Air directly to the rail will cut your emmissions, will also backfire at decel(no big deal). Air injection is KEY as stated by ellington, and as you found. I am wondering why air was not getting to it destination before. The asv is operated by a vsv which is operated by the smog computer. The computer gets input such as speed, temp, airflow and uses these inputs to switch things like your vsv which will switch your avs and allow or prevent air from getting to either the manifold or cat. I am curious, Southbound, What your asv is doing w/ the engine running and if you were to take a ground to the ground terminal of the #1 vsv. W/ the computer and sensors working properly, the vsv should open the asv anytime you are not idling or decellerating, and allow air to enter the manifold. If you by-passed the computers signal to the vsv, and gave it straight ground ( it should already have the +12) it should open the passege from the asv to the manifold. If it DOES open w/ ground provided, I would say there is a problem w/ the wiring to various senders or the senders themselves or the actual processor. Getting the air to the rail will get you into a good emmissions range, but it will also allow air into the stream while under decel or idle, which will give some backfire. So now that you've got that sticker on your plate, lets take the time to find the root of the problem. Take a look at your electrical diagrahm, it will show the various senders and sensors that will give the computer its information. A problem could exsist w/ any of the sensors, but you could eliminate wiring connections and individual sesors by chasing each one from its point of ground back to the computer. Keep us posted here. Just cause you passed doesn't mean all is well.