The cost of a mission, whether robotic or human, to service the Hubble, is at least $1.5 billion, maybe $2 billion, according to NASA figures cited in a meeting this week of the House Science Committee.
http://www.wired.com/news/space/0,2697,66486,00.html?tw=wn_tophead_4Hubble was conceived in the 1970s and built in the 1980s, and like the Shuttles, its technology reflects its era. It may be that the best thing to do with Hubble is to use it up, and then move on...rather like you did with your old Windows 3.11 PC.
An international team of researchers led by Johns-Hopkins astronomers says instead of putting more cash into Hubble, we should put the money into a successor. Several new instruments intended for a Hubble upgrade have already been built and paid for, and have been sitting in storage, awaiting a refurb mission. These could be used in a new telescope designed with up-to-date technology, producing major savings.
http://www.spaceref.com/news/viewpr.html?pid=16050Hubble's main lens, as most people know, was ground incorrectly, and had to be corrected with a helper lens installed later by astronauts. While this correction restored much of the original design function, it did not restore it all, and also introduced lens artifacts into images that have to be digitally subtracted. In other words, Hubble not only uses old technology, it is also a bit hobbled.
The Johns-Hopkins team says that building a new telescope would take about five and a half years, and cost under $1 billion. That price includes the cost of launching on a Delta rocket (no Shuttle needed).
Without servicing, Hubble will last another two to five years. In the worst case, if we followed the Johns-Hopkins team's recommendation, we might be without Hubble-like functionality for a few years. In exchange, we'd get a far superior telescope, for less total money, and no more reliance on Shuttles.
I was a big proponent of servicing Hubble but I've changed my mind. Sometimes less is more.