Thank you all for the tips and info.  Much appreciated.
 
I've been doing some research on buiding a home recording studio, and have learned that my RealTek onboard audio isn't going to cut it for my needs.  It's fine for gaming, but not for what I want to do as a musician.  I want to future-proof my setup now, but I'm on a meager budget, so I'm getting a Focusrite Solo and a SoundBlaster Audigy RX.
 
@bitflipper, thanks for the tip about the TTS-1.  Unfortunately, SONAR's online doc and what few YouTube videos I've seen about using softsynths has left me even more confused, so I have a couple of questions.
 
1) Once I've recorded my 220 with the TTS-1, how do I convert it to audio data that SONAR can export?  I've tried exporting directly to Audio, but SONAR returns this error message:
 
"The specified selection did not contain any audio data.  Please ensure that the selection contains audio data you wish to bounce.
 
There's a section in SONAR's Knowledge Base that covers three different ways to convert MIDI data to audio (Google sonar converting midi to audio, it's the second search result).
 
The third method outlines your tip, but the directions aren't very clear.  On Step 3, it says to route the MIDI tracks to the appropriate output, and talks about the easiest approach it suggests.  After following its directions, I still get the error message I referenced above, which leads me to my second question:
 
2) Do I need to set up an audio track in conjunction with the TTS-1?
 
To that point, there's a function in SONAR called Bounce Clip(s).  Nothing I've learned has been very clear about what that function does or how it's supposed to work, so I've been trying to experiment in an effort to stumble on a solution to my problem.  When I bounce the TTS-1 track to Clip(s), nothing seems to happen.  I've tried inserting an audio track and selecting the TTS as its input, and then bouncing the TTS to Clip(s), but nothing happens.  I finally get a dialogue box when exporting to Audio, but the MP3 I saved has no sound.
 
Cakewalk should consider creating a series of video tutorials for beginners like myself.

Work through tutorial #4.
All soft synths including TTS-1 must use an instrument track or at least one audio and MIDI track. Recently, this was discussed at length here.
 
There are a series of beginner videos created for Music Creator. MC7 was  product based on SONAR X3. They are found here. They are still useful, even the older MC series videos. The user interface resembles the current version of SONAR.

First- you don't need the sound blaster- it may actually cause conflicts with the Solo's drivers,, don't install it. You can use the solo for all your audio and with the option of your on board for other things . Creative drivers are well known to screw things up. 
 
Your export didn't work becuase you need to select all the tracks you want to export. So CTR A before export. 

I'm getting closer to solving my export problem.  Here's what I've figured out so far:
 
1. Insert a General MIDI track using the TTS.  Set its Input to my 220, and its Output to the TTS.
2. Insert an Audio track and set its Input to the TTS.
 
At this point, I've confirmed that both tracks are receiving input when I play a few notes, however, I've encountered another snag that I'll describe in a moment.  Next:
 
3. Arm the Audio track to record, and record a few notes.
4. As per Cactus' advice, I select both tracks and export the Audio to an MP3.
5. I play back the MP3 in iTunes and hear the sweet sound of the notes I played.  SUCCESS!
 
So.  Now I can finally record my performance and export it to MP3.  Yay me.  Unfortunately, I can't hear anything while I play.  I can fix that by setting the TTS' output to the Microsoft MG Wavetable Synth, but then the Audio track goes deaf.  I thought I could simply rely on my 220's built-in speakers while recording, but I've also discovered that it's rendered silent when I connect it to my PC.

Instead of recording the audio from the TTS-1, record the MIDI track. It be rendered to audio anytime.
Also use input echo on the MIDI track to hear it play before and during recording. This will rely on connecting speakers or headphones to the Focusrite Solo (or the PC if using the internal sound chip) and routing SONAR audio output to the interface or PC. Hardware setup instructions are here.
 
Edit: I see you already have SONAR configured to use PC sound chip for audio.

To hear sound while you're recording, make sure the "input" button is clicked on the track you are recording. That way SONAR will feed the "source" (TTS1) audio back out through your interface and into your speakers (Note: If you ever do this with a microphone, use headphones instead of speakers). If you are still using the computer's built-in sound card, there might be quite a long delay between playing your keyboard and hearing the sound. That will be fixed with an ASIO interface.
Disclaimer: I'm a guitar player, not a MIDI expert. I hope I've identified your problem, but that's about as much as I've ever learned about the black art of MIDI recording.

scook
Instead of recording the audio from the TTS-1, record the MIDI track. It be rendered to audio anytime.

 
That hasn't been my experience, unless I'm mucking something up.  Please define "rendered" as it applies to SONAR.
 

Larry Jones
To hear sound while you're recording, make sure the "input" button is clicked on the track you are recording.

 
What input button are you referring to?
 

Larry Jones
If you are still using the computer's built-in sound card, there might be quite a long delay between playing your keyboard and hearing the sound. That will be fixed with an ASIO interface.

 
There's actually virtually no delay.  Maybe a few milliseconds.
 

Larry Jones
Disclaimer: I'm a guitar player, not a MIDI expert. I hope I've identified your problem, but that's about as much as I've ever learned about the black art of MIDI recording.

Sadly, no dice yet, but you're right about this process feeling like a dark art.  I can't believe how complicated it's proving.
 
Just to clarify for everyone, I haven't gotten any extra hardware yet.  I'm still using my onboard audio.

MIDI will be converted to audio on export, just have to make sure the appropriate tracks are selected (or when no tracks are selected everything will export). Once recorded, the synth may be frozen or the track bounced to create audio clips in the project. Tutorial #9 covers export. Here are some help pages that cover these topics
http://www.cakewalk.com/Documentation?product=SONAR&language=3&help=Mixing.06.html
http://www.cakewalk.com/Documentation?product=SONAR&language=3&help=Mixing.24.html
http://www.cakewalk.com/Documentation?product=SONAR&language=3&help=SoftSynths.12.html
 

A couple of more things.
 
Do yourself a favor and disable the Microsoft GS Wavetable synth in SONAR preferences. Use TTS-1 instead (or any other soft synth). When SONAR has no MIDI output defined in preferences, the program will automatically add and assign TTS-1 appropriately when opening MIDI files. Remember to save the MIDI file as a Cakewalk project if you wish to retain any changes not related to MIDI such as FX, automation, audio and synth assignments other than the default TTS-1 assignments. T
 
Always use a Master bus. The default Basic template bundled with SONAR is configured to use one.
 
 

When in doubt, listen to scook, not me. But here's info about the input monitor button. I've never found a clear, simple diagram of where all the controls are in the SONAR interface, maybe because it's so customizable. But the link will at least show you what the button looks like. It's in the control area to the left of the track.