As far as I understand it they are pumping the equivalent thermal energy as an actual operating core would produce into the reactor casing and assembly using a specially designed electrical insert that uses external power.
This isn't a fission test. It's a mechanical engineering test. Still cool.
From my understanding the fission part is "easy", the hard part is taking the heat and converting to electricity efficiently and reliably.
The Mars Curiosity rover among other probes use a thermocouple based generators to create electricity from the fission heat[1]. Extremely robust , no moving parts, but extremely inefficient just a few percent converted to electricity the rest to heat. The heat is very useful on mars though due to the low temps, kinda like your combustion engine in your car, the waste heat can be used to provide useful heating, but still most needs to be rejected and it only puts out about 100 watts of electricity.
This new one looks to be striling engine based which means its going to have much better efficiency since it is a more standard heat engine with moving parts, prob 20-40%. Hence the higher output. However that comes with much more complexity and things to go wrong over long term use.
I still wonder if there is much more efficiency to be had from thermocouple's if the research effort was put into it, it would be similar to solar cell improvements over the last few decades. You just don't hear about thermocouple R&D and if it where improved to say a typical solar cell efficiency of 18-20% it would open all sorts of doors.
In fact the fission[1] part in such small scale is hard. Curiosity and other probes using thermonuclear batteries harness alpha decay[2], which is relatively easy as it happens whether or not you want it to happen, but creates also much less power. In that case the hard part is to obtain material with appropriate half-life and other properties. Perhaps best material for this is Pu-238, which is far more expensive to create than weapons grade plutonium.
I don't think "thermonuclear" is the right term for the batteries in probes. Wikipedia says "thermoelectric"[1], as the electricity is generated directly by thermocouples.
"Thermonuclear" usually refers to the sorts of fusion reactions found in stars and modern nuclear warheads.
Sorry I wasn't using the strict term of fission[1].
This reactor says it uses U-235 which would be full fission similar to the US SNAP-10A[2] or Russian BES-5 RTG[3]
So yes the fission part is more complicated than a P-238 alpha decay RTG. Perhaps I mischaracterized the R&D complexity on the reactor portion, although it has been done before.
When writing the answer, I didn't even consider, that those reactors need to be fast reactors. In retrospect it is obvious, but makes controlling the power even harder.
There are fundamental physical limits of quantum nature to Seebeck effect based thermocouple efficiency.
The maximum what you can get can't be higher than the hypothetical ideal quantum diode made from a material pair.
There are much more down to earth alternatives that beat both stirlings and thermocouples on reliability and specific power density: thermionic converters, thermoaccoustic generators, AMTEC converters, radiophotovoltaic (works only in 0g)
"Thermoelectric efficiency depends on the figure of merit, ZT. There is no theoretical upper limit to ZT, and as ZT approaches infinity, the thermoelectric efficiency approaches the Carnot limit. However, no known thermoelectrics have a ZT>3."[1]
There seem to be some in promising materials with a ZT of 2.2[2] which is around 20% similar to a solar cell. It seems again if more R&D where applied we might be able to make that cheap and practical.
1. Work function - how much energy is needed to let an electron move from one material to another.
2. Quantum tunneling - electrons will tunnel back to lower charge density region, and preventing them from doing so is effectively impractical with modern day tech.
If they were any actual thermocouples with 20% efficiency, they would've nuked piston engines long time ago.
Fission isn't necessarily easy. One problem is that the reactor behavior changes when you add the generators to the reactor. These components can reflect neutrons back to the reactor, which changes how much power the reactor produces. That being said, 'easy' or 'hard' are really not the best way to think about flight ready hardware. Actually making hardware and guaranteeing that it works involves so much more than just understanding nuclear fission. We can come up with a design on paper, but without testing we don't have assurance that it works.
Simply putting the reactor together and testing it are more complicated than they seem. This is because of three things matter can reflect neutrons back, humans are harmed by radiation, and humans are the ones assembling/transporting/testing the reactor. We have to be careful putting the reactor together so that it doesn't spew neutrons and harm the humans assembling it. We need to develop ways to transport the reactor so that it doesn't spew neutrons. When we test the reactor, we have to set up our test so that humans won't have to get anywhere close to the reactor for a couple of months. There's also other issues like training people to do these things and NASA cooperating with the DOE to do the testing. These are 'simple' things, but they still need to be done.
They already completed the equivalent thermal power tests last year[0], and you most certainly don't need to do electrical tests at the Nevada Nuclear Test Site. How much progress they've actually made at this point is not clear from the article, but slides from a presentation they posted yesterday detail what they plan to test[1]/maybe already tested. They need to test the individual reactor components to measure reactivity, then they need to put the power generator on and see what happens when it's just critical, but not producing any heat, then they bring thermal power up to 4 kilowatts, then they do a full power test.
All of these tests involve fissile materials, so to quote Red Alert, "Gentlemen, it's a nuclear device."
This isn't a fission test. It's a mechanical engineering test. Still cool.