It's not just a matter of DPS, it's DPA as well. It's proven that by the time Shield Cannon has finished animating, you have already completed an Aura Strike, a Bash, and are part-way into another Bash. Going by base attack values at Rank 5, in the same animation frame...
Shield Cannon: 88% Attack Damage.
Aura Strike: 24%+21% Attack Damage | Bash: 25% Attack Damage, +1 Aura | Bash (partially animated), an effective gain of... +11%? Attack Damage, needs correction.
We'll pretend I'm right on the effective gain of the third, partially animated bash. Your totals become 88% Attack Damage for the single Shield Cannon, and we'll say it crit at the accepted 2.4x Critical Value of Battle Manual to produce 211% Damage.
Aura Strike + Bash + Some of a Bash at base value is 81%. We'll say, for the sake of argument, Aura Strike managed to critical, increasing this value to 144.
Clearly Shield Cannon wins, right? Well, let's also say that, while less plausible, Bash and Partial-Bash also crit. Your total then goes to 205.4%.
Shield Cannon still wins, right? Well, then you need to factor in you also have 1 and a "half" Aura ontop of the Shield Cannon user, which means by the time they've activated Bash, you've used your third Bash and have completed an Aura Strike. You're also ahead on Bashes, meaning Aura Mastery procs.
This proves, although my percentage-based math on the total 'gain' of the partial Bash is probably slightly off in some direction, Shield Cannon is a DPS loss in a sustained fight, but it's burst capabilities are with out equal.
If you're looking to maximize DPS in any fight lasting longer than, we'll say 30 seconds, do not use Shield Cannon unless you can ensure (You can't) it always Crits. If it does not crit, the DPS loss becomes even more substantial.
THAT, is how a DPS rotation works.
PS: I ignored the potential FURTHER gain of a maxed out Headcrush over Shield Cannon. Yeah, it gets even more absurd. Or, has Jelly noted earlier, the DPS loss comes at >10< seconds in.
Edited by Reppu, 11 May 2013 - 08:14 PM.