Multi turn wave spring 2

[Wuzhee]

Hi Everyone,

I'd like to model a multi turn wavespring for a test project.
The only info I found is a pdf catalog containing some basic knowledge. Is it possible to model the wave spring using the catalog data?

I haven't come across Curve from Equation or Laws before, so this is completely unknown territory for me.
My first approach was making an equation for the spline of the wave then project the curvature to a cylindrical surface then sweeping the section. After some math stuff I got stuck with the equation (calculus isn't my strength).

Here's the spring parameters:
Free height (FH): 5,72 mm
Housing dia (Dh): 18 mm
Shaft dia (Ds): 13 mm
PCD: 15,5 mm
Thickness: 0,3 mm
Width: 1,83 mm
Waves: 3,5
Turns: 3

Can anybody post a possible solution, or some tips where to proceed?

 

[jackk]

You might be able to find a similar spring online, and then download a 3D CAD model, and even purchase the part. (much faster and easier than modeling it yourself)

https://www.mcmaster.com/wave-springs/stacked-wave-disc-springs-7/system-of-measurement~metric/

 

[Wuzhee]

My goal isn't the final product itself, but to learn and apply the methods and know-how to create complex shapes like these wave springs.

 

[3DDave]

I'm not a Catia user, but if it has the ability to create a 3D curve by equation then that's where I would start. It is easiest if there is a Cylindrical coordinate system - R, Theta, Z.

Typically there is a variable that goes from 0 to 1; usually "t".

R will be constant and represent the radius to the centerline of the rectangular section of the spring.

Theta will be Number of turns * 2 * pi * t, so as t goes from 0 to 1, the turns on the curve will go around Number of turns times.

Z is the tricky one for this case. You want a number of Waves per Turn, but also a generally helical advance.

So Z will be Finished Height * t + sin(Theta * Waves per Turn) (Finished * t is the helix); (sin() is the wave)

If you need a guide curve to create the sweep section, create it with exactly the same equation, but at a small increment to R so that it is horizontally offset.

Making it flat on the ends requires more than I can recall off the top of my head, though just grafting on little pieces at the ends might be easier.

 

[LWolf]

Use "Curve from Equations" found in GSD workbench.
x- and y-laws are used to define the circle y=Radius*cos(2*PI*X*1rad*4), y=Radius*sin(2*PI*X*1rad*4)
whereas z-law will define the helix, as 3DDave mentions:
y= Height*x+sin(2*PI*x*1rad*`number of waves per turn`*4)

regards,
LWolf

 

[Wuzhee]

Hi LWolf,

This looks excellent. Thanks for your visual explanation.
What's the 'X' between the () in the equation? (2*PI*X*1rad*4) Is it the radius of the pitch circle? Or it's a variable?
How should I create the law? Do I need to make a formal parameter named X (real)?

 

[LWolf]

both x and y are real parameters, created within law definition dialog box

regards,
LWolf

 

[Wuzhee]

With my parameters input I got this curve.
However on a real wave spring the curve cannot intersect so I have to flatten out the waves so that they just touch tangentially (taken into account the spring thickness which is 0,3 mm).
Also, changing the wave numbers to a real number makes a nice curve, but my springs parameters are given.

 

[Wuzhee]

Finally made the spring look nice. It was just basic high school math, nothing more
Any number <1 makes the sine wave flat, so I experimented with different values. The top and bottom shims are parameterized and I manually adjusted the wave amplitude and height to the nearest extent. The start and finish of the curve is not optimized but I don't really know or bother to make it perfect.

Here's the spring with a helical advance:

And here's the spring made out of planar wavy functions. In this instance the wave count must be integer otherwise there would be a break in the curve.

Thanks to 3DDave and LWolf for providing their knowledge on the topic.

 

[LWolf]

So, what parametervalues did you use?

regards,
LWolf

 

[Wuzhee]

I just fiddled with the multiplier until I got a good enough result. X and Y are as you wrote in your upper comment.
It's for myself studying and also to represent a possible design for a current problem in our product.