This method relies heavily on the use of the dot product. Let's assume that we know one point that lies on the plane. Let's call this point P = (x0, y0, z0). Let's also assume that we know the plane is perpendicular to the vector .
What do we know that we can use to find the equation of the plane? Notice that if we find a vector that lies in the plane, it must be perpendicular to since the plane and the normal vector are perpendicular. Thus, the dot product of these two vectors is zero. This is an equation, if only we can find a vector in the plane.
Suppose that the point Q = (x,y,z) also lies on the plane. The displacement vector from P to Q is then a vector in the plane. The dot product of this vector, , and is zero, so this will be our equation.
If we rearrange this equation, moving all of the constants to one side, we get where b = (n1 x0 + n2 y0 + n3 z0)/n3. Note that in the first formula in the line above, the coefficients of x, y and z are the components of the normal vector. This is true for any plane whose formula is put into the form above. For example, the plane z = x +(1/3) y - 1 has normal vector . This can be seen easily after putting the equation into the form 0 = x + (1/3) y - z - 1.As another example, what is the equation of the plane passing through the point (-1,1,0) with normal vector ? From the normal vector, we know immediately that the equation has the form
x - y + 2z = b.
By plugging in the point, we can compute b as b = (-1) + (1) + 2(0) = 0. Thus the equation for this plane is x - y + 2z = 0.Note that there are many normal vectors to a plane. Multiplying a vector by a scalar only changes the length (and possibly orientation if the scalar is negative) of a vector. Thus, , , and are all normal to the same plane.