Originally Posted By: paul
and the air that actually touches the walls is probably only around .005 percent of the volume of air that leaves the air tank.

Sure. But I said that was for a long pipe. If you keep making it longer, the wall friction will keep increasing, while the end cap force will decrease by the same amount. Eventually the wall friction will dominate.

further than the length of the pipe.

and one more thing , if the air is released very slowly
the pipe could actually travel several times its lenght before the air tank empties or equalizes.

However you do it, after the pipe has traveled its length, there's no possibility that any of the air could still be going in the same direction it was when it left the tank, without having turned around at some point. That's where the length-of-pipe limit comes from. Just try to draw a picture of that, it's geometrically impossible!

The slower you release the air, the more mass is left in the tank while you're trying to accelerate it, so the less distance it can go before the released air _has to_ change direction. This reduces the length-of-pipe limit. The length-of-pipe limit can only be approached when the mass of the released air is much higher than that of the pipe and any un-released air.