The cost of failure of a space mission is millions or even billions of dollars and, more importantly, could lead to loss of life. So testing spacecraft is an important part of developing new ways to travel into space. Those that carry personnel, such as the Orion capsule — part of NASA’s Constellation programme to return astronauts to the Moon — must function properly in the severe conditions of space to ensure the health and safety of the individuals on board.

The Space Power Facility (SPF) at NASA’s Glenn Research Center’s Plumbrook Station in Sandusky, Ohio, is the world’s largest space-simulation chamber used to test spacecraft. The cathedral-like chamber is 37 m tall and 30 m across, with a volume of around 23,000 m3. The facility — known as a thermal vacuum chamber — can reproduce both the temperature and pressure of space. The latter can vary from around 10–3 mbar at 50 km from Earth to 10–6 mbar at 170 km away.

Craft for testing can have a mass of up to 270 tonnes and either be built in the chamber or brought in prefabricated via two 15 ×  15 m doors and three sets of standard-gauge rail tracks. The temperature can be adjusted from –150 to 50 °C by a “thermal surface” with an area of about 1500 m2. A craft is not placed in direct contact with the surface, which means that its temperature change is due to radiation heat transfer. The temperature-controlled surface is cooled by a huge container with a million litres of liquid nitrogen and two 2.2 MW gas compressors. Infrared quartz lamps — simulating the radiation heat transfer from the Sun — are arranged around the object to be tested.

The SPF chamber can be pumped down from atmospheric pressure to high vacuum in less than 20 h using a unique technique. “Roots pumps”, which consist of two figure-of-eight-shaped lobes or impellers that rotate at 1500–3500 rpm, are first used to quickly remove the large volume of air. These work not by the principle of gas expansion and compression but by moving a volume of gas. The pumping speed is set by the internal volume of the pump multiplied by the rotational speed of the pump, so a physically bigger pump has a higher pumping speed. Evacuating the chamber is done via groups of Roots pumps in parallel, known as “stages”, which have a common inlet and a common outlet pressure. For example, a two-stage system has a first stage consisting of a group of pumps that remove air from the vacuum chamber, and exhaust it into a second stage of pumps that then pumps the chamber to a lower pressure.

At the SPF the first stage has four Roots pumps arranged in parallel and connected to the vacuum chamber through a 1.2 m diameter pipe. This stage evacuates the chamber from atmospheric pressure to 600 mbar in about 20 min. Two Roots pumps in parallel with a 60 cm inlet piping then evacuate the chamber to 390 mbar in another 20 min before two Roots pumps with 45 cm diameter inlet piping reduce the pressure to about 170 mbar after an additional 20 min. The fourth stage, consisting of two pumps with 35 cm inlet piping, reduces the pressure to 90 mbar in a further 20 min. Finally, six rotary piston pumps in parallel with 15 cm inlet piping kick in to bring the chamber to 0.0133 mbar in an additional 30 min. Rotary piston pumps use a piston to expand and compress a gas, similar to an air compressor. The fifth stage then exhausts to atmosphere.

These five stages of pumps remove 53 tonnes of air, and then they are isolated from the chamber so that the helium cyropumps can come online to remove the remaining 700 g of air. Helium cryopumps operate by freezing the air inside the vacuum chamber and then collecting it on the cold surfaces of the pumps, which are at a temperature of 15 K. When enough ice has built up on the cold surfaces, the pumps must be warmed to release the air. This type of pump can only operate under vacuum conditions. After around 70 min the chamber vacuum level is reduced to 3 ×  10–5 mbar by 10 helium cryo-pumps. The vacuum level is reduced to less than 10–7 mbar in an additional 3–10 h.

We recently tested the air-bag landing system for the Mars Exploration Rover mission by simulating the conditions that would be experienced when landing on the Martian surface. This revealed that the air bags would be torn to shreds during the actual landing of the Spirit and Opportunity rovers. Without ground testing, and overcoming these issues with the help of the SPF, the rovers would have crash-landed on Mars at a cost of around $820m.