make sure it worked before I went and added a lot of detail to individual parts and added more non-functional parts. Overhead cams, 2 valves per cylinder. I still have to do the intake and exhaust manifolds, the lower half of the engine, I’ve got the valve covers done, but they need some work, and the cylinder heads are a bit primitive. Gotta connect the pulleys to the cam shafts.

The model didn’t really take that long, just about 6 hours for the whole thing. Spark plugs and pulleys are from 3D Central, but everything else is home built. The castings don’t have all the draft and fillets and detailed features that real parts have.

I picked an engine for a lot of reasons. First of all, it’s a break from the kind of stuff I’ve been doing. I don’t think I used a single surface feature on this assembly, but I did use some lofts for the intake and exhaust ports. Also used some multibody stuff on the block and shafts. Second, all mechanical people can relate to an engine. We’re just kind of drawn to that stuff like flies to …whatever flies are drawn to. Third, I wanted to know that I remembered how to do some of this stuff. Mechanisms have always been one of the cooler things to design in SW, and I don’t get to do enough of that kind of thing. Finally, I just wanted to work with the software in a way that it worked more than 20% of the time, just to remind myself that it’s not an entirely awful piece of software for all kinds of work.

As I was working through the history features in the block (one of the few parts with any feature complexity), I was thinking of what it would be like to be doing this in direct edit software. I’m going to have some time or get a pay job one of these days to do some work in some direct edit software. That kind of work is going to be a total change of direction for history users. It’s a completely different skill set, but it’s much more intuitive than history. Still, I think you need history for some things, or for best efficiency at some things.

Anyway, back to the assembly. The biggest challenge with this assembly came in the mates department. There is something about the mate workflow that always finds me making a lot of extra and I think unnecessary keystrokes. Sometimes it seems to take 3 bangs on the Enter key to get a mate to preview, and I’ve never understood why the Mate PropertyManager always works like you have a pushpin pushed, even when you don’t. I use smartmates (alt-drag) when I can, with right click accept. I like that workflow. When I have to use the Mate PropMgr, I’m forever banging the keyboard.

Working with a spaceball is something I’ve grown to love again. I like to use the option to move parts in assemblies with it, but with the Mate propmgr open, using the spaceball to rotate a part causes the entire part to be selected instead of a single face. So I had to turn that option off. Rotate parts with RMB drag instead. Not as nice, but at least mates work better.

I sometimes recommend that people make separate assemblies for different tasks. Sometimes people give me surprised looks, because it sounds like a hack method. This assembly reaffirmed to me that different assembly files for different tasks can definitely be the right choice. In this case, if you build an assembly to reflect a good BOM with subassemblies, you can’t get the motion to work correctly. But modeling it with all the moving parts at the top level to get motion is a bit of a boondoggle, with too much detail to manage a the top. For this reason, I wound up with one assembly saved that used subassemblies to group parts into logical groups that helped me put it together faster, and helped by solving fewer mates at the top level, and one assembly with all the moving parts at the top level, with only one layer deep of flexible subassemblies.

When initially putting it together, I saw the opportunity to save some time by using subassemblies. So I had a piston sub, a valve train sub, and then I had to break it down to make rocker arm subs within the valve train sub. And then to get everything positioned correctly, I had to make the rocker arm subs flexible, because they were all positioned differently along the camshaft.

In the end, with 2 levels of subassemblies, and all of the subassemblies using the Flexible setting, I could see that motion was simply not going to work. I also noticed a funny quirk. If you use the mate to Origin and Align Axes option (which is one of my favorite unsung newish tools), you can’t get flexible subassemblies to work. You get an error about the part being fully defined. You have to remove the Origin mate and fix the base part. The mate to Origin is still newish, but this is something that should be ironed out by now. Not that I’m bashing, I kind of like stuff that doesn’t work. ;o)

Anyway, to get the motion to work, I dissolved the two valve train assemblies. When I did that, the rocker arm subs in one assembly came in flexible, and the rocker arm subs from the other assembly came in rigid. Of course that blew up the tree, but setting all the rocker arm subs to flexible cured the problems. As you can see in the Assembly Xpert, I’ve got 202 resolved top level mates. I’ve got some Cam mates and some Gear mates. And yes, the whole thing will work by just dragging a shaft or a pulley. Not bad for a day’s work.

Watching the video closely, you can see 2 of the rocker arms flipping on the left side and one on the right. This is the cam mate getting confused, and the follower is flipping through the part to the other side. You just can’t escape flipping, even if you don’t use surface trims or planes. Probably slowing down the motor used for the animation or increasing the frame rate / key frames might take care of that.