I know research has been carried out with Stirling engine-based RTGs, which would trade complexity for a much higher efficiency, but I don’t know what the current situation is with those.
That does make a lot of sense, especially for applications like a Titan probe where there’s no shortage of working fluid and plenty of cooling through conduction. Still, the image of a space probe with a big steam-age-looking piston moving back and forth somehow tickles my funny bone

Am I correct to assume that an actual reactor would be able to get a heck of a lot more power out of that amount of plutonium?

Wrong isotope of plutonium, I’m afraid.
While you can technically burn Pu-238 in a reactor – it’s fissionable (capable of undergoing fission), and, more importantly, fertile (capable of capturing a neutron and turning fissile (with an extra neutron it turns into Pu-239, which is the isotope that’s used in bombs and mixed-oxide reactor fuel)) – it won’t make a reactor fuel on its own.
Furthermore, even if the “right” isotope was used, I don’t think (this is a bit outside my particular field ) the amount used in the MMRTG – 4.8 kg – is enough to sustain a chain reaction.

125 watts doesn’t sound like a whole lot, for 2000 watts of waste heat. I assume that RTGs are used because no supervision (or moving parts) are needed?

Yep. Without moving parts, and relying on relatively simple principles, RTGs are extremely reliable. They’re also much smaller and lighter (an order of magnitude or so) than the smallest reactors ever built.
I know research has been carried out with Stirling engine-based RTGs, which would trade complexity for a much higher efficiency, but I don’t know what the current situation is with those.

@69 Blargh: Am I correct to assume that an actual reactor would be able to get a heck of a lot more power out of that amount of plutonium? 125 watts doesn’t sound like a whole lot, for 2000 watts of waste heat. I assume that RTGs are used because no supervision (or moving parts) are needed?

@70 kansel: Heh, here’s what I got when I did a search for “curiosity” on Cheezburger.

http://4.bp.blogspot.com/–lSNi90D2-Y/TcLToE-PeMI/AAAAAAAABT8/lSYxf9q5YwQ/s1600/AtlasV_SBIRS_MLP3-1.jpg

One mission in the 1960s was lost in a launch explosion; the reactor was recovered intact from the Pacific and flown again on another mission.

As a rad guy, I have to point out that an RTG is not a reactor – RTGs run on plain old radioactive decay. A nuclear reactor is one that contains a nuclear chain reaction: the only fission that occurs here is spontaneous (roughly twice per billion radioactive decays, a Pu-238 atom will spontaneously fission – split – instead of emitting an alpha particle) and unwanted, as Paul pointed out in comment #49.
Although to add to/correct his post, Pu-238 does emit some gamma radiation through its “normal” decay as well. But it’s not a lot, and it’s low-energy, which makes it easy to shield against (which is why this particular isotope is used). It gets worse over time (timescale of years) as other radioactive decay products build up, though.

The bigger problem is with the injector. Conventional injectors don’t work well at low flow rates. However, it’s possible to design injectors that do work well; the pintle injector in the Lunar Module Descent Engine (LMDE) worked over a wide range of flow rates, enabling that engine to be throttled to 10% of max thrust.

Space X’s Merlin uses a pintle injector, btw.

Engines operated in the atmosphere also suffer from flow separation if the nozzle exit plane pressure is too low. This means that if they are highly throttled they either need variable geometry nozzles (plug nozzle, for example), or they need to be very underexpanded at high thrust. Engines operated in vacuum do not have this problem.

Also for information on RTGs up to the mid 1990s see
http://atomicinsights.com/article-file
& scroll down to September 1996 for a few articles.

Solar is best for many space applications, but on a planetary surface where you have to deal with night or if you’re a long way from the sun something nuclear is the way to go.

This is a good example: http://youtu.be/LhnU3KB_fi0

The air is pretty thin at 40 miles but we still hear the bolts fire. And even then, the SRBs are still traveling upwards but we continue to hear SRB fuel bits (I’m assuming) roll around inside the metal body. The sound isn’t traveling by way of air but by vibrating the actual SRB itself. The mic pics that up.

It’s like tapping on a hard surface, we can hear it a bit but if we rest our ear on that same surface, it’s much louder. So even in deep space, as long as your ear was on the object, you’d hear it if you were banging it with a hammer. Sure, there’s that whole suffocating/decompression issue but dangit you’d hear it before you bought it!

I’ll end by saying that to garner interest in these things, people need familiarity. You can better understand how sound travels, or doesn’t travel, in space after your interest is gained. Seeing a soundless video doesn’t make sense to us land dwellers where sound via air is plentiful. Finally, as a creative type myself, I’d feel so very weird putting a video together with no foley. It’s just plain fun.

Imagine instead of hovering, it had some means of horizontal thrust. It could skim over rough terrain where a rover couldn’t drive and a lander couldn’t see anything, collecting data from the air and releasing a number of impact probes like the ones the failed polar lander had.

I’m a big fan of rovers, they’re cool for all kinds of reasons, but there’s places they can’t go. I can see this system being adapted to get the kind of area coverage you get with a rover in those places.

Imagine how you’d feel if you got the lander safely on the ground, but then your garage door jammed!

Makes sense. Thanks!