Version 2 (Mars Colonial Transporter-March 2014)
In February 2014, Musk stated that the MCT’s payload was a large interplanetary spaceship capable of holding 100 tonnes of cargo and crew.
In March 2014, SpaceX Vice President of Propulsion Tom Mueller confirmed a Falcon 9 style engine configuration of nine Raptor engines on the core stage and one on the upper stage. Mueller also announced an increase in the Raptor engine’s thrust to 4,448 kN (1,000,000 lbf), and the choice of a 10 meter diameter for the rocket.
Render showing the Falcon 9-style Raptor engine configuration for MCT – via Nathan Koga for NSF/L2
The Raptor engine would also be employing a rarely used but highly efficient combustion cycle called full-flow staged combustion. This meant that the Raptor engine would be a staged combustion cycle engine with oxidizer-rich and fuel-rich preburners, each fed by a separate turbine. These would be fed by their propellant gas, allowing more power. Benefits include cooler turbine temperatures and lower pressure, which leads to greater reliability and longer engine life. The disadvantages potentially include even greater complexity.
New NSF evaluations showed this version’s increased thrust would improve the single core’s lift capability to 174 tonnes to LEO. An easy way to visualize this is to imagine the single-core version lifting an almost fully loaded Boeing 767-200ER into orbit.
Version 3 (Mars Colonial Transporter-May 2014)
In May 2014, further details of the Raptor engine were given at the Space Propulsion conference in Cologne, Germany. Raptor’s sea-level thrust had increased to 6,914 kN (1,554,300 lbf). Raptor’s vacuum version had a thrust of 8,238 kN (1,851,900 lbf) and an ISP of 380 seconds. The result was the vehicle’s thrust had jumped to 62,223 kN (13,988,300 lbf), which would have resulted in a payload capacity of at least 270 tonnes to LEO. This is roughly equal to placing an Airbus A340-300 with 335 passengers into orbit.
Version 4 (Mars Colonial Transporter-June 2014)
In June 2014, as SpaceX started Raptor component testing, Mueller revised the Raptor engine’s thrust to 7,414 kN (1,666,700 lbf), which rivaled the most powerful rocket engines ever built. As a result, the vehicle’s total thrust increased a further 7.2% to 66,723 kN (15,000,000 lbf).
Furthermore, details emerged that SpaceX was moving away from considering a multi-core design and was instead concentrating on a single large core, with diameters ranging from 10 to 15 meters under consideration. The major reason for this shift likely was due to the much smaller performance penalty for reuse allowed by a single core versus a multi-core.
Render of the single-core SpaceX MCT – via Nathan Koga for NSF/L2
Following the upgrade in thrust, NSF evaluations showed that any of the single-core variants under consideration would have been capable of expendably lifting more than 300 tonnes to LEO, with roughly half this capacity with reuse.
Version 5 (Mars Colonial Transporter-October 2015)
In January 2015, SpaceX reversed course and suddenly revised the Raptor engine’s thrust down to 2,256 kN (507,100 lbf), though no mention was made of changes in the rocket’s overall thrust or design, other than there would be a lot of engines.
One of SpaceX’s initial hand-drawn sketches of the upper stage was of a simple 15 meter wide capsule rimmed with 15 Raptor engines, each sporting a 2.89 m (9.5 ft.) nozzle. The cargo and crew would sit beneath the propellant tanks. A spherical liquid oxygen tank 13 meters in diameter would sit in the nose, with the liquid methane tank beneath it.  It would enter a planet’s atmosphere on its side, as would all of its successors. This was because side-on entry has good entry, descent, and landing (EDL) performance, while capsule entry made protecting the engines from re-entry heating very challenging.
SpaceX eventually rejected the initial capsule design because it required longer feedlines through the crewed areas, higher mass, and a worse crew environment. Next, the company considered no fewer than four upper stage variants, each with small spherical header tanks for landing. These would help the vehicle avoid flameouts during high g descents due to propellant sloshing. Since then, every known variant of the upper stage has mentioned the possibility of having these spherical header tanks.
Diagram by NSF member Lamontagne.
The first upper stage variant considered was a 15 m (49.2 ft.) diameter spacecraft powered by 15 Raptor engines. The Raptor engines would rim the bottom perimeter of the spacecraft, each angled 15 degrees outwards.
The propellant tanks would fill the center of the vehicle top-to-bottom. The liquid oxygen tank would sit atop the liquid methane tank and entirely fill the vehicle’s nose. A secondary cargo area would rim the bottom of the vehicle’s oxygen tank and would separate the crew from the oxygen tank. In turn, the crew would be housed in a doughnut-shaped pressure vessel rimming the upper half of the methane tank.
The main cargo section would be immediately below the crew, rimming the lower half of the methane tank. A single oxygen propellant pipe would traverse the length of the methane tank. The Raptor engines would have been nestled below the main cargo area. The methane tank would have ended before reaching Starship’s base, leaving a relatively empty structural area in between the ring of Raptor engines.
The second variant featured three Raptor engines arranged in a triangle pattern in between an outer ring of 12 engines. This variant had the oxygen tank in the nose as well. Its methane tank, however, was the width of the Starship. It also had an oxygen pipe going through it. This design’s propellant tanks took up far less vertical space, although it was taller than the prior variant.
A pair of propellant pipes traveled through the main cargo area, taking up most of the second variant’s length. A doughnut-shaped crew pressure vessel would be nestled outside the upper half of the main cargo area. A small secondary cargo area would separate the crew from the propellant tanks above. The main cargo area would be on the base and take up the full width of the Starship.
The third variant was similar to the second variant in layout but was narrower and taller. It had five centrally mounted Raptor engines in a pentagon pattern rimmed by 10 Raptors in a ring.
The fourth variant of the early Starship was even taller and narrower than the third variant. This version was the only one where the propellant tanks were placed in the Starship’s base. Both the methane and oxygen tanks took up the width of the vehicle, with an oxygen propellant pipe also going through the methane tank. The main cargo area was located above the oxygen tank and spread across the entire width of the vehicle. The crew compartment sat atop this and also made up the entire width of the vehicle. A secondary cargo area was located in the nose above the crew.
By the time SpaceX had an MCT design meeting on July 25, 2015, the design had evolved even further. In the meeting, Musk noted the design had to be more efficient than Dragon, wanted cruise ship-style private cabins or groups of cabins for all passengers, and likened it to being more like a yacht than a dinghy. It was decided there would be a similar architecture between the booster and ship.
This meeting also features the first known mention of a tanker, making three flights to fully fill the MCT. The MCT would feature large doors to move cargo and crew to the surface. Other design elements were mentioned, like a potential 900 tonne spherical methane tank. The most radical change aside from the engines was the switch in propellant tank materials.
Each propellant tank would be a linerless composite monocoque tank insulated by aerogel at 800 pascals (6 Torr) strength. The aerogel would possibly feature internal structure to share loads across it. The tanks would be autogenously pressurized by gaseous forms of their propellant at 1.7 bar (25 psi) ullage pressure, which SpaceX engineers found provided a good trade-off between improved reliability and increased mass.
Lithium-aluminum alloy tanks were found to be inefficient at these scales when autogenously pressurized, as they had to be strengthened against buckling. Carbon fiber tanks offered greater geometric stiffness than aluminum alloy, allowing a lower mass vehicle for the same size.
The carbon fiber tank that SpaceX would eventually construct and test – via SpaceX
Musk even mentioned his desire to combine the thermal protection system (TPS) and the structure to save mass. This presaged SpaceX’s eventual decision to use carbon fiber for the entire design. The main reason for this was its low coefficient of thermal expansion would allow SpaceX to attach the ship’s thermal shield directly to the structure, potentially saving even more mass.
In August 2015, Musk announced that Raptor would have an oxygen to methane ratio of 3.8:1. NSF gained further insights in October 2015 about Musk’s plan. The MCT was to be a 180 meter (590.6 ft.) tall single stick design of 12 meters (39.4 ft.) in diameter, mass 4724 tonnes, with an 1125 tonne spaceship-mounted atop a 3599 tonne booster. It would launch off the pad with 62,223 kN (13,998,300 lbf) of thrust from 27 Raptor engines and be capable of reusably lifting 236 tonnes to LEO. That’s an amount equal to a fully-loaded Airbus A330-200F freighter. It would also, astoundingly, make its reusable version the world’s most efficient launch vehicle with a payload equal to 4.99% of launch mass.
The MCT spaceship and tanker would be 60 meters long and were vertical landing lifting body designs capable of docking and transferring propellant. Both would also feature a TPS using PICA 3.0 ablative material. The thrust ratio of the core stage to the spaceship/tanker stages was a low 5:1, while most rockets have thrust ratios of 8:1 or more. NSF experts realized this ratio meant that there would be five Raptor vacuum engines on the spaceship. The upper stage also likely had three centrally mounted Raptor landing engines by this time.
The MCT spaceship itself would feature a large habitat section complete with suites of crew cabins mounted on top. To carry cargo, the spaceship would feature a lower cargo bay mounted above the Raptor engines, with the engines fed via a pair of central tubes that would cut through the cargo bay. This layout would allow the easy offloading of heavy cargo like vehicles or nuclear reactors. The spaceship and its tanker counterpart would be supported by five extendable landing legs that would not fold down like the Falcon 9’s legs. Surprisingly, there were no plans for a launch abort system (LAS) on the spaceship.
The plan’s summary mentioned a height of 180 meters that was likely outdated, as the rocket’s thrust dictated a much shorter vehicle of around 105 m.