As I mentioned before, Japanese were trying to build an "economical" rocket (same way French plan to launch Soyuz to supplement their exising big rocket). Now Daily Yomiuri says that the project is essentially dead:
Due to conflicting interests among politicians, bureaucrats and business leaders, the recommendation to scrap the GX rocket development program has been in limbo for three months. The most regrettable aspect of the present circumstances is that those concerned have been daunted by the situation and failed to hold discussions on the matter.
A methane rocket sounds like a great idea. It has the next best performance after hydrogen, without all the horrors of LH2. Korean company C&S and American Armadillo Aerospace, among others, demonstrated methane engines. Armadillo spent a few hundred thousand dollars, and, since it's an upper stage, Japanese probably could get away with an Armadillo-style pressure-fed engine. So, how did they manage to overspend by billions? It's a mystery to me.
I'm sure someone is going to e-mail me that an EELV class launcher is nothing like Armadillo's toys. Very well, but then Elon Musk developed a kerosene engine for $100 million.
What the heck, Japan?
UPDATE: Regarding the commentary at Steven's, here's a handy table of specific impulse for modern engines and motors (the material is taken from official data, not Wikipedia):
| Motor/Engine | Sea level Isp | Vacuum Isp | Fuel |
|---|---|---|---|
| Shuttle SRB | 269 | - | Solid |
| H-IIA SRB-A | ? | 281 | Solid |
| Merlin-1C | 275 | 304 | Kero/LOX |
| RD-180 | 311 | 338 | Kero/LOX |
| RS-68B | 365 | 420 | LH2/LOX |
| SSME | 366 | 452 | LH2/LOX |
| J-2X | - | 465 | LH2/LOX |
So, even the Merlin, cobbled together by a company with zero experience, is a more efficient engine than any SRB, although I have to admit: I am thrilled just how good modern solids are. The numbers for Ares SRBs are bound to be better because they use a better fuel (HTPB instead of the old PBAN). Still, there's no contest when it comes to oxigen-rich kerolox engines, and LH2 engines are just in league of their own.
With this kind of numbers, why would anyone use solids at all? If we ignore strap-ons that are added as an act of desperation onto an exising launcher, the answer is in the so-called "gravity loss". In order to achieve orbit, a rocket has to attain a certain horizontal velocity component, and nothing else. Thus, any thrust that goes into lifting the rocket up is wasted. Therefore, it may be advantageous to configure a multi-stage rocket which has a rather wimpy, yet efficient core stage, with powerful and inefficient boosters. Such boosters only work a short time to get the vehicle out of the atmosphere, then separate. (It would be even better to have powerful and efficient boosters — if you can).
This scheme is used by Arianne, and exising Japanese rocket H-II (all modifications). It is also used in both of SSA's rockets in Rocket Girls, although I am not sure anyone did any real trade studies for those. Usually, however, such configuration means that the space agency in question does not have know-how or money to develop powerful enough liquid engine, like F-1 used in Saturn-V. French and Japanese probably lacked both. In case of Shuttle, the excuse is that engines have to be packed into a reusable vehicle, so they cannot be too big.
If Ares was developed from scratch, there would be no reason to use solids. However, at least initially, both Ares-es were supposed to be Shuttle-derived, and so they used the abovemenshioned scheme.
UPDATE MOAR: Steven deleted the post I linked above, because I filled his comments with unacceptable invective. I apologize. Yes, really, I'm sorry.
UPDATE 2008/11/25: For an inexplicable reason, I forgot to look up the planned parameters of GX LNG 2nd stage, although the specifications are publicly available at galaxy-express.co.jp. According to the spec, Isp is 323 s over 118 kN thrust. Surprisingly low, but apparently the engine is driven by an unusual pump (perhaps an XCOR-style piston pump). While it would be quite respectable if XCOR did it for a budget of a few of millions, the question in the context of GX is, why not just use a Soviet kerolox upper stage engine?
BTW, Ed Kyle's article on GX lists 355 s. I wonder where he got that number. If JAXA suddenly decided to boost the efficiency so much, it would cause enormous overruns, I bet.
Another thing, the JAXA website says that they backed up the work due to one little setback with a burn-back on the 1/5 scale model. Why is this a big deal?