For each midi track, there is a Midi Output Channel that you can set. Leave the input set to None (omni).
Sonar will then pick up the midi input from whatever controller you are playing, and for a given track will then route the midi back to TTS-1 on the midi output channel you set for that track.
So, if you have Piano and Bass on TTS-1's first two instruments, they will be set automatically in TTS-1 to channels 1 and 2. When you then go to your two associated midi tracks, you would set the midi output channel for the piano midi track to be channel 1, and the midi output channel for the bass midi track to midi channel 2, with everything routed to TTS-1.
Then if you select the Piano midi track, Sonar will by default auto-enable echo on and when you play your midi controller, Sonar will route that output to TTS-1, on midi channel 1, and you should hear piano.
Doing the above for the Bass midi track should then result in TTS-1 playing the bass sound on midi channel 2.
Bob Bone