LEO Extra: A Little History / STS-104, Part 1

To start this post, I want you to think a little bit about water.  As everyone knows, water is a liquid.  In fact, if you think about the word “liquid” just for a couple of moments, you probably had an image of water in your head.  Water is liquid; liquid is water.

Okay, but what about ice?  “Yes,” you say, “but ice is ice and water is water.  That’s why we have two different words.”  That’s right.  Our language has been made to fit our experience, but we all know, of course, that ice is just frozen water.  And, of course, we know that steam is water made hot enough to boil and become gaseous.  We have three different words for three different states of the same chemical stuff: H2O, two atoms of hydrogen bound to a single atom of oxygen.  On our planet, at typical, habitable temperatures and given our atmospheric pressure at the surface where we live, water is liquid and the other states – gaseous and solid – are generated from there as deviations from “normal conditions.”  And that’s good since we otherwise wouldn’t exist as a species and, more importantly, nobody at all would be reading this blog.


But let’s suppose that we lived on a planet where the typical ambient temperature was, say, 300 degrees Fahrenheit, but everything else was mostly the same.  Please ignore, for a moment at least, all of the other issues arising from such a scenario and imagine what our sense of “water” would be.  All of the H2O that we would know of in our everyday lives would be gaseous.  The only way that we could get liquid water would be to chill some of the gaseous stuff down far below “normal” ambient temperatures, down below its boiling point.  And making solid water would require chilling even further, roughly to 270 degrees below “normal.”  If this was our world, then we probably wouldn’t have three separate words like “ice, water, steam.”  We would likely instead talk in terms of “solid H2O, liquid H2O, and gaseous H2O.”


Now, let’s suppose that rather than a hot planet, we lived on a planet with much higher atmospheric pressure (again, with everything else pretty much like it is on Earth).  In that case, we’d still have a general sense of water, steam, and ice, but our transition from liquid to gaseous would occur at a higher temperature.  Suppose that our atmospheric pressure was 1000 pounds per square inch (as opposed to our pleasant 14.7 psi on the surface of Earth), then our water wouldn’t boil until it reached well over 500 degrees Fahrenheit.  That’s 300 degrees higher than what we’re used to on Earth.  Kitchen stoves on this hypothetical, high-pressure planet would be using some serious energy just to make a bit of pasta for dinner.

So, what does all this supposing have to do with rocket engines?  Well, it has to do with thinking about the cryogenic fluids that we use for propellants.  When dealing with cryogenics, you have to think in terms of these topsy-turvy situations where things “boil” at four hundred degrees below zero and, in a rocket engine where we produce very high pressure situations, that boiling point in terms of temperature can be entirely situational or, above a certain pressure point, completely go away.  And, specifically, this is the background that you need to understand the curious case of STS-104.


STS-104 was a Space Shuttle mission that launched in July 2001.  It was a mission to Space Station and the orbiter was Atlantis.  Although I’d worked on the preparations for that launch and had participated in the flight readiness review at the engine project level, I was actually in Boston when it launched.  My wife had a veterinary conference and I tagged along because, well, Boston is just a really cool city to visit.  I remember sitting in the Boston Common, on a park bench, reading the newspaper article about the launch with some satisfaction knowing that I’d been involved in the process (along with, of course, hundreds and hundreds of other folks who, like me, justifiably took pride with each and every launch).


STS-104 was a special launch from a rocket engine perspective.  It was the first launch that included one example of a Block 2 Space Shuttle Main Engine.  This hardware configuration for the SSME was more than a dozen years in coming to fruition.  When I’d first started work supporting the Shuttle Program back in 1990, we were flying the Phase II version of the engine.  Gradually, over the subsequent decade-plus, NASA and Rocketdyne and Pratt & Whitney worked together to make that great engine even better, more reliable, safer.  The culmination, through several separate designations, was finally the Block 2 configuration.  Over the years, I had worked as an analyst during many of the stages of its development and, towards the very end, had been a genuine Datadog for the final certification testing of the complete package.  For that final process, we were turning around data reviews day in and day out as we stepped through two 22-test series at a rate as great as five tests every two weeks.  It was hectic but exciting.

The final piece that made the Block 2 configuration complete was the addition of the new high pressure fuel turbopump (HPFTP).  The following is a few fun facts about this remarkable machine:


Both the older turbopump and this new, Block 2 version had roughly the same performance from an engine power balance perspective, but the new one was safer.  That was the whole reason for the development effort.  Why was it safer?  Because over the twenty years preceding that point, we’d learned all kinds of stuff in terms of potential failure modes, effects, and mitigation methods.  Indeed, even the older, pre-Block 2 turbopump was not exactly the same one that had first propelled STS-1 in 1981.  We’d made small modifications all along to ensure that flight was as safe as possible, but the Block 2 design was a complete overhaul, an entirely new component.  Also, some of this additional reliability and safety came from the simple fact that the new HPFTP was stouter than the original.  Due to other modifications across the rest of the Space Shuttle vehicle, we were able to make the engine a bit heavier while still meeting mission objectives.  General Rule:  Give a designer a little more mass margin and he’ll give you larger factors of safety.  This additional mass for the HPFTP will be part of the overall story, so hold on to that fact.


Now, back to the story of STS-104.  When we last left our hero, he was sitting on a bench in the Boston Common, smoking a cigar and reading the newspaper.  He was all smug and pleased with himself in knowing that he had something to do with a positive story in that newspaper and in every other major newspaper across the country and some newspapers around the world.  I don’t think that it’s bragging to admit that it’s a pretty good feeling when something like that happens to you.

It wasn’t until I got back to work a couple of days later that I learned of the anomaly.  At first, it wasn’t even recognized as an anomaly.  It was more of an oddity in the data.  But all oddities have to be explained and as we dug into it we came to realize that it was indeed a real event, not some kind of data glitch, and that it was something we’d not predicted.  And then we figured out that it was something that needed to be remedied or future flights of the Block 2 engine would be in jeopardy as would the entire Shuttle Program.  Coming to resolution of this issue would consume much of the next few months of my professional life, but we did find a solution and we continued flying.  I will tell you more about it in Part 2 of this story in my next posting.

Post script:  Go out to Wikipedia and look up STS-104.  You’ll see there a little note provided about the Block 2 engine and an in-flight anomaly.  It doesn’t say much.  I’ll share with you “the rest of the story” (as the late Paul Harvey would say).

Inside the LEO Doghouse: RS-25 vs. J-2X

Nobody is confused by the fact that we don’t use a Ferrari 458 Spider sports car as a dump truck. Nobody is astonished that a Toyota Prius did not qualify for the Indianapolis 500 race this past May. And nobody whom I know drives a Caterpillar earth-moving truck back and forth from home to work (…but, I have to admit, it might be really cool to try – Outta my way, I’m coming through!).


We’re not confused by these things because most of us have automobiles and we are generally familiar with the notion of different vehicles being designed, built, and used for different purposes. In a number of different articles I’ve repeatedly stressed the notion that form follows function and function follows mission requirements. The mission requirements for a Ferrari are different than those of a Prius or those of a giant piece of mining equipment and so the resulting products are dramatically different.

The same concept of differentiation applies to rocket engines. That’s obvious, right?

On one end of the spectrum, you have something like the F-1 engine used for the Saturn V launch vehicle. It had a thrust level of 1.5 million pounds-force of thrust and a specific impulse of about 260 seconds (sea level). It stood nineteen feet high, at the base of the nozzle was over twelve feet in diameter, and it weighed over nine tons. On the end of the spectrum (at least the spectrum that we deal with within LEO), you have the RL10 which, depending on the specific configuration, puts out less than 25 thousand pounds-force of thrust but has a specific impulse over 450 seconds (vacuum). If you have an RL10 without the big nozzle extension, the engine is just over seven feet tall, about four feet in diameter, and it weighs less than 400 pounds.



Yes, that’s a flood of numbers, but let me make it a bit more graphic. If we wanted to get the same thrust level using RL10 engines as was obtained on the S-1C stage of the Saturn V (which used a cluster of five F-1 engines), then you would need 336 RL10 engines. That would be an interesting vehicle configuration indeed. Alternatively, try to imagine the Centaur upper stage – the typical use for the RL10 – with something as big F-1 hanging off the end. The whole stage weighs less than five thousand pounds (dry) and is just over forty feet long. If you tried to apply 1.5 million pounds-force of thrust to something like that, then in just fractions of a second, the whole stage would be a shiny metal grease spot in space.


So that brings me to the subject of this article. I want to compare the RS-25 and J-2X engines, currently the two primary products of the Liquid Engines Office. These two engines are not as radically different from each other as are the F-1 and the RL10, but the differences are substantial and meaningful. Here is a quickie table that will give you many of the basic characteristics of the two engines:


I know, I know. That all looks like a meaningless, banal listing of numbers. Specifications rarely seem interesting unless or until you know the stories behind the facts. So, let’s discuss the stories.

First of all, they’re both hydrogen engines. Why? Because they both need to have high specific impulse performance at high altitude and in space. The difference between these two engines is that the RS-25 is a sustainer engine whereas the J-2X is an upper stage engine. The RS-25 sustainer mission is to start on the ground and continue firing on through the entire vehicle ascent to orbit. The J-2X upper stage engine mission is to start at altitude, after vehicle staging, and propel the remaining part of the vehicle into orbit. Also, an upper stage engine can sometimes be used for a second firing in space to perform an orbital maneuver. This difference in missions accounts for the difference in raw power. The RS-25 is part of the propulsion system lifting a vehicle off the ground. It needs to be pretty powerful. The J-2X is the propulsion system for a vehicle already aloft and flying quickly across the sky.


The difference in missions is also largest part of the explanation for the different engine cycles used. In past articles, I’ve discussed the schematic differences between a gas-generator engine like the J-2X and a staged-combustion engine like RS-25. The staged-combustion engine is more complex but it generates very high performance. You may look at the two minimum specific values, 450.8 versus 448 seconds, and say that these are not very different, but remember that the J-2X cannot be started on the ground. If we tried to start the J-2X on the ground, the separation loads in the nozzle extension would rip it apart. The RS-25 achieves this very high performance without a nozzle extension because, well, it had to in order to fulfill the mission. Note that a ground-start version of J-2X would have a minimum specific impulse of something like 436 seconds.

Something else that is significantly different between the two engines is their throttling capabilities. The J-2X can perform a single step down in thrust level. This capability can be used to minimize vehicle loads or as part of a propellant utilization system since the throttle is accomplished via a mixture ratio shift. The RS-25, on the other hand, has a very broad throttle range. Why? Two reasons.

First, because during the first stage portion of any launch vehicle ascent, the vehicle experiences what’s known as a “max Q” condition. Perhaps if you’ve ever listened to a Shuttle launch in the past you’ll hear the announcer talk about “max Q” or “maximum dynamic pressure.” This is the point at which the force of the air on the structure of the vehicle is greatest. It is a combination of high speed and relatively dense air. Later, the vehicle will be flying faster, but at higher altitudes, the air is thinner. Thinner air means less pressure (the equation – thank you Mr. Bernoulli – says that dynamic pressure is proportional to the air density and to the square of the vehicle velocity). Thus, to minimize structural loads on the vehicle, the engines are throttled down deeply for a short period of time, and then brought back to full power. An upper stage engine operating only at high altitudes never has to face a max-Q condition. Second, a sustainer has to be big enough to contribute to the lift off of the ground, but at higher altitudes, after the vehicle has been emptied of most of its propellants, with too much thrust you’ll get too much acceleration. If you had no way to throttle back the engine thrust levels, then the vehicle would accelerate beyond the capacity of the astronauts to survive. An upper stage engine does not generally start out with as much oomph so the throttling needs to lessen acceleration loading is not as great.


Lastly, let’s talk about differences in engine control. Engine control typically refers to the parameters of thrust level and mixture ratio (i.e., the ratio of propellants, oxidizer to fuel, being consumed by the engine). When we talk about thrust, we are talking about throttling as discussed above, yes, but also thrust precision, i.e., the capability of the engine to hold tightly at a particular thrust level. When we talk about mixture ratio, we’re generally talking only about the notion of precision (but below in a post-script I’ll tell you a little more with regards to RS-25). Well, what would cause an engine to stray away from a fixed operational condition? Two things: boundary conditions and internal conditions.

The most obvious boundary conditions are pressure and temperature of the propellants coming into the engine. A sustainer engine can see a wide variation in propellant inlet conditions due to variations in vehicle acceleration. This is most dramatic during staging activities. An upper stage engine won’t typically see these wide variations. This is why it was very, very useful (almost necessary) for the RS-25 to be a closed-loop engine. A closed-loop engine uses particular measurements for feedback to control valves that, in turn, control engine thrust level and mixture ratio to tight ranges. The RS-25 holds true to the set thrust level and mixture ratio regardless of propellant inlet conditions. The J-2X, on the other hand, is an open-loop engine. The thrust and mixture ratio for J-2X will stray a bit with variations in propellant inlet conditions. Note that this “straying” is predictable and is built into the overall mission design. Because the upper stage engine won’t see the same wide variations in propellant inlet conditions, this is a plausible design solution.


The different control schemes for the two engines are also the reason as to why the noted thrust and mixture ratio precision are different in the table above for the RS-25 and the J-2X. Every engine runs slightly differently from firing to firing. These are usually small variations, but they are there. This is part of the “internal conditions” factor in terms of an engine straying from a fixed operational condition. A closed-loop control engine can measure where it is with regards to thrust and mixture ratio and make corrections to accommodate and compensate for slightly different internal conditions. An open-loop engine like J-2X cannot make these accommodations and so it will have a wider run-to-run variability even if everything else remains the same.

Note that we could have made J-2X a closed-loop engine. We made the specific decision to not go that way based upon a cost-benefit analysis. Simply put, closed-loop is more complex and, therefore, more expensive to develop and implement. We conducted a trade study, in conjunction with the stage development office, and decided that the benefits in overall stage performance did not justify the additional development and production cost. For RS-25, given its mission, it really had to be closed-loop from a technical perspective to enable the Space Shuttle mission. Plus – thank goodness for us today – the RS-25 control algorithms are validated and flight-proven as we head into the Space Launch System Program. That’s a nice feature of using a mature engine design.


So, that’s a top-level comparison of our two engines that we’re managing for the SLS Program. They have a number of common features, which is not surprising given that the SSME design grew out of the original J-2 experience forty-some years ago and the J-2X was developed, in part, with thirty-some years of SSME experience behind us. But they are also quite different machines because they were designed for different missions. No, this is not a case analogous to the comparison of a Ferrari and a dump truck. It’s more like, perhaps, a Bugatti Veyron and a Lamborghini Aventador. Each is just a remarkable creation in its own right (…and I’d probably be reasonably happy with either in my garage…).


Post-Script. A quick note about mixture ratio control and RS-25. You will note that the mixture ratio is shown as a range in the table of characteristics up above. The RS-25 can be set to run at any mixture ratio within that range. This is a nice accommodation for stage design efforts today as part of the SLS Program, but that’s not why the range exists. The original design requirements for the SSME included not only the provision for variable, controllable thrust level in run, but also for independently variable and controllable mixture ratio during engine firing. This fact, in turn, explains the rather unusual engine configuration of having two separate preburners, one for the fuel pump and one for the oxidizer pump. I’ll tell you why.


Think back to basic algebra. Remember when you had to solve for a number of variables using several equations. The mathematical rule of thumb was that you had to have as many independent equations as there were variables or else you could not arrive at unique solutions for each variable. The same principle is applicable here. With two separate, independently controlled preburners (and therefore independently controlled sources for turbine power), you can resolve to independently control two output parameters, namely thrust and mixture ratio. That’s pretty cool. But here’s the interesting historical part: We never actually used the shifting mixture ratio in flight. As the vehicle matured, it was decided that the mixture ratio shifting capability was not needed. But the design and development of the engine was too far down the road to backtrack and simply. Thus, we have a dual-preburner, staged combustion RS-25 engine.

Post-Post-Script. A number of years ago as part of the SSME project, for some specialized development testing, we did actually invoke the capability to shift mixture ratio in run on the test stand. So we have demonstrated this unique capability on an engine hot fire. There just wasn’t ever any reason to use it as part of a mission. Interesting.