I went through the same thing. Here's my thoughts, which others of deeper understanding can correct if I'm wrong...
Generally, when using it for effects processing, you will set up a bus and put some effect on it. But you'll mute the dry signal. E.g. if it's a reverb, you'll mute the dry output of the effect and just output the reverb tail stuff. So you would use the output of the track to send the main part of the signal, then you'd do a Send to send it to one or more effects buses, which will only add the effect output to it.
This way, you can control the dry signal of the track separately (via the track fader) from the amount of f/x you are adding to it via the send (the level of the send.) So you would use them both, but for different reasons. Leave the f/x buss fader at 0dB and use the send level of each track to adjust the amount of effect that track will get (the hotter the signal from that track that you push through it, the hotter the reverb tail - or whatever is is - will be.)
You'd normally turn on the Post button, which means that the level of the send is also affected by the fader level. So you use the send to control how much effect you are adding to the track, but then it also will go up or down with the level of the track (which you'd probably want since the louder the track the more of the effect you'd need to maintain the relative volumes of the two paths.) But you might end up adjusting the fader and then adjusting the send level if you want to change that relationship.
So it might look something like:
track (output) -> master bus (output) -> main faders
(send) -> reverb bus (output) -> master bus (output) -> main faders
(send) -> chorus bus (output) -> master bus (output) -> main faders
post edited by droddey - 2007/04/29 22:19:11