Enabling Contact Sets locks motion study

[Anonymous]

Hi

I have been working with Fusion for a while now and decided to try the joints and motion capabilities.

The model I'm working on is an engine, as can be seen in the image below, the crankshaft and pistons are working fine but I'm not able to get the camshaft to function correctly.

There's only one valve right now which I'm trying to get to work properly before moving to the next ones.

What is happening is that when I do not add any contact sets the valve collides with the camshaft during motion study (and thus the valve stays motionless):

The movement of the camshaft is limited to rotation, the movement of the valve is limited to sliding across its central axis, which means that it can only go up or down.

When creating a contact set between this camshaft and the valve, Fusion seems inable to process the movement needed to avoid a collision of the valve and the camshaft. The only movement that should happen is the valve moving down, which is not happening.

I can add a motion link between the two though which, I think, means the parts are not too limited in their movement (while the camshaft rotates the valve goes down steadily by the defined mm per rotational degree of the camshaft).

How can I solve this issue? Thank you kindly!

Just fyi, I checked these threads:

http://forums.autodesk.com/t5/design-validate-document/valve-camshaft-motion-difficulties/td-p/64053...

http://forums.autodesk.com/t5/design-validate-document/unclear-how-to-create-relationship-between-an...

But somehow these steps do not work here.

[sachlene.singh]

Hi @Anonymous,

That sure is a neat looking assembly! I'm excited to see how it turns out with all the joints and motion working.

Contact sets seem like the right way to go, but with complicated profiles like gears or cams it gets tricky for Fusion to accurately and quickly calculate the motion, so it just limits a lot of motion that shouldn't be limited.

I have always worked around this by keeping it slightly simpler for the software. I add a 'slider' joint to pistons, with their cylinders/housing, two revolute joints at each end of the connecting rod and a rotating crank shaft of course. That usually accomplishes the motion I'm looking for.

To restrict motion of the pistons/valves, add some limits to the slider joint and that was Fusion knows exactly where to stop. That usually works well for me.

Let me know if this helps. Also,feel free to share the assembly if it doesn't and I can give it a shot.

[Anonymous]

Thanks for the reply, I'm not quite sure I can use this advise since the camshaft is not connected to the valves.

The pistons were already working fine with a slider and two revolute joints like you said, but I cannot work this way with the camshaft since nothing should be connected between these two elements.

They should only touch each other to push the valve down, no more than that.

Is there any other way I can make the valve react to the collision with the camshaft?

[Anonymous]

I tried again and again and again and finally it seems to be working a tiny bit.

As can be seen in the screencast Fusion is having a hard time getting to animate it, is there any way to improve the calculation speed?

My CPU is at about 95% for Fusion alone, so I guess it's purely the processing speed that matters? Or can I deactive some elements to increase the framerate?

[jeff_strater]

In my experience, contact sets only work for very small models.  I'm not surprised that you are seeing performance problems with a model of this size.  I don't think that deactivating components will help, to be honest.  Unfortunately, I feel that the best approach for this kind of mechanism is a Motion Study.  It's a bit tedious to set up, but if your goal is to simulate the engine function, this will get you there.

What is missing in Fusion, IMO, is some kind of "cam follower" joint.  This is the Idea Station item for this: cam-joint-mate.  Give it a vote!

Jeff

[Anonymous]

Thanks Jeff for the answer!

What exactly do you mean with "the best approach for this kind of mechanism is a Motion Study", isn't that what I've been doing or is there another way to accomplish this?

I voted for the idea as that is indeed what I'd need, but I'm afraid that won't come in the near future, will it?

I added the other valves and right now I have about 0.2 fps during the motion study.

Also, I tried a cloud render of the motion study, to see if the cloud thingy was better at calculating the movement, but I got exactly the same result back (but this time in video format instead of a motion study, hooray), does this mean that even with a very powerful machine, I won't be able to simulate this notch-follower motion?

[innovatenate]

I think @jeff_strater is referring to the Motion Study command located in the Assemble drop down menu.

I hope that helps!

Thanks,

[Anonymous]

I've been using the motion study, as you can see in the screen cast above.

I hoped there would be an option to let it simulate every frame in the background, and when it's finished it creates a video of it.

Right now it tries to play in realtime, which is not possible due to (for fusion at least) large amount of calculations, would there be a solution to let it calculate every frame in the background?

[jeff_strater]

There kind of is an "offline motion study render", if your design contains a motion study.  You need to use the Render command, set to Video.  

However, that is not what I was saying.  I was recommending staying away from contact sets at all for this design.  If it were me, I would manually simulate the mechanism entirely within a motion study.  That is, edit joint positions in the motion study so that they simulate the valve components following the cam, and not rely on contact sets at all.  Even if the cloud render works for the video, it will likely be very slow.

Jeff