First of all, make a drawing of ALL your MIDI in/out/thrus on ALL of your MIDI gear that you want to use. Just make some crude rectangles, and draw little circles, naming the gear and the circles. Starting with the computer going to the main ports of the MIDIMan I/O, see how many things you can connect using in/out/thru. At this point, since you want to sync to TASCAM, it would probably be best to connect this first in the chain OUT of a MIDIMan port (probably 1), to avoid inevitable MIDI delays.
By the time you read through this next (lengthy) part, you MAY realize that you may have to swap out a thing or two by the time you finish. You MAY just not have enough equipment to do everything easily, although you can probably, eventually, use everything.
I THINK this is getting closer to a solution, but you have a LOT of outputs from the keyboard, drum machine, computer, Nanosynth, tape deck, etc.
Do you have a mixer, other than the one in the TASCAM? Eventually, I suspect, all of this is going to have to make it either into the computer, the tape, or both. And then, finally, into the computer for final mixing to a stereo file. I'm guessing you only have in your computer a soundcard that will only allow stereo recording through a mic/line input (of which you should have set to line in).
If the TASCAM can access all 8 inputs at once, and can mix to a stereo output, there are several things you can use that for. If it can, and you dump all 8 outputs of the drum machine on it, the tracks will filled, forcing you to dump them into the computer. You COULD use it to record the drums to record to the computer as a stereo track, but this would probably be redundant, and just add noise. Could you use the TASCAM as an 8-in/ stereo-out mixer, to submix the drums? In this way, you could theoretically run, say, the different drum outs through different outboard processing (compression, reverb, etc.), into the TASCAM, mix, and output to record in stereo. Does the drum machine internally offer a stereo out mix?
As you can see, there are a lot of options already.
Personally, I would use the TASCAM for submixing to stereo (if it's possible, barring another mixer), and only use tape maybe for non-electronic sources, such as vocals, guitar, etc. The drum machine, Nanosynth, keyboard, etc., I would probably run directly from their outputs to the computer for the crispest, cleanest sounds.
SO, I don't know how you are going to monitor all those outputs, unless you have another mixer with a lot of inputs. It may be worthwhile to invest a couple hundred dollars in a small mixer. Of course, then you'll wanna hear it properly, so an amp and decent monitor speakers will eventually come into play, but you could theoretically just do it all through headphones or a home stereo/speakers, though it probably won't translate well to a final mix.
Anyway, if you end up putting ANYthing on the tape, what I would do is don't put audio into the computer until AFTER anything on the tape has been submixed and recorded to the computer, unless it's clear that the tape will follow the computer properly. The only reason I can get away with dumping all 7 tracks (track 8 is SMPTE) of my 80-8, is that I do it all at once through the Delta 1010 with 8 A/D inputs.
Now, the Nanosynth, keyboard, and drum machine will follow Cakewalk using Sync to Audio and outputting MIDI sync, so you can wait until late in the project to record them in as audio. I would just monitor a basic MIDI file of the project through the soundcard using it's instruments. With the tape acting as Master (MIDI should follow properly), record all the tracks you want onto tape. This way, the tape tracks and MIDI are always in sync. Once you have all your tracks recorded to tape, now is the time to dump them to the computer. Unfortunately, if you don't have an interface that provides at least 8 I/O's, you are going to have to submix the tape to stereo to record to the computer. This will prevent ANY further editing or EQ'ing or tweaking to any of the individual tape tracks.
UNLESS, the tape is able to follow, via MTC or MMC, Cakewalk, which will be then acting as Master set to Sync to Audio. If you can figure out how to do that, you dump the tracks a pair at a time to their own tracks in Cakewalk, thus allowing you to further process them individually, if desired.
OK. So now you have all your tape tracks into the computer. Cakewalk is set to sync to Audio. It's outputting MIDI sync. Now you run your keys, Nanosynth, etc. in, one (or a stereo pair) at a time.
You can now use, or mute, or delete any of the MIDI tracks that you started out with to monitor through the soundcard. (I would have made a backup called something like "TheSongBU" before recording any audio. That way you can always recall it later and start over from scratch).
Sheesh....time for breakfast.
Hopefully, this gives you some ideas that may be useful.
HDB