Space Exploration Technologies Corporation (SpaceX) January 2005 Update

Oct 2004 through Jan 2005 Update

Testing, Testing and More Testing

The past four months have been focused on thoroughly testing all aspects of the vehicle, which will continue right up until we launch in a few months. The commitment I’ve made to all the engineers and our customers is that we will not launch until every engineer and technician in our company is two thumbs up. As you might expect, the more you test, the more problems you uncover that have to be solved. This results in a long tail on the end of development, analogous to the beta test period for software, but it is a lot better to solve problems on the ground than risk failure in flight.

In this process, I think I’ve come to realize what makes orbital rocket development so tough. It is not that any particular element is all that difficult, but rather that you are forced to develop a very complex product that can’t be fully tested in its real environment until launch and, when you do launch, there can be zero significant errors. Unlike other products, there is no chance of issuing a bug fix or recall after liftoff. You are also forced to use very narrow structural safety margins, compared to an aircraft or suborbital rocket, to have any chance of reaching orbit at all and must hit a bull’s eye when you do.

Having seen us go through the wringer to make this work (and it’s not over yet), I have a lot of respect for anyone that has tried to develop a serious launch vehicle.

Our test facility in Texas (structural test stand on the left and propulsion test stands on the right)

Engine Development Drawing to a Close

After two and a half years of hard work, we have essentially completed development of Merlin, our main engine, and Kestrel, our upper stage engine. It was an arduous journey, taking a lot longer and costing a lot more in R&D dollars than originally anticipated. Fortunately, recurring cost is in line with expectations and we have no plans to raise the price of our launch vehicles. Being able to amortize engine development costs over both Falcon I and Falcon V has been key to making the business case close.

It is worth noting that Merlin will be only the second American orbital booster engine developed in the past twenty-five years. The other one is Boeing’s RS-68, used on the Delta IV, and the one before that was the Space Shuttle Main Engine. For its part, Kestrel will be the first all new American upper stage engine in thirty years. More details are discussed below in the technical appendix.

Click for Merlin full duration run video

Reusability

I am increasingly confident of the reusability of the Falcon I first stage. We will soon exceed 200 cryogenic pressure cycles on the first stage tank mounted in Vertical Test Stand 1 in Texas and there are no signs of fatigue. The stage is also constantly wet by the water deluge system and by melting ice from the LOX tank, but is showing no significant corrosion.

Although we keep telling people that Falcon I has a reusability percentage roughly equal to that of the Space Shuttle, the only other semi-reusable flying today, it still gets referred to as an expendable rocket from time to time. I am quite confident at this point that Falcon I will not only be majority reusable, assuming the parachute opens, but that the economics will work out such that we may be able to make a modest reduction in price from the current $5.9M, which assumes no reusability value.

Launch Schedule

The updated launch schedule will now be posted on the Company Description page and we also have a FAQ page.

Futron Reliability Study

We asked Futron to perform a design reliability study, comparing all currently available US launch vehicles. It was limited to US vehicles, because that was where their database was most accurate and complete. The study examined all failures of the past twenty years by sub-system to derive a statistical reliability for that sub-system. Then, by adding up the failure rates for each sub-system in any given vehicle, you can determine its design reliability.

As common sense suggests, the simpler the vehicle, the less that can go wrong and the higher the reliability. According to this approach, Falcon I has the same design reliability as the simplest of the Boeing Delta IV and Lockheed Atlas V rockets, and Falcon V is considerably better (by virtue of an engine out capability).

Surrey Satellite Technology Limited (SSTL)

SpaceX recently acquired a 10% share of SSTL, one of the world’s leading small satellite companies. SSTL shares a similar philosophy of radically lowering the cost of space activity, so it makes sense for us to establish a closer relationship. It is very important that there be both low cost launch and low cost spacecraft to achieve a fundamental improvement in the economics of space. One without the other will not change the game.

No strategic deal was signed as part of the investment. SSTL will continue to launch on whichever rocket offers the best value and, by the same token, SpaceX will treat all satellite customers equally without preference to SSTL.

Falcon V Progress

Tanks – Aluminum for the tank walls and domes will be arriving from Alcoa in a few weeks. The Falcon V tanks will be welded using a comparatively advanced process called friction stir welding. The first tank will be stir welded along the longitudinal seams, where the highest stresses occur, and VPPA welded circumferentially. Our circumferential stir welding machine is due later this year, at which point the whole tank will be stir welded. From what I understand, that will make Falcon V the first rocket with fully stir welded tanks.

Engines – The engines for first flight of Falcon V are in fabrication and should come off the manufacturing line around June. These engines are a slightly improved and higher thrust version of Merlin 1A, appropriately called Merlin 1B. Static fire of the Falcon V first stage is planned for this summer on our large tripod test stand.

Avionics – Falcon V will have a triple redundant version of the Falcon I avionics and will move more in the direction of digital sensors and controls. Flight software will be much the same, with the significant differences being dealing with the possible loss of engines in flight and redundancy. We are also taking nothing for granted in assuming that it will be the same as Falcon I and intend to do a full hardware-in-the-loop simulation of the whole system.

Launch infrastructure – Most of our Vandenberg and Kwajalein construction and ground support equipment allow for launch of both Falcon I and Falcon V. We obviously need to build a new mobile launcher that supports a 12 ft wide rocket, but otherwise it will be just a question of adding more propellant storage tanks when it comes to launching Falcon V.

Elon

TECHNICAL UPDATE

Propulsion

Merlin

The good news is that we are done with core development and have a solid engine. After a lot of difficult development work, several component redesigns and the occasional RUD (rapid unscheduled disassembly), the flight engine, including our turbopump, combustion chamber, injector manifold, valves and controls, is working. The only fly in our ointment is that the pintle injector is not delivering quite the hoped for level of mixing efficiency at higher pressures (780 psi) and flow. We seem to have hit a ceiling at about 94% combustion efficiency, which is 2% below the target spec.

To make up for it, we have boosted nominal thrust from 72,000 lbs to 77,000 lbs (sea level). The higher thrust to weight on liftoff makes up for the drop in Isp, so payload performance to orbit is about the same. Nonetheless, it is a bit disappointing. Final vacuum Isp, including gas generator losses, will be around 304s rather than 310s. There is still some hope of tuning injector geometry and squeezing out more performance during the qualification program, otherwise improvements will have to wait until after first launch.

For the long term, we are evaluating other injector types, in particular a modified co-axial unique to SpaceX. Design is underway and we should have hotfire comparative test data this summer.

The next major engine development for SpaceX is the Merlin 2, where we will aim for a significant increase in thrust and chamber pressure. Merlin 2 will serve as an exact scale version of the F-1 class (>1,500,000 lb thrust) engine we intend to start developing in a few years. Target performance numbers will be released in the spring.

Kestrel

Kestrel, at the much lower 135 psi chamber pressure and only 7,000 lbs of thrust is performing well. All indications from development are that it will pass qualification at slightly above projected Isp, probably 327s rather than 325s. Our upper stage delivers more velocity increment than the boost stage, so this is really good news, helping to offset the Merlin underperformance. We have now done several full duration runs to demonstrate good ablative life, including putting 2.1 mission duty cycles on a single chamber.

One remaining challenge with Kestrel is efficiently bulge forming the niobium nozzle, which seemed like a fairly straightforward problem at first. Our approach to date has been to seam weld the niobium-hafnium sheets into a cone and then place the cone on a hydraulic expander to form the Rao contour, a technique that has been used since the days of Apollo. Unfortunately, this requires near perfect welds. We are getting to the right Rao contour by this method, but it is painstakingly slow and labor intensive. This method is not great for the long term, so we’re developing a new proprietary SpaceX technique in parallel.

Propulsion Test Stands

We now have three propulsion test stands in simultaneous operation and a fourth being built up for operation this summer. Horizontal test stand 1 and vertical test stands 1 and 2 are shown below:

Horizontal test stand 1 (subsystem tests)

Vertical test stand 1 (Merlin integrated engine tests)

Vertical test stand 2 (Kestrel testing)

For Falcon V stage hold down testing, we will be using our large structural test stand (VTS-3). This is rated for up to 3 million pounds of thrust and, if all goes well, we will one day use most of that capacity 🙂 It is hard to get a sense of scale from this aerial photo, but this test stand is of epic proportions, with foundations that go deep underground.

Large test stand (VTS-3)

Structures

As with the rest of the rocket, the past four months have involved constant testing of every aspect of the vehicle’s structural integrity, verifying both design and manufacturing processes. The four major vehicle load cases are:

1. Tank pressure
This is a minor factor for the first stage, where the design is stiffness driven. The first stage flight tanks we have installed in our test stand have now undergone over 160 pressure cycles with no sign of fatigue. Our upper stage, which is pressure driven and therefore has much tighter pressure margins, has undergone several pressure cycles, including a proof case to 24 psi above nominal operating pressure.

2. Ground winds
These turn the rocket into a giant lever arm when it is mounted on the pad just prior to launch. Failure modes are all at the base of the vehicle. We can currently withstand ground winds of up to 46 mph. If you can stand upright on the launch pad, it is O.K. to launch.

3. Maximum dynamic pressure at maximum angle of attack (max Q-Alpha)
This causes a bending moment and failure would most likely show itself as either a buckling failure in the avionics bay, which is just above the 2nd stage tank, or in the inter-stage.

4. Maximum g load at first stage burnout
At first stage burnout, the vehicle is accelerating at about 6.5 g’s and the upper stage is exerting about a 65,000 lb load on the inter-stage. Failure would show itself as buckling somewhere between the top of the first stage tank and the bottom of the second stage tank. The vehicle is essentially in vacuum at this point and moving at about 8,000 to 9,000 ft/s, so there is no significant aerodynamic load.

Only one major structural test remains (#4 above) and Falcon I will be a go for launch from a structures perspective.

In doing our structural qualification, we increased loads on the qualification article all the way to failure for max q-alpha, showing that we have about a 23% margin over limit load for that case. Limit load is the worst projected flight load if everything bad happens at the same time, which is very unlikely. For example, a 777’s wings at limit load would deflect almost 2.5 stories! Our intent was to design Falcon for launch in any conditions short of a weather system with a name (e.g. Hurricane Bob), so as a consequence we have excellent structural margins for normal weather conditions. Sitting on the launch pad waiting for good weather is expensive.

One of our biggest concerns, given how often it has been an issue with prior launch vehicles, is separation events. Both our stage separation system and fairing separation system passed qualification last year, but, as further verification, we will be conducting a series of 5 fairing and stage separations with a zero anomaly requirement.

Our primary structural test stand, shown below, is a seventy foot long steel brace with hydraulic actuators at various points to simulate flight loads. Here we are mounting the aluminum upper stage (coated in gold alodyne) to the carbon fiber inter-stage.

Main Structural Test Stand

Avionics

All flight avionics are now built, tested and integrated into the avionics bay. The test regime included thermal cycling, vibration, vacuum, shock, acceleration and salt fog. Issues were encountered and addressed in each test regime. Most of the tests were done in house, with our own vibration table, vacuum chamber and thermal chamber, with shock, acceleration and salt fog done at outside test facilities.

Click for video of avionics hardware undergoing a 13 g (twice flight g’s) load test

We also finished our hardware-in-the-loop simulator, which is a complete electrical and thrust vector control simulation of the vehicle. Using this simulator, we can show that our flight software can control the vehicle through ground and upper level winds, respond to gust and trans-sonic buffet, and reach the intended orbit.

Hardware in the loop simulator

For Falcon V, as part of the human rating of the vehicle, we are developing a triple redundant flight computer, using a more advanced backplane than Falcon I. The sensor and control system for Falcon V will also be considerably more digital, allowing for easy switchover from one computer to another without complex cross-wiring.