Ryan, holding the OS2.
OS2 incorporated a major innovation, the brainchild of Rusty Cook, my brother-in-law. I've dubbed it the "slider/reducer". It's awesome. Truly. Plainly put, this gem is a small rocket nozzle that kicks in part way into the ascent, providing a longer "burn" time.
Here is a 3D rendering of Rusty's slider/reducer (click for larger image).
(Rendering courtesy Render-Cam Images)

Here is how the slider/reducer and the main nozzle work together. The slider starts in a position higher up in the tube, but as the water beneath it is expelled out the main nozzle, it settles and seats down against the end-cap (as shown), and all of the water left above the slider/reducer exits out the small port made from threaded lamp rod. Hard acceleration can be seen even as the rocket is passing 100' altitude.
(Rendering courtesy Render-Cam Images)
Here are the real parts... first from the nozzle end...
... then from above the slider/reducer.
Here are the parts slightly separated.
Modifying the PVC reducer to serve as the end cap/main nozzle on the Sherline lathe (an AWESOME small lathe). The outside diameter must be turned to be a snug fit into the flourescent tube. Not too tight (or it will split the tube) and not to loose (it might pop out).
Here is diagram of the best way I have found to cap flourescent tube covers. A good air seal comes from putting a bead of silicone (GOOP) into the tube, and pressing the cap into place, contacting the silicone. PVC cement is for *structural* strength, not for air integrity. It should be applied ONLY to the nozzle, evenly, yet sparingly. I often wipe off some of the cement with my finger just to make sure it's even and sparce. I then press the part into the tube in one single twisting motion. If you stop only part way in, the cement will set (yes, it's that fast). Strapping tape is added to the outside of the tube because it seems like a good idea :-).
Two slots were added with the table saw, as part of the hold-down system for the rocket.
Here is the "intake" end of the nozzle... My theory is that having a tapered hole (on the pressurized side) makes for more efficient thrust.
The Top of the tube was capped with another custom turned PVC piece. This time, a cap. After making a center hole through it and a PETE bottle cap, threaded lamp rod was used to provide a port which serves to allow air to transfer from the secondary air chamber (a standard 2-liter PETE). Again, PVC cement was used, along with strapping tape to further strengthen the area.