Shipboard fusion plants generally come in one of two configurations, a single loop torus or a some variation of multiple loops. Most of that is determined by how you want to use space within the hull and how much computational and "environmental" control you are willing to put into the design.
Single loops are easier to manage and generally take less monitoring and crew but take up more space within the hull, while multiple loops are actually smaller due to the reduced size of the individual rings, but, since the flow moves through a more complex path, they require a lot more management because you have to force the flow through all of the gyrations the smaller structure imposes. There are theoretical arguments in favor of multiple plants, but, under current design limitations, they tend to be ignored due to not yielding enough output when compared to the amount of space they take up. There are other issues, including a version of the square/cube law in relation to efficiency, but the space issue tends to be the deciding factor.
By kicking the whole fusion chamber up into hyperspace and using the inherent pressure of the that dimension to augment the gravimetric fields made by a jump drive, we should be able to get a "pinch" more potent than anything previously in human history. With no physical chamber on the ship, the entire active power plant will be no larger than the jump drive hardware. Since the ship won't be "wired" for a functional jump, the majority of the space that a traditional jump drive uses gets reclaimed. All the changes to the navigation, hull and skin of the ship simply won't be installed.
Safety procedures amount to monitoring the process and turning off the jump drive if needed, losing the entire fusion bottle in hyperspace. With the connection to the ship severed, a runaway fusion process won't be able to do any damage, it'll just dissipate into the background energy of hyper.
As a side note, we'll need enough capacitor storage to be able to do a system restart. It wouldn't do to be dependent on a shore power jumper cable.
