We Need a New (Cheaper) Upper Stage Engine

During my run for the US Congress, I did quite a bit of research on space policy and in the process I found out something a bit surprising: one of the most commonly used engines for expendable launch vehicles is also one of the most expensive!
Every time a Delta-IV rocket takes off, it uses an RL-10 engine to propel its second stage to orbit.   The retired Delta-III also used the RL-10 for the second stage.  The RL-10 also powers the Centaur upper stage, used by Atlas rockets and previously used by the Titan series of rockets.  The Centaur, which uses one or two RL-10 engines has been in use since the 1960’s and is a mainstay of US high energy upper stages.  Six RL-10 engines were used to power the early versions of the Saturn-I rocket.

The RL-10 is also under consideration as a component of future launch systems such as the Space Launch System, which is currently under development by NASA and between one and four RL-10 engines are planned for the Advanced Common Evolved Stage, a new upper stage planned for future launch systems.

It’s not hard to see why the RL-10 is such a popular rocket engine.  As a liquid-hydrogen engine, it’s perfect for upper stages and for use as an earth departure stage.   It has excellent specific impulse, amongst the best of any rocket engines of its type.   It can be restarted in flight and is extremely reliable.   Early versions produced 66 kN of thrust and more recent variations now push 110 kN, which makes it perfectly sized to most payloads.

Considering that the RL-10 has been in production for a half-century and is one of the most prolific engine types, one might think it would not be terribly expensive.

However, the RL-10 is actually one of the most expensive engines out there, costing a whopping $38 million per engine.

Granted, space hardware is never cheap, but even by rocket engine standards $38 million a pop is extremely high.

As compared to other rocket engines:

The Space Shuttle Main Engine – The Space Shuttle Main Engine is a much different animal than the RL-10.  For one thing, it produces 2,278 kN of thrust, more than 20 times that of the RL-10.  It’s not restartable in flight, because it was not designed for use as an upper stage engine.   The SSME has a vacuum Isp of 452.3 seconds, only slightly less than the 462 seconds of the version of the RL-10 used in the Delta Cryonic Second Stage.  It should also be noted that the SSME was never designed to be cheap, as it is reusable and the cost of each engine is spread over many missions.

The SSME costs about 50-60 million dollars per unit (for an engine 20 times more powerful and reusable)

The RS-68 – The RS-68 is the most powerful liquid hydrogen-fueled engine ever produced.  It is a much different engine than the RL-10.  Since the RS-68 is designed for use as a first stage engine, it is not restartable.  Unlike the SSME, it is not designed for reuse.   With 3,370 kN of thrust, it is more than thirty times as powerful as the RL-10.  The engine has a relatively low Isp of only 410 seconds.  Designers chose to use a simplified construction method for the engine bell, utilizing  a combination of ablative cooling and simplified cooling channels to make the engine cheaper than regeneration-cooled engines, sacrificing some efficiency in the process.

The RS-68 costs about 14 million dollars per unit.

The J-2X – The J-2X is an engine design intended for next-generation heavy lift and beyond earth orbit spacecraft.  It’s a redesign of the J-2 engine used on Saturn-V rockets.  Like the RL-10, the J-2X can be restarted in flight.   Its Isp is comperable to the RL-10 at 448 seconds.   The J-2X can produce 1,307 kN of thrust, almost twelve times as much as the RL-10.

Although only static test units have been built, the projected cost for the J-2X is twenty five million dollars – much less than the RL-10 for an engine of much greater thrust.

Given that there are few in-production engines to compare the RL-10 to, it’s hard to make a perfect apples-to-apples comparison, but the cost is clearly very high, even by rocket engine standards.

To put this in additional context:

A Centaur upper stage is used for heavy lift payloads with the Atlas-V rocket and is commonly used to propel payloads out of earth orbit in order to visit distant planets or the moon.  The two-engine version of the centaur has $76 million worth of RL-10 engine alone, for just the upper stage.

Launching a Delta-IV medium rocket costs about 140 million dollars.  Of that cost, about $38 million is just the cost of the RL-10 used in the upper stage.  The remaining cost of about 102 million covers the fuel, the transport and assembly of the rocket, the electronics, payload fairing, RS-68 engine used on the first stage etc.   That would seem to make the RL-10 the single most expensive component of the rocket.

The Advanced Common Evolved Stage has a great deal of potential.  As a large, modular, upper stage, it could be used to propel manned and unmanned spacecraft out of earth orbit.  It could become a highly capable, scalable stage for a variety of missions and has the capacity to be used as an in-orbit space tug.  It has also been proposed as the basis for a propellant depot. With four RL-10 engines, however, the ACES will cost more than $150 per unit, just for the engines alone!

But why is the RL-10 so expensive?

The answer to this turns out to have a lot to do with the age of the engine.  The newest versions of the RL-10 have updated the design of the engine’s nozel and incorporated new turbo pump technology, but overall, it is the same basic design that began testing in 1959.   As such, the design of the RL-10 does not benefit from newer assembly techniques.

The engine bell of the RL-10 is made of 360 individual tubes, which are cooled by liquid hydrogen.   This massive assembly of tubes is put together by hand over a bell-shaped mandrel.  Each tube is manually brazed by a worker until the engine bell is complete. The construction techniques used have not changed much since the 1950’s and do not make use of automated assembly systems, requiring a large amount of human labor to braze all of the tubes together.  If that does not sound expensive enough, the brazing process requires the use of pure silver.  (There are a number of reasons why the design requires silver brazing, including the high thermal conductivity of silver and the similar expansion coefficient to that of the stainless steel tubes.)

Despite the RL-10’s excellent efficiency, reliability and versatility, when it comes to the economics of its construction, there is clearly much room for improvement.

Possible alternatives to the RL-10:

In the late 1990’s, Pratt and Whitney began development of a new engine, intended to supplement the RL-10 and provide enhanced capabilities, while incorporating new production technologies.  The RL-50 was to have more than twice the power of the RL-10 in an engine of similar size.  During the early phase of development, several changes were made to the initial design of the RL-50.   The updated design, which was more powerful still, was dubbed the RL-60.

The RL-60 was intended from the start to be a direct replacement for the RL-10.   The dimensions are nearly identical, allowing it to easily be installed in existing RL-10 upper stages like the Centaur and Delta upper stages.   With maximum thrust of 290 kN, a single RL-60 would be capable of replacing two RL-10 engines.   The RL-60 design also meets or exceeds the efficiency of the RL-10, with vacuum Isp of up to 470 seconds.   Most importantly, the RL-60 would use the latest manufacturing techniques to reduce the production cost of the engine.  Unlike the RL-10, the RL-60’s regeneration cooled engine would be produced using a metal “sandwhich” production method developed by Volvo Aerospace.  Fabrication would replace hand-brazing with robotic laser welding.

By 2003, development of the RL-60 appeared to be going extremely well.  Several successful static tests had been conducted and the engine had met all performance goals. 60% of the components for the final version of the RL-60 were complete and Pratt and Whitney was confident the engine would be ready to fly by the end of 2005.

Unfortunately, shortly after these announcements, the RL-60 project seems to have slowed down and then been shelved.  It was not officially canceled and indeed the engine’ continued to be listed on Pratt and Whitney’s website for several years after, but no further updates were given on its status and all indications are that active development has been suspended.

No reason has ever been given for this, but it appears that lack of support and interest from NASA and the US Government is the primary reason why Pratt and Whitney stopped perusing the RL-60.  NASA and the US Air Force are the primary customers for the upper stage engines made by Pratt and Whitney, and the government seems to be uninterested in supporting the development of a new engine and perfectly happy continuing to use the old and expensive RL-10.

The loss of the Space Shuttle Columbia in 2003 resulted in a complete reevaluation of the future plans of NASA.  Over the past several years, plans for NASA have been in a state of constant flux, starting with the ambitious but underfunded Constellation Program, which was then revamped several times before being completely terminated.  Along this tangled path of proposed, canceled and repeatedly revamped and uncertain proposals, NASA seems to have lost all interest in improving the upper engines of current ELV’s and thus, there’s little reason for Pratt and Whitney to continue to develop the RL-60.

Yet while the seemingly-shelved RL-60 stands out as the best direct replacement for the RL-10, it is certainly not the only design with great potential.  Some of these, however, would require significant changes to upper stage designs or the fuel types used.

The Raptor engine is currently under development by SpaceX as a high energy upper stage engine.   Initially, the design called for a liquid hydrogen-fueled engine, but more recent revisions have shifted toward the use of methane for fuel.   This would require some changes to the design of upper stages like the Centaur, but does offer some advantages.  Methane liquifies at a higher temperatures than hydrogen, making it easier to handle and has a higher energy density, although a lower energy to mass ratio.  Thus, while the fuel would be heavier, the tank it is stored in could be smaller and lighter.

The thrust produced by the Raptor engine remains unknown, as the engine remains only an early design concept.   SpaceX has stated that their goal is for the engine to have an extremely impressive Isp of 380 seconds.  Given that SpaceX has a proven history of producing innovative, efficient and low cost engines, this is certainly one to watch.

Another interesting alternative to the RL-10 is the Chase-10 engine, produced by Darma Technologies, a relatively unknown engine producer.   Like the Raptor engine, the Chase-10 is fueled by liquid oxygen and methane.   The Isp of the Chase-10 is stated to be only 321 seconds, making it less efficient than the RL-10.   It also produces slightly lower thrust than the RL-10 at only 97 kN.  The Chase-10 has the interesting feature of being reusable.  This makes little difference in current upper stage designs, which are all intended to be expendable, but does present interesting possibilities.

An upper stage powered by the Chase-10 engine would therefore have less capacity than one powered by the the RL-10.  However, this could be offset by the fact that the cost of the engine is dramatically less than the RL-10.   The Chase-10 has a per unit cost of only three million dollars!  Even considering the limits to the engine’s performance, the fact that an engine in the class of the RL-10 could be produced at such a low price is quite amazing.

Why this should be a major priority for NASA:

It’s unfortunate, but right now, the future of NASA is not certain.  The agency has been plagued by a series of programs and initiatives which were started, only to be cut back or canceled entirely before reaching the point of deployment.  A great deal of money has been spent in the development of vehicles that never made it past the drawing board and on components for launch vehicles that ended up being canceled.

Developing better upper stage engines, however, is one area where this should not be a problem.   Regardless of what direction the space program ultimately takes and what kinds or missions the future will see, there is no doubt that upper stage engines will be required.   Since we already use the RL-10 so heavily, a cheaper, and ideally more capable engine, will have a tangible effect on the cost of launching payloads, especially those beyond earth orbit.   This is true for both NASA and the US Air Force, the two current primary users of the RL-10 engine.

An engine like the RL-60 is also inherently flexible and, as such, can be applied to many missions.  A cluster of four or more offers an alternative to the J-2x engine for manned missions beyond earth orbit.  If incorporated into the Advanced Common Evolved Stage could make it an even more flexible and capable stage for use in manned and unmanned missions.

Since high energy upper stage engines have limited use for commercial space launch systems, it is unlikely that the private sector will have sufficient incentive to develop and deploy next-generation engines like the RL-60 without the backing of NASA and the US Air Force.  Thankfully, there has finally been some renewed interest.  This year, NASA and the US Air Force joined forces to commission a new project to study advanced upper stage engines.

According to Dale Thomas, an associate director at the NASA Marshall Spaceflight center:

We know the list price on an RL-10.if you look at cost over time, a very large portion of the unit cost of the EELVs is attributable to the propulsion systems, and the RL-10 is a very old engine, and there’s a lot of craftwork associated with its manufacture.

I could not agree more! Now, please, lets make this a priority and actually get the engine flying before some short-sited politician pulls the plug on it!

Source: