Last week, Sierra Nevada conducted the first free-flight test of its Dream Chaser passenger carrying spacecraft. During the test, involving a non-spaceworthy atmospheric test vehicle, Dream Chaser was carried aloft to a 12,500 foot altitude by helicopter and released to perform a glide test back to the runway at Edwards AFB. The purpose of the test was to determine the aerodynamic handling qualities of the design, controllability, stability and other characteristics of the spacecraft during the final phases before landing. A necessary goal, but not strictly a test objective, was for the test article to finish its flight undamaged on the runway.
Much ado has been made of events that occurred after the vehicle touched down. As we know, the left landing gear did not fully deploy. This resulted in a stable landing on the nose and right rear landing gear only. Initial rollout was smooth as Dream Chaser’s computers compensated for the offset loads. Eventually, the speed of the vehicle slowed to the point where the onboard computer could no longer keep the left side above the ground. It impacted the runway, whereupon the vehicle entered a combined skid and roll down the runway. The added drag on the left side and lack of steering caused Dream Chaser to skid the side and off the runway, where it stopped, upright but with some outer damage.
(Watch a replay of the Sierra Nevada post-test media teleconference with details of the flight)
It is important to note that the landing was a controlled landing. Never was the vehicle unstable and it did not land uncontrolled. There was no visible internal damage and the craft was still powered up and operating otherwise normally. By definition, the term “crash” means uncontrolled at the time of landing. By definition, it was a “hard” landing and not a “crash” landing either technically or qualitatively. The difference between the two depends on the combination of the sequence of events, whether it was under controlled flight, the level of damage and expected survivability had it been carrying passengers. By all accounts, the vehicle remained under control through touchdown and even as it skidded on the runway, the control systems were actively providing inputs and meeting the definition of control. The damage was primarily or solely exterior and no major systems were seriously damaged or destroyed. Had passengers been onboard, they would have walked away uninjured. That is a textbook hard landing, not a crash… sorry skeptics.
In the days that followed, a level of controversy and debate has erupted over whether the test was a success or failure and why a private company isn’t being as open with technical details as NASA, which is required by law to be open. Sadly, but predictably, the debate being promulgated by extremist supporters on each side (NewSpace and its critics) has been framed in all-or-nothing terms which bear little in common with engineering reality and ignore the fundamental purpose of conducting a test program in the first place.
On one side, there are those who say the Dream Chaser Approach and Landing Test (ALT-1) was not a success because of the landing gear failure. Essentially, those individuals are discounting the entire free flight portion of the test, which by all accounts more than met its test objectives. They are, in effect, saying the whole flight was a failure because the craft skidded on the runway. It is not terribly uncommon for a passenger-carrying aircraft to have a perfect flight only to experience one of its landing gear failing to deploy. Many times, the pilot must land the plane on only two of three or one of two landing gear – exactly the same as Dream Chaser. The plane rolls out and eventually one side tilts to the runway and the aircraft skids to a stop. Sometimes it ends up skidding off the runway. And the passengers are safe even though the plane takes on some damage. Was the flight a failure? No. Nobody ever makes the preposterous claim that since the plane skidded on landing that the flight should be characterized as a failure. Yet, that is exactly what some people are claiming about Dream Chaser. Those individuals are engaged in a double-standard by which Dream Chaser is required to operate outside of what engineers call “reality”.
On the other side are those who claim the test was a complete success with just a minor problem on landing. That too is incorrect. Downplaying a system failure doesn’t diminish the fact that a failure occurred. One of the landing gear failed to deploy and forced Dream Chaser to execute an off-nominal landing profile. “Off nominal” is not a good thing. It means there’s a problem that needs to be corrected; it’s as simple as that. The vehicle could as easily have rolled end over end and been totally destroyed. If it were an operational mission, at the very least Sierra Nevada would be facing the NTSB and likely grounding by the FAA. And so one cannot characterize the test as a 100 percent success.
But the flight met almost all of its test objectives. So, while the “flight” can’t be characterized as a complete success, that does not mean the TEST can’t be characterized as a success. The test is what occurred during flight to gather data on the vehicle. The flight itself was not “the” test. The two overlap but they are not one and the same and that’s what non-engineering pundits and protagonists fail to acknowledge.
In 1955, noted author and journalist Martin Caidin interviewed Gen. Donald M. Yates, who was commander of the Easter Test Range (Cape Canaveral) during the height of the missile race from 1954 to 1960. Caidin raised subject of what constitutes a failure, and the inaccurate perception in the media and public of what outwardly appeared to be failed tests but which, in reality, met most or all of their goals. Here is an excerpt of that interview. It still holds true today.
Martin Caidin: “General Yates, much has been said and written of the situation that exists here at AFMTC (Air Force Missile Test Center) in relation to so-called failures, and missiles that seem to fail, but are actually almost or wholly successful in their designed test objectives. I’ve used Atlas launches as examples; can you give me several others?”
Gen. Yates: “There’s no doubt that too many of our tests have been reported to the public as dismal failures. As you know, this is by no means the case – the Atlas is only one instance. The purpose of our tests is to gather data. Sometimes the period in which we can do this is measures in seconds or minutes rather than hours. The sole criterion of our tests is whether or not we are able to gather all the data we are seeking within this time span. The apparent performance of the missile or the distance it covers in its flight has no bearing on the situation. The Polaris is a good example. We fired one of these things in which the powered phase lasted exactly twenty-three seconds. After that, the rocket disintegrated. Aerodynamic forces acting on it after burnout literally tore it apart. But we were interested in gathering information on only the powered phase of the flight, and we obtained almost 100 percent of what we were seeking. That test, however, was reported as a failure. Viewers saw the spent shell of the missile disintegrate, and reported the test as unsuccessful. It wasn’t.”
“Now, testing determines flaw. Flaws in design or performance cause malfunctions; these can be demonstrated in many ways. In our tests they usually result in premature flight termination either by command or by accident. We learn something by all of them; sometimes, everything we’re looking for, at other times only part of the answer.”
“In one of our first Bomarc missiles the bird “failed” in the early part of its test. We found the failure to be a small potentiometer in the pitch-rate gyro system. We found it; we corrected it. And the engineers were happy about it.”
“One of our so-called unsuccessful Matador launches wasn’t even a launch. The bird was still on its launching platform when someone noticed a fuel leak. It looked like a fire that could destroy both the missile and its launcher, and perhaps killed some people. The test conductor literally kicked it off (using the booster only) its platform without its main engine.”
“We had a heartbreaker in one of our Navahos. the bird was climbing perfectly when suddenly the missile sent itself an engine shutoff signal. Nothing had happened, except that the fail-safe system built into the missile proved too safe. That’s a chance you take; the price you pay to meet the requirements for range safety.”
“We had a launch of a Thor missile that blew up during its climb. It was reported as a failure, but the bird actually had functioned perfectly. A million-to-one chance caused its destruction. As the missile climbed, the Range Safety Officer received an incorrect reading from his ground instrumentation; a technician had incorrectly wired some circuits. Immediately, the RSO pushed the destruct button. We found no flaws in the bird, but we certainly, if unintentionally, proved the efficacy of our ground-destruct system.”
Herein lies the problem. One side has taken the position that unless everything works perfectly, the entire test is a failure. The other side chooses to ignore, or at least compartmentalize, the obvious failure of a critical system as though that was something outside of the flight event. It’s all or nothing to both sides.
Well, if your position is that everything must be perfect or the whole thing is a failure, then you are patently incorrect and promoting an engineering impossibility. If your position is that the test was a 100 percent success because the post-landing rollout was not, strictly speaking, part of the test regime, then you are incorrect as well. If the vehicle is damaged, it’s not a perfect flight regardless of whether or not the test was a success, period.
The fact is, from an engineering standpoint – and in an engineering test, that is the only thing that counts – the Dream Chaser test was like almost all other tests of aerospace vehicles throughout history. It was neither a complete success nor was it a total failure. Attempts to claim otherwise in order to score political points is both disingenuous and damaging to the overall cause of space development.
If a test meets 98% of its objectives, then it’s 98% successful because that is the purpose of the test. For an operational vehicle, the standard is different – does it meet the business objective of safely delivering its payload. Even then, components and systems may fail but if it doesn’t result in a loss of the mission and its objectives, it’s not a total failure. Why are people making it all one way or all the other, as though the only two technical possibilities are 100% success or 100% failure? That is not engineering reality. In the engineering world, most often there is a combination of success and failure and rarely is something completely successful or a total loss. AT the end of space shuttle mission STS-51D, Discovery experienced multiple brake and tire failures that could have led to the loss of the vehicle. The brake problems were serious enough that post-Challenger, landings at KSC were stopped for several years while the problem was addressed. Was that a failure? YES – the brakes failed repeatedly. Were the missions failures? NO – the mission objectives were successfully accomplished.
The same standard should apply to Dream Chaser or any other vehicle.