Todd May is not the typical NASA rocket boss. For starters, he never worked on the space shuttle program or any of the agency’s previous, half-hearted attempts to replace it.
A materials engineer by training, May spent nearly a decade of his 21 years with NASA working on the international space station before moving into space science. He helped lead the Marshall-based team credited with finally finishing and launching one of NASA’s perennial problem children, the $700 million Gravity Probe B mission to test Einstein’s theory of relativity. He then moved on to manage NASA’s Discovery and New Frontiers programs of cost-capped solar system exploration missions.
Today he is in charge of NASA’s $1.8 billion-a-year effort to field a human-rated, heavy-lift rocket slated to make its unmanned debut in 2017. He says his previous experience instilled in him the importance of meeting program schedules and budgets.
May spoke recently with SpaceNews Deputy Editor Brian Berger.
The Space Launch System (SLS) has its roots in the Constellation program, which was canceled on sustainability grounds. How is SLS more sustainable?
SLS is not quite the comprehensive program that Constellation was. Constellation included Ares 1, Ares 5 and a fairly large and ambitious lunar lander program. We’ve replaced Ares 1 with the Commercial Crew and Cargo Program. For SLS, we are using heritage hardware in smart ways to make the program more affordable and sustainable.
NASA’s overall budget outlook appears flat at best. How does that affect SLS?
From the outset, we realized you can’t do a bunch of parallel developments to get to the first flight. So that’s one of the things that helped define the architecture we have today. We’re minimizing the number of developments.
Where are you currently focusing SLS development dollars?
The core stage. We still have to get through preliminary and critical design reviews. We’re using surplus RS-25 engines from the space shuttle program. There’s no development curve there. The engines are sitting down at Stennis Space Center ready to be installed. But the core stage takes some development. It’s the same diameter as the shuttle’s external tank, but it’s also taller because of the addition of a fourth RS-25. The side-mounted boosters we’re using for the initial flights have 30 years of shuttle flight, and we made a lot of progress on the five-segment version as part of Constellation. So those are coming-down-type development curves that offset the core, which is actually the one thing that has to ramp up. But even there, we’re using reusing a lot of capability at the Michoud Assembly Facility and a lot of the same supplier base and avionics base we were using under Constellation.
Are you worried that core stage development issues will keep SLS from flying in 2017?
As any good manager should be, I’m always worried. We have more than six months of schedule margin for the 2017 flight. But we also have a number of things that are close to becoming critical path issues.
Can you give me an example?
Here’s one that might surprise you. Even though we’ve already got the RS-25s down at Stennis, we’re actually going to take the old controller, set it off to the side and make the J-2X [upper stage] controller the common controller for both the core stage and upper-stage engines as a cost savings measure.
If all goes well, when will NASA start bending metal on the SLS core stage?
We will have most of our tooling in at Michoud this year. We’re actually already running development welds and machining techniques here and at Michoud to prepare for the actual flight hardware.
NASA says the RS-25 engines you have on hand are worth more than $1 billion. But how much do you actually save in development costs by using them?
If you wanted to go with a brand new engine, you would have to spend the dollars to develop that engine. And not only that, you can’t really develop the engine and the vehicle in parallel. You tend to want to get your engine developed and fairly far along and then design your rocket around the engine. The first launch in most of those scenarios was after 2020. And you had to develop the engine, the core and sometimes the boosters, too, in parallel. That starts to stack up. When the reality of the flat budget comes into play, it pushes everything to the right even further. The total cost to first launch is also a lot higher.
Why not go with the RS-68 main engine, which is flying today on Delta 4?
It would have cost nearly a billion dollars to human rate that engine. There were other considerations as well. Going with the RS-68s drove you to the 33-foot [10-meter] diameter tank to be able to get the lift capability. That drove you to a different type of tooling than we have down at Michoud today and a different footprint at the pad.
What else is NASA doing to control SLS costs?
It starts with an understanding that we are in very austere budget times. On the government side, we’re going with a leaner insight/oversight model. We’ve reduced the government work force 36 percent from where we were in 2009. We’ve also, instead of bringing the whole standing army to bear on the contractor, we’ve empowered individuals to have conversations directly with a contractor counterpart. We’ve made the interface a lot more efficient.
The other thing we’ve done, particularly on the core development, is reduced the documentation required from the contractor by 80 percent. We’ve also reduced the percentage of deliverables that are Type 1 and made them Type 2 or Type 3. This puts more of the responsibility and the risk on the contractor side, which is what they said they wanted when we asked how we could best build this rocket.
Does that mean you’re cutting corners?
We aren’t compromising on safety or reliability. We are starting with technologies like the RS-25, which is the safest human-rated liquid engine in the world, and then we are going with the segmented solids, at least to start off, with 30 years of flight heritage. So we’re not cutting corners there. And if you take a company like Boeing with their background in the Delta 2, Delta 4 and international space station, they know how to build safe hardware. The question is whether or not everybody on the government side also has to check off on that design.
Would you still be able to reduce NASA oversight if you weren’t using as much heritage hardware?
The irony is that the heritage hardware — not the RS-25s, because those were already delivered to the government, but the J-2X and the boosters — actually have more deliverables on their contracts than the new core design because those designs were fairly far along and in some ways it actually costs money to undo requirements because those requirements have already found their way all the way down into the supplier base. So when you are starting fresh on a new core you have an opportunity to be more aggressive. The heritage stuff actually has more requirements on it.
You’re using the Delta 4 upper stage for the first couple of SLS flights. Where’s the incentive to switch to Apollo-heritage J-2X?
We are probably pushing the limits of what we are calling the Interim Cryogenic Propulsion Stage, or ICPS, on those initial flights. If we want to go deeper — depending on where we want to go, whether we want to land on the lunar surface or do Mars — a larger upper stage is ultimately needed. The ICPS is another way of avoiding simultaneous development curves because we can take an ICPS almost off the shelf.
How much will NASA spend on SLS development and what will it cost on a per-launch basis?
As an agency, the way we work is we don’t make cost commitments externally until we get through the preliminary design review and what we call Key Decision Point C.
And where are you now?
We’ve been through the systems requirements review, systems definition review and Key Decision Point B. Now we’re heading into a season of preliminary design reviews and that’s for the various pieces. That culminates next summer in a vehicle preliminary design review, which then feeds into Key Decision Point C next fall.
Are you still looking at one launch a year?
Yes, we are designing the way we build this so that it can be fairly optimized around a single launch per year. In a perfect world, we would fly maybe two launches a year. You get above that, though, and you actually start to stress some of the infrastructure.