Space faring was once a science fiction fantasy of the far future.  Now we live in that future and space activity permeates our economy.

The civilian space economy is surprisingly large.  According to the Satellite Industries Association, in 2018 the global space economy amounted to $360 billion.  Most of these revenues were derived from satellite information services and related ground equipment.  Satellite manufacturing and launch services amounted to $26 billion of the total.

This commercial economic activity mainly results from a multitude of very large satellites.  These satellites typically weigh tons and cost hundreds of millions of dollars, up to billions of dollars, each.  Most of these involve broadcast communications, which require them to be placed in geosynchronous orbits.  GPS satellites, GLONASS, and similar navigation satellites, in mid-Earth orbits, are government investments for military and public service use.

Deployment of the SBS-4 communications satellite (photo credit: Picryl)

An array of big satellites at high altitude is not the only interesting architecture.  Another approach is to have a large constellation of satellites in low orbits.  With enough satellites one can have continuous line-of-sight linkage from low orbit to any point on the Earth.  These dense low earth constellations have multiple advantages which make them suitable for cellular and internet communications.  Several companies, most notably SpaceX, OneWeb and Amazon, are taking this approach.  Their constellation sizes each involve thousands of relatively small satellites.

But small satellites are limited in what they can do.  There inevitably will be a desire to improve performance.  This means that even low earth satellites will grow in size to give the desired upgrade in performance.

Today, satellites and other spacecraft are manufactured on the ground and then launched into space.  Unfortunately, the launch environment is very punishing:  It has killer noise, killer shocks, and is broiling hot.  Equally unfortunate, big things are more structurally fragile than small things.  And, the more fragile something is, the more expensive it is to design, build and launch. 

On a per pound basis, therefore, large satellites may be more expensive than small satellites.  This is not the whole story, however, because serial production of large numbers of small satellites has a major beneficial impact on cost.

Typically, a large satellite costs about $400,000 per pound.  This includes amortizing the development cost over a small production run.  The technology of such a satellite is not more exotic than many high technology products that are not to be sent into space.  Cell phones, digital cameras, fighter airplanes, their munitions, and racing cars are comparable technology and cost about $2000 per pound, or less.  This suggests that it may be possible to reduce the cost of spacecraft down to $2000 per pound, or less.

Indeed, this is already happening.  I mentioned that small satellites are far less fragile than large ones.  Produce enough of the small satellites and their cost comes in line with ground-based devices.  The numbers released for the SpaceX Starlink constellation suggest that these satellites cost less than $2000 per pound.

So, how might we build large satellites at the same cost as terrestrial technology?  Something fundamentally new is needed.  It appears that final assembly of large satellites in manned orbital factories is the answer.  Let us explore this possibility.

There are a number of advantages to on-orbit final assembly.  Subassemblies can be shipped in protective boxes.  Properly packaged, even relatively fragile components can easily survive the highly stressing launch environment.  Orbital assembly eliminates the size and shape straight jacket that currently limits design configuration options.  Perhaps most important, orbital assembly will lead inevitably to mass production of standardized modules with major cost reduction as a consequence.

Any unmanned spacecraft, whether satellite or deep space explorer, consists of two functional parts:  the payload and the bus.  The payload is an instrument package that customizes the function of the spacecraft.  The bus provides the platform on which the payload resides.  In effect, the bus provides the “public utilities” of the spacecraft.  These include power, thermal management, communications, orientation control, and propulsion. 

Several commercial companies now provide standard buses of various sizes.  Because these buses are in serial production their development costs have been paid off.  Typically, these buses cost twenty to sixty thousand dollars per pound - the hostile launch environment, and the need for extreme reliability, being the major cost drivers.

It is important to recognize that a spacecraft which is assembled in space will likely look radically different from one assembled on the ground.  The compactness that today is needed for a satellite to fit on top of a rocket will no longer be required.  A space-assembled vehicle might possibly consist of an open truss onto which are mounted standard mass produced, low cost, plug-and-play modules. 

This brings us to the design of the orbital factory.  In the early days of aviation, it was discovered that the wings forward tails aft configuration that we have today was the best.  Like the airplane, a space factory’s configuration is driven by its function.  This is just the old industrial design slogan that Form Follows Function.  Thus, we may expect the factory to be much like the following:

The core of the factory will be a cylindrical pressure vessel.  A cylindrical form is needed so that the pressure vessel can be launched into orbit with strap-on rocket boosters.

The crew will work entirely inside the factory in a shirt sleeve environment.  Only in this way will their work be efficient.  Any outside work will be done with remote teleoperation manipulators. 

The cylindrical pressure vessel will be partitioned by an airtight bulkhead into a habitation section and a workshop.  Because it will be necessary to get an assembled spacecraft out of the workshop, there must be a suitably large door.  This will cap the end of the workshop’s section of the cylinder.  Logically it follows that the workshop will actually be an airlock.  In order to store the air from the workshop airlock, there must also be attached a very large air storage container. 

After the pressure vessel is launched it will be outfitted with various applique fittings such as protective tiles, solar panels, heat radiators and other odds and ends.  A separate manned service vehicle will be necessary for this outfitting.

An assortment of small utility vehicles, such as space tugs and remote manipulators, will complete the factory’s tool kit.

Other than the cost savings through assembling spacecraft, there are a number of other cost advantages to this factory concept.  In particular, two are important:

“No long screwdrivers” constrains today’s designer.  This means that once it is in orbit a satellite’s maintenance or repair is impossible.  An orbital factory eliminates this constraint.  A newly assembled spacecraft may be tested outside the factory before it is towed to its operating location.  If any part is not working properly it can be replaced before sending the satellite off.  Today, long service life means using costly ultrahigh reliability components and great redundancy.  With orbital factories it is possible to retrieve and refurbish these vehicles.  The cost savings should be considerable.

The cost of engineering development would also be greatly reduced.  Engineering spacecraft is very different from engineering terrestrial systems.  Terrestrial items are engineered through a sequence of prototypes which can be field tested for bug detection and eradication.  Today a spacecraft cannot be prototyped and field tested.  The consequence is a great deal of costly checking and cross checking of design concepts during development.  This requires time and a large staff of people. 

Engineers using an orbital facility would revert to the prototype style of low-cost engineering development.  Particularly important:  provisional designs can be inexpensively tested in their natural operating environment of space.  Vacuum testing on the ground is expensive and unrealistic.

Transportation to orbit will also fundamentally change.  Big satellites now require big rockets.  Modular assembly only needs small rockets.  Reusable large rockets have dropped delivery costs from $10,000 per pound down to around $1,000 per pound for a full manifest.  The industry believes that payloads could be delivered for $100 per pound, perhaps with the next generation of vehicles.  Smaller rockets make this goal much easier to achieve.  The space factory could be supplied with a small DC-3 class of vehicle having only a 3000 to 6000 pound payload.  Analysis suggests that ultimately the delivery cost might be as low as $20 to $30 per pound.  This is roughly the same as inflation adjusted airfreight in the middle of the twentieth century.

With their final assembly taking place in space, spacecraft costs might be reduced by orders of magnitude – especially for large spacecraft.  In response, space activity would increase by orders of magnitude.  Industrialization of space promises truly large-scale exploration and exploitation of the resources of the Solar System. 

Mr. Richards has half a century’s experience working on space and satellite projects.  Before retirement he was an Engineering Fellow at Raytheon Space and Airborne Systems Division.  There he was Chief Architect for Advanced Concepts and Architectures.  This article is a digest of presentations made at several AIAA conferences and to the Air Force Research Laboratory.