It also (in the case of LOX/RP1 rockets, anyway) uses the stuff you already have large, mass-efficient tanks full of and which you already need to pump around
Yes, but...
Turbopumps in large rocket engines are
powerful. As in far-more-powerful-than-a-speeding-locomotive powerful. A few messages back I computed a net mechanical power of 17.5 megawatts for the pumps in
each F-1 engine. A typical Diesel locomotive generates only 2.2 - 2.5 megawatts. The turbopumps in the five F-1 engines on a S-IC stage required more power (87.5 MW) than all six Diesel engines on the late Costa Concordia combined (75.6 MW, total weight 1,000 tons).
Even the 100% efficient generation of that kind of power for several minutes requires a seriously large amount of fuel (and oxidizer). The less efficient the generation, the more fuel and oxidizer you'll need. Providing a separate fuel and/or oxidizer supply for the turbines might well be worthwhile from a mass standpoint
if the turbine can operate more efficiently than it could by burning the main propellants.
Bob B. commented that staged cycle engines can achieve higher Isps by operating their main combustion chambers at higher pressures than gas-generator cycle engines. The F-1 engine (which uses the gas-generator cycle) would seem to bear this out; it achieves a sea level Isp of only 263 sec as compared to the RD-180, which uses a staged combustion cycle and achieves 311 sec at sea level. That's a
big improvement.
Edited to add: Checking the F-1 documentation, I see that the actual turbine output power was 41 MW. I got my figure of only 17.5 MW from the actual work done by the pumps on the propellants; I knew the actual turbine power had to be considerably greater to overcome the usual losses. This further strengthens the case for more efficient turbines even at the expense of a dedicated fuel/oxidizer supply.
Edited again to add: Hey Bob, wanna figure the performance of a Saturn V in which the F-1 engines have been replaced with ten RD-180s?
