[serialk11r]

drmiller, you're not compressing air though...that calculator gives the amount of power it takes to compress a certain amount of air from atmospheric pressure to 5psi gauge pressure. 5psi gauge pressure means 5psi above atmospheric. This does not describe what is happening with the throttle.

What happens at the throttle is there is atmospheric pressure on one side, and air is flowing across to the manifold with lower pressure. The pressure drop is going to be atmospheric pressure minus manifold absolute pressure, aka vacuum. The air escapes through the throttle opening, gains velocity, and then has all that energy dissipated into the air itself as heat.

A simple estimate for pumping loss is to consider the piston moving against a constant amount of pressure. The manifold pressure is lower than atmospheric, so it takes at most the amount of air displaced by the piston * the pressure acting on the piston to pull the air in, as pressure * volume = work (more accurately it's \int PdV, but whatever). But now that the piston is filled with below atmospheric pressure air, some of the work used to pull the air in is returned to the piston as it travels up. Previous posters have used this upper bound estimate to calculate a number for pumping loss. Admittedly pumping loss is significant, but not as great as friction or even cooling losses. For example, if we look at BMW's Valvetronic engines, BMW claims a 10% overall fuel efficiency increase. Since the EPA tests run at low loads, the amount of throttling loss saved should be significant, yet they only saw a 10% increase. Now of course 10% is nothing to scoff at, but considering that the part load efficiency is still pretty bad, there's clearly a lot of other things at work. At idle I believe throttling accounts for about 60% of power usage, since the throttle is used to consume the engine's excess output, but you don't need very much load before the throttle is second to friction.

A good way to measure pumping losses would be to compare the manifold temperature to the temperature that air would be at if adiabatically expanded to the manifold pressure. The difference in thermal energy is the pumping loss.