topic Re: Roller bearing motion in Fusion Design, Validate & Document Forum

Roller bearing motion

joergen_nilsson — Tue, 13 Aug 2024 12:41:10 GMT

I tried to simulate a correct roller bearing motion with no success. How do I configure my joints letting all the balls rotating around the inner ring at the same time as the outer ring rotates relative to the inner ring?

Further, I tried to break the links to the bear balls and get an error message due to the circular pattern. But I'm not able to remove the circular pattern either.

Seems to be a pretty advance motion to simulate and I appreciate all help I can get.

Re: Roller bearing motion

TrippyLighting — Tue, 13 Aug 2024 12:54:59 GMT

Motion Links and a little math.

If you can share the model we can be more specific.

Re: Roller bearing motion

Drewpan — Wed, 14 Aug 2024 00:10:27 GMT

Hi,

How accurate do you want this to be and for what purpose? If it is a "nice simple" learning exercise then you are about

to find out there is nothing nice or simple about it. If you want to simulate a roller bearing in a model then a simple

rotating joint is all you need.

Without seeing your model it is a bit hard but I don't think a circular pattern is going to work. I think you may have to

model each bearing and as @TrippyLighting suggested, use some maths to work out each individual motion link on

each of the bearings. I should imagine it will be messy and time consuming.

You will first have to work out how one bearing will work. Then you will have to work out how adding another one will

fit and if that will work or clash somehow. You will also probably have to do this for many many rotations because the

smallest error will cause a crash that is not necessarily obvious with a single rotation. The you have to repeat this for

each and every bearing you add till you are done.

I think that this particular nice simple learning exercise will end up being too hard and complicated. A little like a

computer, all ones and zeros - simple right? If it is that simple then build me one step by step so I can understand it.

Not so simple.

Cheers

Andrew

Re: Roller bearing motion

jhackney1972 — Wed, 14 Aug 2024 00:51:07 GMT

You are mixing your types of bearing, one place you say Roller Bearing then in another your talk about Balls in your bearing. I went with the Ball Bearing. You can cut your work in half by downloading a ball bearing from McMaster-Carr that has ball cages. You can used them to quickly create your required joints and then hide them in the assembly. A Motion Link to half the rotation speed between the race and the balls makes it more realistic. I have attached my model.

Re: Roller bearing motion

joergen_nilsson — Wed, 14 Aug 2024 06:13:46 GMT

Ok, my intention was just to teach myself and my students more about joints and I thougt this would be a nice challenge.

Re: Roller bearing motion

joergen_nilsson — Wed, 14 Aug 2024 06:16:18 GMT

Yes, english is not my native language and I know I mixed up roller bearing with ball bearing but the problem should be similar no matter what. I will study your model and come back laster.

Re: Roller bearing motion

Drewpan — Wed, 14 Aug 2024 07:20:46 GMT

Hi,

I would strongly recommend that you do the embedded tutorials in the Fusion Documentation and also some of the

Self-Paced Learning to help you to learn fusion faster and better. They can be found here:

It is also much easier for the forum to help you if you attach your file AND a screenshot of what you want to achieve

and what the problem is. You can create a file to export like this:

Time spent on the tutorials and self paced learning will not be wasted.

When designing, it is much simpler to use a simple joint instead of an assembly. While @jhackney1972 has a good point

about inserting a McMaster-Carr component if you were actually designing a more complex assembly like a drive it is

a good solution. The problem with inserting many of these external parts is that it can cause fusion to lose

performance. It is similar in say, modeling the threads on a fastener, you can do it but it will affect performance.

The question to ask yourself is what am I trying to achieve? I can model everything right down to the washer on a bolt

and it will work and I can test it. Or I can assume that a bolt is going to be used but just use a simple joint instead to

represent what the bolt and washer assembly will do. I mention this because having a full bearing like the one that

John has used is representative of the real world but all you wanted to demonstrate was a Revolute joint.

For just learning about joints, I would keep it simple to start with. Once you get your head around it then set them

homework to model a bearing and see what happens.

Cheers

Andrew

Re: Roller bearing motion

joergen_nilsson — Wed, 14 Aug 2024 09:14:13 GMT

Yes, I get that.

But it would be interesting to see a more general way to simulate the rollers/balls rotate around their own axle while at the same time rotating around the bearing centre. Almost similar to how the planets move around the sun.

Re: Roller bearing motion

TrippyLighting — Wed, 14 Aug 2024 13:10:07 GMT

The motion @jhackney1972's model shows is only half of the motion of the ball bearing.

It shows how the outter bearing race moves around the inner race. It shows how the balls move around the center of the bearing race(s)

It des not show the actual rotation of the bearing balls around their center-of-mass axis.

The latter is a bit more difficult to achieve with that particular model. Usually the origin of a sphere would be centered in the sphere. However in this case that isn't so, so the model requires some modifications.

The blue highlighted sphere is the geometry of the ball component and it isn't centered around the origin.

Re: Roller bearing motion

MichaelT_123 — Thu, 15 Aug 2024 04:50:34 GMT

Hi Mr Joergen_Nilsson,

The subject is an interesting one ... particularly if one wants to look deeply into the intricacy of this wonderful mechanism which behaves "Almost similar to how the planets move around the sun"

As some Interlocutors pointed out the (F360) model should reflect what you try to investigate/study.

Planetary movement ???
Yes, it is the most common approximation of bearing kinematics, however if the groove geometry in which balls roll is taken into account, it is easy to discover that the trochoidal movement of the surfaces of balls/tracks is compromised as their radii ratios changes along a groove height.
The mathematics of the calculation of the slippage should not be too complicated. The Boston MIT entry exam recqs should do ... Mr Drewpan.
But there is more to ball movements. Their rotation/roll axes are not strictly aligned with a bearing axis. Slight imperfections of inner/outer grooves or external forces make them wobble/process semi-randomly … so they wear out mostly uniformly; one of not too many examples when imperfections are beneficial.
Hence, depending on your curiosity… our CAD model should reflect the above … and perhaps more.

Regards
MichaelT