I'm thinking that what he might be doing is something like this:
 
1. Using 3 virtual audio outputs as sends for external inserts on each track.
2. They are mixed into a single virtual audio input. 
3. This input feeds a track containing Amplitube and outputs to another virtual audio output. 
4. This output is split into 3 virtual audio inputs...
5. ...which are the returns for each track's external insert.
 
 
Do I win the prize? 
 

drewfx1
 
Do I win the prize? 
  

At the point where you are talking "around" someone on an internet forum, you've borrowed from the playbook of a teenage girl and no, you do not win.   

This is where a "Splitter/Joiner" Pro Channel module would be amazing. 
 

200bpm
Craig, you are closest to understanding this, but this you do not get:  I am able to send 10,20, 30 separate DI tracks to a single instance of Amplitude because I am using Totalmix to make it seem like a single VST is a piece of hardware.

 
SONAR can send as many DI tracks as you want to a single VST, so I don't understand why you need Totalmix to do this. You can do this while recording from multiple DIs, or playing back the tracks created by multiple DI overdubs. Or from any number of arbitrary tracks that just happen to exist, regardless of how they got there.
 

It would be great if there was a "Insert effect" that you could set virtual I/O to/from a VST in the rack and route multiple tracks to it.

 
Basically, that's the definition of a bus, which you can set pre-fader if you want the level to the VST to be constant and depend solely on the send control and not the channel fader. You haven't specified where the output of this "insert effect" would go, but you can assign the VST output to a bus instead of the master and route it anywhere you want.
 
I can't help but think you're missing some very fundamental element of routing in SONAR that would make your life much easier if you knew it. That's why if only I knew the goal you want to accomplish instead of what you think you need in order to do what you want to do, I could be more helpful. Your goal is not really to have a "VST external insert," your goal is to achieve some particular routing to accomplish some particular recording goal which has not been explained or defined, but which you think would be solved by a "VST external insert."
 

The way FX inserts in sonar work are single input, but theres no reason multiple tracks cant be routed to a VST as long as they are not playing at the same time.

 
With what I've described, you can send multiple tracks to a single VST whether they're playing at the same time or not. I still don't understand how you claim to be able to vary the outputs of three separate tracks after being processed by a single instance of AmpliTube, which tends to support my concern that you're missing some very elemental aspect of signal flow. But hey, if you have something that works for you, run with it.

 
 
And I thought that re amping was as simple as running a dry guitar track out to an amp(s), miking the amp and re recording the resulting signal to a new track. I used to do this with 4 track recordings but those other mikes were recorded live to the mix down deck.  
 
 
 
 
 
 

Anderton

200bpm
Craig, you are closest to understanding this, but this you do not get:  I am able to send 10,20, 30 separate DI tracks to a single instance of Amplitude because I am using Totalmix to make it seem like a single VST is a piece of hardware.

 
SONAR can send as many DI tracks as you want to a single VST, so I don't understand why you need Totalmix to do this. You can do this while recording from multiple DIs, or playing back the tracks created by multiple DI overdubs. Or from any number of arbitrary tracks that just happen to exist, regardless of how they got there. 

Yes, this is correct.  But when you send multiple tracks to a bus with a VST, you mute the DI track and listen instead to the bus.  That is OK, but the channel faders don't do anything at that point and you cant vary the level of the individual tracks.  They are all "one" track at that point, and because they are going out the same output, there is no way to add different post effects to that many->"one" track.
 
 

200bpm

Anderton

200bpm
Craig, you are closest to understanding this, but this you do not get:  I am able to send 10,20, 30 separate DI tracks to a single instance of Amplitude because I am using Totalmix to make it seem like a single VST is a piece of hardware.

 
SONAR can send as many DI tracks as you want to a single VST, so I don't understand why you need Totalmix to do this. You can do this while recording from multiple DIs, or playing back the tracks created by multiple DI overdubs. Or from any number of arbitrary tracks that just happen to exist, regardless of how they got there. 

Yes, this is correct.  But when you send multiple tracks to a bus with a VST, you mute the DI track and listen instead to the bus.  That is OK, but the channel faders don't do anything at that point and you cant vary the level of the individual tracks.  They are all "one" track at that point, and because they are going out the same output, there is no way to add different post effects to that many->"one" track.

 
I get that, which is exactly how I explained it to John. 
 
To refresh your memory, I said: "Obviously he's not getting three separate audio streams through a single instance of AmpliTube 3." 
 
You said: "Yes I am.  Just took some creativity.  Sonar does not normally allow this but RME totalmix has some cool features."
 
What I don't get (and I also don't get why you are so adamant about not explaining it), is how you can send multiple inputs to AmpliTube and claim to have come up with some type of creative routing method that allows you to control the level of each audio input's output signal, individually through a channel fader, after it exits AmpliTube?
 
Can you please answer my question? It's also okay to say you thought that's what you had done, but on subsequent analysis, you realized you were wrong. It would also be okay to say you've come up with some ultra-high-frequency, clock-based time division multiplexing scheme that allows AmpliTube to process individual slices of the input signals on a time-divided basis, complemented by a time-division-based de-multiplexing system that follows AmpliTube and reassembles the slices back into individual streams, which the demultiplexer can then send to different channels for output level adjustment. Any one of these would explain what's going on. 

ok,  from reading all the posts this is what I understand.
 
He has 3 tracks with Bass recorded on them.
He's found a really kewl tone in Amplitube3 that he wants to use to re-amp the basses.
He doesn't believe his computer is powerful enough to have Amplitube3 in every FX bin.
He's decided the best solution is to put Amplitube3 in a bus but is struggling to route the bus output back to the original channel...
 
 
The CORRECT way is to put Amplitube3 in the track FX bin set the tone then bounce audio to clip/track, drag Amplitube3 from the track into the FX bin of the next track.  This ensures all the settings are exactly as you last touched them, even if you forgot to save.  This does not result in lots of copies of Amplitube3, it moves it from one track to the next.  Rinse and repeat for each track.
 
Result:  All tracks have the re-amped tone from Amplitube3 printed to the track as if it were recorded like that.  All tracks can be mixed/comped/edited/whatevered as normal with no overhead from Amplitube3, which can be deleted.
 
 
 
 
 
 
 
 
I think this is what I said back on page one...

What I don't understand is why do you have to mute the three tracks and "listen" to the buss"? If you mute the tracks you can't listen to the buss.  
 
200BPM read JohnnyV's post carefully it has some useful information in it. 
 
One thing I do know is the people that have been trying to help all know what they are talking about. They are very careful with the terms they use and have done all sorts of routing using Sonar and done it using hardware too. 
 
BTW it doesn't matter if its DI (direct input) or miced . Its just audio to Sonar. It couldn't care less how you got it into it (Sonar) and being DI doesn't make any different from any other audio.
 
Do not use the word "send" when what you mean is output.  Each actually has a meaning that uses different ways to rout.
 
I get that you don't want to use a tracks fader to adjust the volume because it will impact the signal of the tracks output. Mute is not the answer either. However you may be using the word mute differently than we do.   
 
If what you are trying to do is use the same FX on more than one track use a buss to have the FX in its FX bin which is an insert point. Graig is right about not being able to unmix once you mix the three tracks via a buss. 
 
We will still be here trying to help even though we have problems with clear communication. 
 

Ok, having read all this, it sounds to me like what he's doing is pretty reasonable - he wants to share 1 effect plugin across several tracks, and wants to be able to tweak the post-FX levels with those separate track controls.
 

 
(There could also be per-track pre-Amplitube FX as well, left off to save space.)
The thing that has confused most people, because it was seemingly only mentioned in the first post, is that there is "only one track feeding per time". So the fact that all 3 tracks are getting the same audio back from Amplitube doesn't matter - 2 of the 3 tracks are muted/archived. No need for time-sliced demultiplexing. ;)
 
If this was an external piece of hardware it would be trivial to do with the External Insert plugin, but there doesn't seem to be an equivalent for software. And since Sonar people don't think in terms of Send/Returns (as they only really have Sends), this usage doesn't seem 'normal'.
 
He could do the same thing via routing to a bus with Amplitube set as an insert - but then he doesn't get to have the 3 separate post-Amplitube fx bins and fader levels.
 
A workaround would be to have each track output to the single Amplitube bus, then that sends to 3 separate buses. But now he has to mentally tie his 3 input audio tracks to the 3 output buses. And his pre-Amplitube and post-Amplitube FX (if any) are in completely different places - an awkward workflow.
 
Some have mentioned that he could bounce each track - but that would stop him from experimenting with Amplitube settings in the context of tweaking the different pre- and post- fx and levels.
 
If needed to do this, in Sonar, I would probably just put everything in one track on separate take lanes. They're not really designed for that, but since we lost layers, that's the best we have. Clip automation would handle differences in gain and similar.
 
As I find quite often, REAPER appears to offer this sort of routing in the box - create a new track with your insert on it, then any track you want can Send to it and Return from it. I couldn't get this to work just now, but the functionality appears to exist there. I hope Sonar start to offer better routing options in future for this sort of thing.