Parametric Surface from f(x,y)

mahir

I'm trying to build upon the Parametric Curve FS to create a Parametric Surface. The intention is to be able to input a function Z = f(#x,#y) and have that create a lofted surface. However, I'm running into problems getting the z equation to accept 2 inputs (x,y). @ilya_baran, @kevin_o_toole_1, maybe one of you can take a look at my code to see where it's getting hosed? I tested the rest of the code, and it seems to work well. It's just the part where I try to calculate Z as a function of X/Y that isn't working too well. Thanks.

https://cad.onshape.com/documents/57a4e196e4b04351c9604db5/w/9e4f338e07807b410d29e1d9/e/95f3ca6fb50c41817cfd6cae

kevin_o_toole_1

The error you're getting (Cannot call a function that takes 1 arguments with 2) is because z in an Expression, and Ilya defined Expressions to be functions of a single input.

Since you want functions of two inputs, great. You'll just need to change all the overloads defined to use two arguments like (x, y) instead of (t), everywhere in the file (the "all" button in the ctrl-f find dialog may help you )

Then x_ should be function(x, y) { return x; }, and y_ should be function(x, y) { return y; } 

Perhaps you'll want to rename Expression to something less generic like BiExpression to better communicate that it won't work with other numbers of parameters.

ilya_baran

My original thought for multiple variables was to have the input be say a 2 element array and x_ be function(t) { return t[0]; } which lets the original Expression type be kept.  But Kevin's approach should also work.

kevin_o_toole_1

Ooh, yes. Ilya's approach is better. Do that instead

mahir

Thanks for the input, guys. I implemented @ilya_baran's array method, and it worked perfectly

mahir

On a related note, I'm having trouble with order of operations when overloading the unary (-) minus operator. Right now, -#x^2 is evaluated as (-#x)^2 instead of -(#x^2). Not a big deal, but most people would assume PEMDAS would be followed without extra parentheses.

mahir

I retract. After a little research I realized that unary operators take precedence exponentiation. Live and learn.

mahir

Here's the finished product.

https://cad.onshape.com/documents/57a62df8e4b03de4c9fcf154/w/f6d71f3d862b522902efcd68/e/ece65382fcd820281a9ee704

lemon1324

Nice! There's some ripples from the interpolation between profiles, but that can't be avoided.

mahir

Yeah, you can increase the number of curves, but the computation time scales as O(n^2). Good enough for now, I suppose.