Sketching 3d shapes with looking at an equation

1. Sep 27, 2008

korr2221

i was wondering what are some more popular ways to graph or get an idea of how a shape of an function would turn out to. the way i've learned it is to set certain variable to zero to get a certain plane. for example if i wanted the the xy plane of a function such as x^2+y^2-z^2 = 1 i would need to set the z = 0 and then solve for the y and solve for the x plugging random numbers to see how the shape is for that plane and do the same for the other two planes. which is the xz and yz. is there any other way to graph such an equation w/o using a computer and maple or something?

2. Sep 27, 2008

tiny-tim

Welcome to PF!

Hi korr2221! Welcome to PF!

Yes … this equation obviously has rotational symmetry …

so put x2 + y2 = r2, and you get:

r2 - z2 = 1 …

that's an easy 2D curve, which you can then rotate about the z-axis.

3. Sep 27, 2008

korr2221

Re: Welcome to PF!

sorry if i am incorrect, but isn't this just about the same as what i've said? i don't see any difference... unless you're trying to make a point that in this case because it's shape has symmetry the positive xy plane and -xy plane is th same because of it's shape but sometimes that isn't always the case? sorry i don't understand where you were going with you reply above...

4. Sep 27, 2008

tiny-tim

Hi korr2221!

No … your method, putting z = constant (i assume that's what you meant ), gives you horizontal slices, which as you remarked still leaves you a problem in deciding the shape.

My method uses a single vertical slice (well, half-slice), which you can just rotate.

5. Sep 27, 2008

korr2221

ah yes.... i sorta get what you've said now... but i still don't fully understand it... what 'phrase' should i google to find more about this method? what are you basically doing? turning the z plane into r^2? does your method always work? or only for this special case?

6. Sep 27, 2008

HallsofIvy

Staff Emeritus
One method I like to use is to sketch their cross sections in the coordinate planes. For example, in the xy-plane, z= 0 so x^2+y^2-z^2 = 1 becomes x^2+ y^2= 1, the unit circle. In the xz-plane, y= 0 so the equation becomes x^2- z^= 1, a hyperbola. Finally, in the yz-plane, x= 0 so the equation becomes y^2- z^2= 1, a hyperbola.

Sketch those on three separate pieces of paper and put them together to get an idea of the shape.

That together with drawing the "level curves":
When z= 0, that is x^2+ y^2= 1, a circle with center at (0,0,0) radius 1
When z= 1, that is x^2+ y^2- 1= 1 or x^2+ y^2= 2, two circles with centers at (0, 0, 1) and (0, 0, -1) with radius $\sqrt{2}$
etc.
which gives you a "geodesic map" idea of the surface should give a very good picture of the surface.

7. Sep 28, 2008

tiny-tim

Hi korr2221!

I can't think of anything to google …

it's really just a matter of simplifying

we started off with three variables (x,y,z), which gives a 3D image that is difficult to visualise …

I reduced it to two variables (r,z), which gives a familiar 2D image …

I suppose the general rule would be to try to reduce the number of variables while keeping the whole original equation! (which you didn't … you sliced it! ).

And no … sorry … it won't always work!

(you knew that didnt you? )