The tail structure is the reactors and radiators, while the EM-Drives are on the wings near the crew cabin.
While EM-Drive is still being verified as a real thing, pondering its application is still possible. One thing we should discuss is the power supply needs for the 1 gee space-craft and the EM-Drive parameters needed for that performance. Say we have a 1,000 ton spaceship and we accelerate at 1 gee. Thus 10 MN thrust is required. At Eagleworks’ observed performance of 1.2 millinewtons per kilowatt, the power to get 10 MN thrust is 8.3 TW, which is probably unrealistic for a 1,000 ton ship to supply. We might be able to do it with a laser bank beaming at a high-efficiency collector, but it’s still herculean by modern standards (we’re just starting to make sustained 100 kW lasers.) It’s still much better than the laser power needed for a purely photonic drive, some 3 petawatts, but it’s a long way from our state of the art.
Of course the performance increases with the Q-factor. Tune the cavity and make it superconducting. If we take the NASA EM-Drives and pump the Q factor to ~30 million, then about 2 GW power is needed for the sustained 1 gee thrust. A fast-spectrum reactor with a thermal output of ~ 6 GW and ~35 % thermal conversion efficiency would be a first pass design to supply the power. Assuming ~100% burn-up the fuel used over 20 years masses 4.2 tons. If the reactor mass was limited to ~200 tons, then it’d need to supply power at 10 kWe/kg of reactor mass, which is very high performance. A gas-core or magnetic collimator fission-fragment reactor might be up to the job, but both are somewhat futuristic.
Fusion reactors are presently *hoped* by propulsion engineers to be able to supply power in the 100 kWe to 1 MWe range, though no working fusion power-reactor has yet been demonstrated. As the joke goes, fusion has been “just 20 years away” for the past 60 years and always will be…
I wonder just how high a superconducting EM-Drive’s Q-value can go? If we assume a Q of 3 billion, then the power required is 20 MWe, which is more than any reactor ever orbited but well within known design parameters. Ship systems would be some additional fraction of that supply and two reactors would be needed for redundancy and maintenance, along with a significant amount of storage. If we assume a more modest specific power of ~400 W/kg, and two reactors, then total mass for 25 MWe supply is 250 tons for the reactors and 50 tons for the ‘battery’ bank. Thus 700 tons of starship remain to allocate.