The problem I see here is that there a several ways to accomplish this, but you haven't given us enough information to determine which one way you intend - i.e. what your purposes are.
I'm not sure what the reference to the Mackie mixer purports - it's not a computer audio interface, you'd need another device in order to get synthesizer's the audio back into the computer. Also, although the XP-80 can play multiple parts simultaneously, it only has a single stereo out.
Most folks would forego the Mackie in a recording situation, it's not needed and can only get in the way. You'd capture audio for each instrument to be played through the XP-80 individually, each recording going to a different audio track. Send the output of these audio tracks to a bus. Adjusting the output level of that bus affects all those tracks at once. Alternatively, you could group the faders of those audio tracks and any adjustment to one of them affects them all.
Now in a midi-playback situation, say in a live performance, you aren't at the mixer. Any change in volume there is done by the synthesizer itself in response to a change in midi volume message (assuming the FOH leaves the audio return from the XP-80 alone). This is, I assume, rarely done. (You'd normally, I expect, modify your keyboard playing, i.e., less key velocity...) To do it in Sonar, you'd simply group the midi volume faders together, and any adjustment to them would affect them all. Naturally, you'd have to have write automation enabled and the volume parameter showing in each of these midi tracks for sonar to record your midi volume changes.
The short answer is, most folks do the volume changes in the audio mix. For recording, do it in Sonar by using a bus.