First just want to say this is an impressive bit of machinery. I really like it especially since I also have an 86 GSXR 1100.

unk
OK so I played around with Power point and did a quick spreadsheet (the problem was intriguing :dunno). I attached a somewhat more complicated design, but it is certainly more balanced.
I won't claim to understand all (any) of the issues with trying to couple the two motors but assume the crank to crank spacing is required to get an integral number of links within that spacing. The idea is that the front lead motor will pull the most and the straight section of chain will establish the relative timing (I realized this just before I posted the pic :sad). The motors are spaced exactly N links apart and with N links on the sprockets they are in sync (nominally at least). If the chain stretches you have to slide the motors further apart.
That must be the reason for trying to pull out all the chain slack on the bottom. Of course is this is very unbalanced . Additionally the adjuster is pushing "in" and so the small idler has to spin so much faster as well as the increased chain deflection angle.
Here are some highlights:
1.) Lower chain deflection angles on the chain adjusters because there are two adjusters. Slight increase on the lead drive sides with a decrease on the follower drive side
2.) Idlers have reduced load because the chain is wanting to fly out away from the idler rather than pushing against it (this might be a minimal effect).
3.) I believe that there is sufficient adjustment to get 1 full 530 link of slack adjustment on the leader drive side with less chain deflection. Then the lower follower side can be set for slack adjustment. Depending upon the symmetry of the idlers, it may be close to identical.
4.) Since there is more space inside the chain, larger idlers can be used reducing the idler RPM in proportion to the chain in #of teeth
5.) The adjusters are angled for improved fit as well as to better oppose the drive side chain tension even though the angles are reduced.
6.) By simply using the timing marks on both motors, the tensions can be set to synchronize the motors without moving (crank shaft spacing) either motor as required in the current version.
I also threw in some guides (inside and outside in blue). Also little squirter's for an oiler would be easy to add as show rather than going to a full bath.
I have now way of knowing, but I would certainly think that this configuration would be much more stable and once the engines were synced (using timing marks and tensioning) that the bike would run longer in sync as there is likely to be less chain wear and stretch and even the chain wear and stretch will be symmetrical and nominally not change the timing (as the existing will if there is any stretch). It is easy to do a quick after run alignment adjust if required and you are off and running again. :burnout
Nice bike what ever you happen to decide.
Posplayr
P.S I'm just looking at the pic again, It seems now you might have enough room to sync the cranks instead of the transmission and then run a separate chain and gearing for the tranies if you think a single transmission will handle the power.