Let's see how it works!
....and why it fails!
DISCLAIMER: I am not a mechanical engineer, so my explanation could be wrong on some accounts. Feel free to correct me. Also, I would not recommend doing this, as there are spring-loaded parts inside, as well as oil vapor, which could cause a fire. Used motor oil is known to the State of California to cause cancer and birth defects. I took all necessary precautions while cutting this open (gloves, eye protection, fire extinguisher nearby, etc.). Please keep in mind this is an AOS which, to my knowledge, had NOT failed in any way. 
First, a little lesson in how a Positive Crankcase Ventilation (PCV) System should work.
With the engine idling, there's high intake manifold vacuum (suction) and low crankcase pressure (due to the low cylinder pressures, and consequently low crankcase blowby pressures). This causes the PCV system to draw very little crankcase emissions into the intake, since very little exist.
At cruising speed, intake manifold vacuum is lower, but crankcase pressures are higher, due to the engine actually moving a load. This results in a medium amount of crankcase emissions being sucked into the intake.
At wide-open throttle (WOT), there is close to zero intake vacuum, but crankcase blowby will be relatively high. This results in the maximum allowable crankcase emissions to be sucked into the intake.
During an intake manifold backfire, the PCV valve acts as a check valve to prevent the vaporized oil in the crankcase from igniting, which could cause a crankcase explosion. Intake manifold pressure (the opposite of vacuum) forces the valve closed, as the intake would have higher pressure than the crankcase in this situation.
Now let's see how the AOS accomplishes this:
As shown in this pic, crankcase air enters the small-diameter passage (the one with the orange o-ring at the very bottom) and enters a large chamber. As the crankcase oil vapor enters the large central chamber, the pressure decreases (due to the significant volume compared to the tube). As the pressure drops, so does the velocity. Both the pressure and velocity reductions contribute to the condensation of the oil vapor back into droplets. These droplets would fall harmlessly to the base of the large chamber and would eventually drain back into the engine. These same droplets would have a difficult time making it past the tube at the top of the center chamber, if they were lucky enough to remain suspended in the airstream.
The top two arrows show the direction of the oil-free crankcase pressure as it leaves the chamber.
The arrow on the bellows shows a second path for crankcase pressure (possibly that which is developed in the cylinder head valvetrain area?) which is entirely seperate from the crankcase ventilation path. As the pressure enters the bellows, the only escape path becomes the tube outlet to the left. Above this tube outlet, inside the AOS body, the path suddenly narrows, giving the pressure nowhere to escape. This may induce a small amount of turbulence into the exiting stream. Also, because the narrowed path is directly inline with the bellows, I'm concerned that bellows "blow-out" may occur to due pressure spikes under certain conditions. Could this cause a weakened bellows to fail? The pressure that exits the bellows path takes a convoluted tube somewhere else (presumably the intake).
Let's move to the top half of the AOS, which contains the diaphragm to control the escape of crankcase pressure.
Crankcase pressure enters through the rectangular hole at the base of the diaphragm housing. Intake suction is drawn through the hole in the center of the housing and out through the tube connection to the intake pipes near the throttle body. Access from the rectangular hole to the suction hole is restricted by a pair of spring-loaded plates and a rubber diaphragm. You can see one of the springs (this happens to be the large one) in the pic.
Shown here is the underside of the removed diaphragm housing lid. Don't mind the tear in the rubber diaphragm, I did that during the removal. You can see the large plastic disc attached to the diaphragm. This plastic disc is supported by the large spring in the previous pic. The plastic disc does not seal against the body of the disphragm housing, only against the diaphragm itself. Also shown in this pic is the small spring which actually controls the passage of pressure through the housing (along with a small amount from the diaphragm itself).