It’s been over six years since I started working on the J-2X development effort. I missed the very first day that the notion of a J-2X engine was conceived, but I was only two weeks late to the party. So, I’ve been with the thing almost from the beginning. And throughout that entire period, whenever I get the chance to talk to people outside of our small, internal rocket engine community (…for very understandable reasons, they don’t let us out much), the single, most frequent, recurring, and ubiquitous question that I hear is something along the lines of this:
“How come you guys are spending so much time and effort recreating an engine that flew nearly fifty years ago?”
That is an entirely fair question. I am not a volunteer. As generous and as charitable as I like to consider myself, I do accept a paycheck. So do my coworkers. So does our contractor. Thus, all this work to develop J-2X isn’t free and, as I said, the question asked is therefore a valid point of discussion.
To a certain degree, I tried to answer this question by way of analogy in a J-2X Development Blog article posted a year and a half ago (December 2010) about a 1937 Ford Pickup truck. But analogies and metaphors can sometimes be abstruse. Let us eschew obfuscation and arrive expeditiously to the point: What makes J-2X different from J-2?
The J-2 rocket engine, developed by Rocketdyne and the NASA Marshall Space Flight Center, was qualified for flight in 1966. Between August 1966 and January 1970, 152 engines were produced. Between 1962 and 1971, some 3,000 engine tests were conducted. The J-2 engines were used for the second stage of the Saturn 1B vehicle and the second and third stages of the Saturn V vehicle. (Note that I wasn’t much involved in the original J-2 project considering that it was concluding just as I was figuring out that whole reading thing in First Grade. Remember Dick and Jane, Sally and Spot?)
The most significant differences between these two engines can be found in their performance requirements. I suggest that these are most significant because it is these differences that lead directly to a majority of the physical design differences between these two engines.
That’s an increase in thrust level of over 25%. And the specific impulse increase is on the order of 6%. While that doesn’t sound like much, in the realm of rocket engines, given that the J-2 and the J-2X are using the same power cycle, it’s huge. It means that we’re pulling staged-combustion or expander cycle levels of performance from a gas generator engine. That’s really something special.
From requirements flows form. Or, as stated by architect Louis Sullivan (mentor to Frank Lloyd Wright): “Form follows function.” You don’t design and build a rocket engine a certain way because it’s neato. It’s designed to meet requirements that fulfill mission objectives. It’s not like a 1959 Cadillac where stuff was added just because it looked really cool (picture below courtesy of the Antique Automobile Club of America Museum in Hershey, PA).
In order to get that kind of boost in performance for J-2X, we had to do two fundamental things: (1) move more propellant mass through the engine, and (2) use that propellant more efficiently. To the first point, J-2X pumps into itself and expels out approximately 20% more propellant per second than did J-2. That translates into needing a whole lot more pumping power. Here’s a comparison of power requirements for the J-2 and J-2X pumps (as a point of reference, a typical NASCAR engine generates about 750 horsepower):
That’s between 80% and 90% more power for J-2X as compared to J-2. The reason that you need so much more is not only the need for greater flow, but also the need for more efficiency in usage is manifested as higher discharge pressures. I’ll explain this further below. But first, let’s talk about the hydrogen pump just because it’s an interesting story.
Back in the day, when J-2 was first being conceived of, the technology of how exactly to pump liquid hydrogen was still being developed. The RL10 engine existed already, but it was about 1/10th the size of J-2. Some work had been done with pumping hydrogen as part of the NERVA nuclear thermal propulsion development effort, but not everything learned there was widely distributed. This relative lack of information resulted in J-2 having a liquid hydrogen pump that was, in reality, an axial compressor. You see, the problem is that liquid hydrogen is so light that it kinda sorta acts as much like a gas as a liquid. I’ve heard it described as being like whipped cream but less sticky.
So, do you pump it like a liquid or like a gas? You typically use axial compressors for gases. That’s what you use in turbojets for airplanes. And you can get it to work with liquid hydrogen, as J-2 clearly demonstrated, but it’s not the best solution. One of the issues is that a compressor has some unfortunate stall characteristics where the effectiveness of the pump can plummet during the start transient. This is caused by what is known as the start oscillation that always happens in liquid hydrogen engines. Picture this: Prior to start, everything up to the valves that hold back the flow on the hydrogen side is chilled down to liquid temperatures (typically 36 to 39 degrees Fahrenheit above absolute zero). Then the valves open during start sequence and the liquid hydrogen suddenly comes into contact with relatively warm downstream metal. The result is similar to what happens if you sprinkle water into a hot frying pan. In other words, the liquid boils immediately upon contact. In a rocket engine this causes a transient “blockage” as this voluminous plug of newly formed hydrogen gas gets pushed through the system. In terms of the pump, this sudden “plug” downstream results in a transient, elevated pressure at the pump discharge and this can cause the pump to stall, especially if it’s an axial compressor. In order to overcome this effect, they had to precede the J-2 start sequence with several seconds of dumping of liquid hydrogen through the whole system to pre-chill the metal downstream of the valves.
Okay, so that’s not too much of a big deal, but it was a nuisance. By the end of the 1960’s, it was clear to most folks that the better way to pump liquid hydrogen was to use a centrifugal pump and that’s the way we’ve done it ever since (including on the J-2S engine, which was an experimental engine tested in the early 1970s as a follow-on to J-2). With a centrifugal pumps, you get to avoid the stall issues inherent with an axial compressor and you get a more compact, powerful machine. Which is good considering how much more power we need to pull out of the pump for J-2X.
In addition to changing from axial to centrifugal, we had to make a number of other changes to the turbomachinery. In one place, we used to use on J-2 an Aluminum-Beryllium alloy. Well, you can’t use Beryllium anymore since it is considered too dangerous for the machinists working with the metal on the shop floor. In particular, Beryllium dust is toxic. And since we really like the guys working on the shop floor (as well as following the law), we had to go to another alloy. Also, we redesigned internal seal packages and rotor bearing supports using the most modern analysis and design tools and methods. In short, there’s not much in the turbomachinery, both fuel and oxidizer, that wasn’t reconsidered and redesigned to meet the imposed requirements.
Now, the other reason that we need 80% to 90% in addition to pumping 20% more “stuff,” is the fact that we had to get that stuff to higher pressures. Why? As discussed in a recent previous blog article, if we go to a higher combustion chamber pressure, then we can have a smaller throat and, with a smaller throat, we can have a larger expansion ratio without getting too out of hand with engine size. And, because of our extreme specific impulse requirement (remember: form follows function), we need that very large expansion ratio. So here are the top-level thrust chamber parameters:
The J-2 main combustion chamber was built from an array of tubes braze-welded together. When you needed the walls of that chamber to be actively cooled, this was the most common way to make combustion chambers “back in the day.” This is a fine method of construction, but it is kind of limited in terms of how much pressure it can contain. For the Space Shuttle Main Engine project in the early 1970’s, we needed the capability to handle a much higher chamber pressure and so we (i.e., Rocketdyne working in coordination with NASA) developed what is called a “channel-wall” construction method. So, to get the higher performance using the higher chamber pressure, we had to abandon the tube-wall construction method for the J-2X main combustion chamber and use a channel-wall main combustion chamber similar to the Space Shuttle Main Engine.
The main combustion chamber is on the top end of the scheme to get the larger expansion ratio. On the bottom end, we had to add a large nozzle extension. On the J-2, the nozzle consisted of another tube-wall construction. For J-2X, we have a tube-wall section that is actively cooled and then we have the radiation-cooled nozzle extension beyond that. The reason for transitioning is because the nozzle going out to a 92:1 expansion ratio has a diameter of nearly 10 feet and a tube-wall construction that large would be unreasonable heavy. In other words, from the vehicle perspective, the engine would be so heavy that its weight would offset any benefit from performance. The radiation-cooled nozzle extension is significantly lighter.
That make it sound easy, doesn’t it? If you want more performance, just strap on a big hunk of sheet metal and call it a nozzle extension. I wish that it were that easy. First, you need to figure out what material to use. Metal? Or maybe carbon composite? Plusses and minuses for both. Then you need to learn how to fabricate the thing light enough to be useful. And then you have to make it tough enough to survive the structural and thermal operating environments. In the pictures immediately above you can see a sample panel of how the J-2X nozzle extension is made and you can also see one of these samples sitting in a test facility where we blasted the panel with high velocity hot gases to partially simulate nozzle flow environments. The panel has a coating that enhances the radiation cooling so not only does the panel itself have to survive the environment, but so does the special coating.
Other things that we’re doing to get more performance out of the engine include the use of a higher density main injector and the use of supersonic injection of the turbine exhaust gases into the nozzle. When you talk about “injector density,” what you’re talking about is the number of individual injectors stuffed into a given space. Up to a point, the more injectors that you have, the better mixing you get, and, from that, the better performance you an extract from the combustion process. The picture below shows some testing that was done early on in the J-2X development effort to optimize the main injector density.
With regards to the turbine exhaust gas, on J-2 it was effectively dumped into the nozzle with the only intent being to not mess up the primary flow. For J-2X, we carefully designed the exhaust manifold and internal flow paths to get as even a distribution as possible around the nozzle and, from there, we are injecting it into the flow through mini throats at supersonic velocity. Here again we are extracting as much performance as we can given the simplicity of the power cycle.
The next element of the engine to consider is the thing that creates the power that drives the turbines…that spins the pumps…that feeds the injectors…that fill the chamber…that makes thrust. In other words, I’m talking about the gas generator.
So, due to the increased power needs of the pumps, the gas generator has to flow twice as much propellant and at higher pressures through the turbines as compared to J-2. The temperatures are pretty much the same since this parameter is mostly limited by material properties of the spinning turbine components. In terms of “form following function” from a design and development perspective, these increased power requirements translated to the fact that gas generator used for J-2 was entirely inappropriate for J-2X. It just wouldn’t work. Rocketdyne had to design a new gas generator based upon work that they had done as part of the development of the RS-68 rocket engine (used on the Delta IV vehicle). In the past, I’ve shown some pictures and even video of the whole development test series that we conducted to validate the design of our gas generator. Below is a representative picture of our gas generator component test bed.
Something not captured in the table of performance requirements way up above is the bevy of requirements imposed on the J-2X in terms of health monitoring and controls functionality. These too resulted in differences between J-2 and J-2X.
The J-2 engine had a sequencer to control the engine. Yes, it consisted of solid-state electronics, but other than that it was pretty much like the timer on your washing machine. The J-2X has an engine controller, which is a computer with embedded firmware and software that allows for a great deal of functionality in terms of engine control and system diagnostics. Some of these diagnostics we call redlines. These are specific limits that we place of measured parameters such that, should we break the limit, then we know that something bad has happened to the engine. The idea is to catch something bad before it turns into something potentially catastrophic. This is all part of the higher reliability and safety standards that have been applied to J-2X as compared to J-2.
The J-2X controller is composed of two independent channels such that if one fails, the other can take over. For critical measurements that inform the controller during engine operation, we actually take four separate measurements, compare them to make sure that they’re reasonable and good, and then use algorithms to perform the health checks. That’s one result of the imposition of more detailed requirements pertaining to reliability and safety. Along these same lines, we also have a number of design, construction, and workmanship standards that were applied to every aspect of the J-2X engine design, development, and fabrication. These standards, in combination with more evolved and advanced analysis tools, have, in a number of cases, further driven design changes away from heritage J-2 designs to what we’d call modern human-rated spaceflight hardware.
In an old J-2 manual, I found reference to a reliability value for that engine equivalent to 2,000 failures per one million missions. The requirement for J-2X is 800 failures per one million missions and, of those, only 200 can be “uncontained failures” meaning that the engine comes apart and potentially threatens other vehicle elements. So, all over the engine system we’re pushing more propellants, operating at higher pressures, generating more thrust, and squeezing out more performance efficiency, and we have to do this in a manner that results in an engine that has over twice as reliable as the heritage design. The result is an engine that is bigger and heavier than its historical antecedent:
So, in summary, here are the components that we had to change to meet J-2X requirements:
• Main injector
• Main combustion chamber
• Gas generator
• Added a nozzle extension
• Swapped the sequencer with a controller
What does that leave? Valves? Nope. Because of the higher flowrates and pressures, we had to drop the heritage designs for the valves and go to a design more akin to the Space Shuttle Main Engine. Ducts? Nope. Once you’ve changed all of these other things, you end up rearranging the connecting plumbing just as a matter of course. Even the flexible inlet ducts were changed slightly to accommodate more stringent design standards.
Form follows function; function flows from requirements; requirements flow from mission objectives. Different mission, different requirements, different function, and a different result. Thus, the J-2 and the J-2X share a name and share a heritage — in many ways the J-2 (and the J-2S) was the point of departure for the J-2X design — but the J-2X is truly its own engine. Lesson learned: Don’t assume too much from a name.
16 thoughts on “J-2X Extra: What's in a Name?”
I am so fascinated by all the intricate details. You really do an excellent job of explaining this so well.
I’ve been reading (slowly), some material on the NASA site about the Saturn V program and how the developed the different rocket engines for the different stages. In particular, they talk a lot about the F-1 (i think i mentioned it in one of my previous comments).
Anyway, they spoke about how they wanted to try to use current technology or “off the shelf” hardware whenever the could, in order to keep moving along. However, with the F-1 being so huge, you couldn’t just make a bigger part from an existing design. Once you decided to have more power, it meant more pressure, more mass, more everything. It changed the whole dynamics, so you’d basically have to re-design everything anyway.
Much the same as you explain above. So, when I read your most recent entry, i had some background as to why a J-2 and J-2X are only similar in name and heritage.
I really hope you engineers love these types of challenges that are put before you. It would drive me nuts. Not only do you have to design a rocket, but you have to make it more powerful, meet the design requirements, it has to work with the whole spacecraft, has to be safe for humans, and on and on.
Pretty soon, you’ll be asked to do all your currently doing, but with one hand tied behind your back! =-)
Keep up the great work and keep the posts coming!!
I second Bruce’s comments: this is fascinating stuff. Thanks for taking the time to write such a detailed post.
There’s one attribute of the J-2 engine that you definitely don’t want to inherit, POGO. How are rocket engines detuned, I think that’s the word, nowadays? Is CAD good enough to recognize the potential and help the engineers avoid it? Are some power cycles (GG,staged,expansion) ‘resistant’ to such things? Questions, questions, endless questions! 🙂
@Bruce: A quick story regarding “off the shelf.”
A number of years ago, a proposed vehicle subsystem used cryogenic helium. This was helium chilled to near liquid hydrogen temperatures. There was a flow control valve within the system and that valve would, necessarily, see these very cold temperature s. The prime contractor (whose name will go unmentioned) liked to brag about using off-the-shelf stuff and promised that their grassroots, down-home approach would bring oodles of savings to NASA on the project.
Early on, a subcontractor made a recommendation with regards to what kind of valve actuator to use for this system. The recommendation was for an extremely expensive, specialized piece of equipment. It had been used before, but it was expensive simply because each one was, effectively, hand-made to order.
“Nonsense!” declared the prime contractor. “We will use something off the shelf, something form the aircraft industry, something that’s pumped out by the thousands of units and so, therefore, cheap as dirt.” So, they connected to the valve a standard electro-mechanical actuator from the aviation industry.
Well, they set up a test facility to prove out their subsystem design. They called in the bigwigs and said, “Watch this!” The cryogenic system had been properly chilled down and was sitting there waiting to be actuated. At the appropriate time, the technician pushed the button and everyone expected the system to begin to flow.
Nothing happened. All of the screens showing data registered stillness.
After a flurry of confusion and hand wringing, the decision was made to increase the voltage. They did. And then they pushed the button again.
They went through this cycle one more time, again turned up the voltage and again pushed the useless button, but they simply could not get the valve to move. So, the bigwigs went off to dinner and other adventures while back in the lab, after a round of recriminations, an investigation was kicked off.
It turns out that the aviation standard actuator that they’d chosen was packed with grease. That’s fine for its intended use, but grease at cryogenic temperatures turns into a brick. They could have applied all the voltage of the Hoover Damn into that actuator and it never would have moved. Not wishing to admit defeat, the prime contractor declared that “slightly modified” off-the-shelf was still so much better than the silly, specialized, and expensive space hardware with which NASA is so enamored.
Over the next year and a half, the contractor engaged in a modification project looking to find ways to thermally isolate the valve actuator from the cryogenic system. In the end and to their technical credit, they got it to work. But the final price tag was substantially more than if we’d just followed the early advice of the subcontractor and used the expensive, not-off-the-shelf valve actuator hardware.
@Mark: Strictly speaking, POGO is not exclusively an engine instability. Rather, it is an instability indicating a coupling between the engine, the feed system, and the stage structure.
We deal with so many abbreviations and acronyms here at NASA that most people assume that “POGO” must be an acronym. It’s funny how many times the question gets asked, “What does ‘POGO’ stand for?” Well, actually, it doesn’t stand for anything. It is a reference to the motion of pogo stick: boing, boing, boing, boing…
POGO is caused when the heavy fluid column of liquid oxygen gets to bouncing in the feedline and that bouncing is enhanced by the engine response and then also begins to couple to the stage structure. With regards to the engine, the question is how it responds to fluctuating inlet conditions. How much does a pressure surge in the liquid oxygen inlet translate to surge in thrust? What is the timing of that response? Does that timing in any way relate to the natural frequency of the fluid column? And/or does that timing relate to any natural frequencies of the structure into which the thrust is delivered? The analysis for POGO is, then, a complex and coupled assessment. The slightly amusing thing about stage POGO analyses, however, is that almost every one eventually reveals the possibility of POGO cropping up to some degree and/or at some point in the mission.
In answer to your question about engine power cycles and POGO, I’ve never heard of any influences of one on the other. But that doesn’t mean that there might not be some level of connection.
In terms of dealing with POGO, the fix is rather standard at this point: You add a shock absorber. Another way of saying it is that you add compliance to the system in the form of a bubble. On the Space Shuttle Main Engine, the bubble is maintained in a sphere attached to the low pressure oxidizer line. The liquid level can fluctuate in and out and the bubble compresses and expands to fill the space to accommodate these fluctuations. Thus, the whole system becomes “softer” thereby effectively de-tuning it away from any dangerous, coupled modes. On other vehicles, the bubble is maintained in the feed system leading up to the engine rather than in the engine itself. You’ll note, for example, that the J-2X does not have a POGO accumulator within its schematic. We will depend upon the vehicle to provide any necessary compliance.
I love your blog and always look forward to your next post.
Do you know if there exists such a detailed record of the development of the SSME?
I highly recommend “Space Shuttle Main Engine The First Ten Years,” by Robert E. Biggs:
That was a fascinating link to the SSMEs! Does J-2X have as intricate a startup sequence as described for the SSMEs? Are the J-2X turbopumps as prone to runaway acceleration?
The start sequence for the SSME is complicated for a couple of reasons.
First, it’s difficult because it is a staged-combustion engine. This means that you have combustion chambers in series. When everything is steady state, serial combustion chambers seem just fine, but you have to ask yourself how you got to that steady state point. In what order did you light those chambers? What was the time between those ignition times? If the preburner driving the oxidizer side of the engine gets up and running too powerfully too early, you can push up your mixture ratios, make things far too hot, and get yourself in a dangerous place. If you light the preburners without getting the main chamber up and going, the pressure ratios over the turbines gets too high and, as you reference, you can over-speed the pumps. So the timing of the three ignition events is extremely important. And you tailor these events by skillfully controlling the oxidizer and fuel flows and hitting the spark at just the right time.
Second, the start sequence for SSME is difficult because you have no “cheat” system. No, nobody actually calls it that (except me), but on J-2X we have a spin-start system. We blow down high-pressure helium through the turbines to get things going and to pressurize the system. The helium spin-start system is like the starter on your car. It provides the “crank.” On SSME, you don’t have any crank. You have to slowly build up pressure by coordinating the spin up of your FOUR turbopumps (yes four, not two like J-2X) while avoiding pump stall and cavitation issues. As a point of comparison, the J-2X is up and into steady state operation by 3 seconds after receiving a start command. For the SSME, it takes nearly 6 seconds before you’ve truly settled into a steady state.
So, starting the SSME is not easy. Yes, we’ve figured it out, but the development of that start sequence was no small feat.
The possibility for runaway acceleration on the SSME high-pressure turbopumps is related to the engine cycle. In a staged-combustion engine, the upstream pressure for the turbines comes from the preburner(s) and the downstream pressure, the “back pressure,” is the main combustion chamber. Say that the normal, steady-state upstream pressure in the preburner is 5000 psia and the chamber pressure is 3000 psia. To first order, that’s a pressure ratio of 1.67. That’s where you want to run the turbine. However, let’s say that during the start your sequence gets messed up and your preburner reaches 5000 psia when your chamber is still only at 1000 psia. Now, your pressure ratio across the turbine is 5, you’re probably choking the flow where you didn’t want to be, and you’re probably accelerating far more than you’d like. So you see, in a staged-combustion engine, this “prone to runaway” situation is due to your back pressure. For a gas-generator cycle engine, the back pressure is, essentially, the pressure where you dump the turbine exhaust gas into the nozzle. In other words, the back pressure, even during the engine transient, is far less of a variable. So, in answer to your question, the J-2X does not have the same potential for runaway situations.
Will many of the J-2X team members then move to RS-25E development, or is that team already going? My understanding was that budgeting for the two programs was serial, which implies the same team might be involved. Should we assume the same lesson from J-2X, and not assume too much (or assume too little effort from an engineering standpoint) from just adding an “E” to the end of RS-25?
448 sec is quite an increase over the 310 sec of the Merlin. NASA still has a tendency of always trying to squeeze out the most efficiency possible, while SpaceX is throwing brute force at the problem, for the lowest cost.
@guest: The Merlin engine is a kerosene engine. The J-2X uses liquid hydrogen. Different missions, different requirements, different designs, and different performance. The Merlin is a boost-stage engine. The J-2X is an upper-stage engine. You will note, also, that early incarnations of the Merlin engine were strongly influenced by work done by NASA on the Fastrac (MC-1) engine here at MSFC.
@Matt: The RS-25 team and the J-2X team are, basically the same team under the current organization for the Space Launch System Program. We comprise a Liquid Engines Element within the program and are responsible for both engines.
With the fact that the RS-25 team is essentially the same as the J2-X team, will we be seeing a blog for that as well at some point in time? I’d be interested in the development of that rocket engine as well. With the RS-25 being the core stage, I’m guessing these will be very powerful rockets. It will be interesting to read about the challenges of “clustering”.
Also, congratulations on the record breaking test of the power pak!!
Does engine nomenclature (J-2X, RS-25) have meaning? i.e., are they codes for some important characteristic or feature of the engine?
@Steve: Most Rocketdyne engines typically have the designation of “RS-xx” (where “xx” is a number). Examples include RS-27 (used for years on Delta II vehicles), RS-68 (current Delta IV engine), and RS-25 (the engine known for years as the Space Shuttle Main Engine). I have been told by the Rocketdyne folks that the “RS” stands for “Rocket System” and was first instituted as a naming convention back in the days when Rocketdyne was part of North American Aviation.
So far as the Apollo era engines go, such as the F-1 and the J-2, I have heard all kinds of suggestions as to where those names came from. Note that there are other, similarly named engines in our history as well such as the H-1 (for the Saturn I vehicle and predecessor of the RS-27) and the M-1 (a proposed mammoth liquid hydrogen engine that didn’t get too far in development. The most plausible explanation that I’ve heard for the use of such bland, non-specific names was that these all started as military “black projects” that evolved into NASA activities. Thus, the suggestion is that they simply retained their original designations. If anyone knows different, I’d love to hear about it.
For J-2X, we started with our point of departure, J-2, and added the “X” as a means for pointing towards our “e-X-ploration” mission. Note that we did have some heated discussion at the time over six years ago about this topic since it is most often the case in NASA history that “X” is used to designate “e-X-perimental” stuff. Others wanted to change the name entirely, to something having nothing whatsoever to do with J-2, just to avoid the kind of confusion that my article attempted to address. We even had a little back and forth bit of consternation about whether we wanted to carry forward the hyphen within “J-2X.” Looking back it all seems kid of funny. In the end, we all saluted and agreed to “J-2X” and to the fact that a rose by any other names still smells as sweet.
Not to go down the rathole, but I think the letter designations came from the Silverstein Committee report on large launch vehicles in 1960 or thereabouts. There were numerous engine programs in the Air Force and ABMA and some arbitrary nomenclature was needed to track them all.
There were several others, including the A-3 (Pratt & Whitney RL-10), the E-1 400,000-pound-thrust engine for early Saturns, and the Y-1, which remains somewhat obscure. Rocketdyne also described an L-series of very large engines for Nova, but those were conceptual only.
@George: Thanks for input and the reference. Interesting reading.
Based on this site, however, the engine names existed before the Silverstein Report was published:
So, I’m still searching for where the non-descript letter-number names originated.
Comments are closed.