Like Don I feel there is some confusion with using a grey or white card to set the exposure and colour balance.
Using a tool such as
http://www.speedgraphic.co.uk/meter_...e/16356_p.html can be used to set the camera exposure and in post processing check and correct the brightness and contrast levels. This is because the steps in grey level are know on the target and the image can be adjusted in post processing to match the the target. Thereby getting the image to match the scene.
Colour balance is about removing colour casts caused by different temperature lighting conditions.
In either case the RGB values of various parts of the image need to be measured and adjusted as required to the desired value.
Norm it good that you are attempting to use reference targets such as the web wheel and grey and white cards. However, I am somewhat confused by your methods in post processing.
You don't say what editing software your using. If it limited to just editing brightness, contrast and saturation then that may go some way to explain your methods. It sound like you are adjusting by eye rather than using your editor as an instrument to measure and adjust.
To start with your method of scanning the white card for a reference to compare against your image is flawed. Your scanner is going to introduce one set of distortions and your camera another set. You will never know which is nearest to being correct.
Your photo has a number of know reference points so make use of them and measure these points and correct accordingly.
On one of your recent posts with the grey step wedge it has a yellowish cast. I have used Photoshops colour balance tool to make some crude adjustments based on reading from the shadow, mid tone and highlight areas of the grey scale wedge. If the grey step levels were know (i.e one stop) then the contrast could have be adjusted to match the target.
On your last post the white balance looks good because the camera had a large area of neutral tone to work with. The thing to remember is that digital camera Auto White balance algorithms are not fool proof and normally work well within a limited range. They do not cope well with mixed light sources.