Goertzel Algorithm Implementation For Fast DFT Of A Few Bands?

I'm working on an application where I need to analyze a sampled wave for the presence and power level of a handful (6 for now) of specific frequencies. For the last several weeks I've been reading up on FFT, and even implemented a .NET FFT algorithm by translating over some VB 6 code I found. Unfortunately, I don't think that the code I translated is working, because it only gives me 150 some elements. (My array of samples has 512 elements, so after FFT, I should have 256, right?) I've tried to debug it, but because I don't understand the math behind it, it's been difficult.

I came to realize that the Goertzel algorithm may be better for my application. I understand it as a fast way to get the power level of a specific band found in the result of a DFT, without all the math necessary for DFT or FFT, simplifying the problem, and saving a lot of CPU cycles all at the same time.On that site there is some sample code for implementation:

Code:
realW = 2.0*cos(2.0*pi*k/N);
imagW = sin(2.0*pi*k/N);

[code]....

Note that in that code, eventually I intend to be able to pass a frequency in Hz, and the current sample rate, but for now, the freq parameter is just what band I want, and samplerate is ignored. In my current FFT, I get values from 0 to 1, but with this code they can be as high as hundreds of millions.