* Update *
 
Since I cannot reproduce the hang reliably, I decided to work on the one other issue I recall encountering before, being multiple projects open and trying to play them each back and forth.  Instant freakout, CPU spike, etc.  Upon much investigation, it turns out to be a combination of 2, and possibly 3, things.  1) Live Synth Pro 2) Having at least 1 of the projects with BOTH Superior Drummer 1 and 2 loaded and 3) MIDI Yoke input selected on a track (even though there is nothing effectively being input).
 
Need to do more experimenting with various permutations of the above to try and learn more detail as to what is causing that problem.  I was aware LSP can be picky on multi processor machines.  But within a single project I hadn't encountered issues as long as I kept the MIDI channels separated.  Clearly though, there's a very strange dynamic going on with respect to these particular synths in the Synth Rack with multiple projects open.  So this may be a clue that might help lead me to figuring out what my main hang issue is all about.  Hopefully....
 
Again, thanks for anybody still reading, and please keep any suggestions coming.....
 
Sonic

Here's another update.  I noticed I was crashing upon close of a project at times and tracked down the cause, which leads me to believe all this may be a Sonar 5 problem.
 
1) Open Project
2) Open and close ANY plug within the console view that does NOT result in Sonar acknowledging a change transpired (no asterisk next to the file name in the title bar)
3) Close the project .....and  CRASH!!!   (Normally it should ask if you want to save changes, but Sonar apparently doesn't realize a change was made, indicated by lack of asterisk in title bar, and crashes)
 
For example this can easily be replicated by simply opening & closing the Sonitus EQ on a track, then closing the project.
 
Leaves me wondering if my problems are simply a buggy S5 design or if there's some strange system interaction going on. 
 
Sonic

Well, no such luck on any of the above clues.  It seems they are entirely unrelated to the hang.  I basically ripped everything out of the Synth Rack except for one drum synth and yet still got a hang (actually twice, alternated between Superior 1 and Superior 2 and got hangs both ways).  I also noticed at the time of hang the CPU meter in Sonar was 0% yet the CPU meter in Task Manager was registering 50%.
 
I'm just about of ideas at this point....
 
Sonic

Did you try enabling (or disabling) read & write-caching? Both inside SONAR & Windows Control Panel for your HDD? And write-cache buffer flushing? In Device Manager for your HDD?
How about just hitting Ctrl>S & waiting a bit prior to closing SONAR? Does that help?

Thanks for the suggestions!
 
I was planning on doing exactly that today, because i have write caching enabled inside Sonar, but not read. So I will try that, and double-check the Control Panel and Device Mgr.   I did however already try waiting on the hang for about 10 minutes, it never returned.  But I have no idea about Ctrl>S is....never heard of that - but will try that as well...

UPDATE:  Ok, so I set the Read Caching in Sonar....but after reading the help file I learned something NOT obvious yet critical: The cache settings don't take effect until after you restart Sonar, even though you get no warning to that effect.  So....any testing I had done previously with these settings was bogus because I would not have restarted Sonar.   Also, I double checked that write caching was already enabled on the OS side (DevMgr).
 
Thus far I have noticed with Read Caching enabled in Sonar that disk usage has gone to ZERO in PerfMon.  So, should the hang go away would that mean we are just masking the problem??  I mean, what would this tell us, that there's a disk I/O quirk?  Which I find doubtful because if there were it would almost surely be more broad-based and not isolated to this one specific hang condition.
 
Anyway, I'll pound on this new config (Read Cache enabled in Sonar) to try and get it to hang and report back....

CTRL+S for saving the Project prior to closing it. Worth a try.

Another couple of things that come to mind:
1. In your motherboard's BIOS there may be more low-level options for the disk that can make a significant impact.
2. I do recall that replacing my ATAPI/IDE HDDs with SATA made a huge difference (as long as the motherboard has the SATA port-I think you have to slip-stream the SATA driver during the WinXP installation so that may not help in this case).
 
The usual tips:
A log off/log on may be necessary after making changes in DevMgr. Reboot is even better.
Defrag your disk.

Don't forget to read all the WinXP Audio Tweak Guides you can, i.e. this is vital:
 
Enable DMA mode for all hard disk drives (WinXP):
1. Check to make sure the transfer mode for your hard disk drive(s) is set to the DMA mode, otherwise the disk access speed will be poor.
2. Go to Control Panel > System > Hardware tab > Device Manager button. Expand the entry labeled IDE ATA/ATAPI controllers. Right click on the Primary IDE Channel item and select Properties and go to the Advanced Settings tab.
3. Check that the Transfer Mode options are set to DMA if available for all channels. Repeat this for all IDE channel items beneath IDE ATA/ATAPI controllers.

Ok, so read/write cache enabled/disabled inside Sonar made no difference on the hang issue.
 
There's no low-level tweaks in the BIOS for disk besides choosing IDE/SATA. Very basic BIOS, intel G41 board.  All the latest updates too.  And i already did ALL the things mentioned above.  One thing I did notice however is that Windows is showing the main OS disk on the IDE Primary Channel, while the Audio disk is on the IDE Secondary channel (along with the CD/DVD drive).  Not sure if that matters WRT this hang problem, but I really doubt it.
 
The hang seems to always be when I stop the transport.  I'm still trying to find some way to replicate the problem, it happens only once in 100+ times.  Seems like it might be related to the display possibly, mouse action or something like that.   I'm right now testing on a project where I ripped most all the plugs out to see if I can still duplicate the hang.   It's really maddening to troubleshoot with no means to reliably replicate the problem.... 

Didn't realize that you had 2 HDDs. Where is SONAR installed? Where is your audio data? I always had problems when the audio data is physically separate from the Cakewalk program files. (Been doing this since mid-1990s).
This is important:
If SONAR is accessing anything on the same IDE channel as the CD/DVD drive, you'll want to find the patch that stops WinXP from checking every few seconds if there is a CD in the drive. Better yet let the CD/DVD drive live by itself on that channel, and put Cakewalk program files & audio dir on Primary. I'm surprised it only happens 1/100 if indeed the audio disk is on the same IDE channel as the CD/DVD.

Ok, I ripped out all the plugs and still hang.  The only thing I have left is ONE DXi wrapped in Directixer (Superior Drummer).  Fails with either Superior 1 or Superior 2 so I don't think the DXi is the issue, unless it is the wrapper (Directixer) that is responsible.  But I have no safe/easy way to eliminate that variable at the moment. I seriously doubt Directixer is the issue.
 
Next I'm going to try a painful start/stop test sequence - but not moving the mouse away from the focus of the area where the transport buttons are and see if that makes a difference.
 
I'm a bit suspect about MIDIYoke being loaded, even though it's not being used, so to speak.  Not sure if maybe I should risk trying to disable it in the DevMgr Devices view....
 
Yes, I have separate OS and Audio disk, always ran that way before no problem.  But that was on a W2K single-CPU/core system.   
 
I just compared another hang dump and it's literally almost identical, whatever is happening is likely the same thing, at least that much is encouraging - meaning it is indeed ONE CONSISTENT problem.
 
I could re-locate the audio disk onto the same Primary IDE Channel as the OS disk as long as it isn't going to confuse Windows and cause a domino effect.  My original thinking was to make use of the separate IDE Channels to reduce any potential conflicts and avoid the two disks competing for use of the same IDE Channel.  Was that a bad decision?
 
Sonic