Atlur has it quite right.
I suggest you take a look at this page:
http://pwg.gsfc.nasa.gov/stargaze/Sintro.htmIt deals with a lot of issues.
The plasma I'm speaking of, does not come from the reactor - if it were so, you wouldn't see anything coming out of the exhaust.
It is a low degree plasma, where all kinds of agregates are found.
You know, plasma does exist in our daily life - have you ever looked at a flame? It is plasma.
The engine I'm speaking of is a nuclear engine, but don't confuse it with a fusion engine.
It merly uses the heat from a nuclear rection to heat a prolusion mass - that matter could be anything!!!
That's a lot less expensive than using hydrogene - or fusion close hydrogene namely.
Atlur is right about the magnetic coil thing - the reason why plasma is used, is that it can be contained with a magnetic field, so you don't need a super heat resistant fan to steer the hot gas - this is also the method behind ion engines.
It has no moving parts, so it can be a lot faster.
This also conforms with Bobbau's new trusters.
As for the your engine problem with capships:
The vents and tunnels I'm speaking of are internal!
So you have a system where the matter of nuclear reaction is transformed into kinetic enegy.
The tunnels are used to capture this momentum and give it to the frame and the hull.
A fighter can't use a system this big, so it uses a more complex system.
It's not the liquid that stops the fighter, it is its own engine.
However, the process is made easier by the liquid core.
It IS a burden in terms of mass - because it has to be heavy.
The reason is, that it holds most of the crafts momentum, when the liquid is set loose (in the core exclusivly!) it takes away the crafts inertia.
By converting that force into electricity - that's why the layer has to be charged - you strip the fighter of its own momentum.
Imagine fishbowl on wheels, filled with water by 1/2.
Now, when you try to stop the bowl, the water will still go in the original direction the bowl used to go.
You would validly think, that it's easier to stop an empty bottle, so far the liquied core seems useless.
However if the liquied was a completly charged collection of positive particles, the story would be different.
If passed through a magnetic field, the particles loose their speed, but in turn generate electricity. That's how your hard drive works actually.
The reason why liquid is used, that it can easly conform to any direction necessary, while building a concrete system that would do so (like the magnetic disk od your HD) would never be as efficient.
BTW it is indeed the He3+ part that makes this reactor a lot more expensive.
That way you don't need small thrusters Atlur, and you don't need any moving parts except for the core itself.