I got introduced to the “alt.space” movement back in my freshman year via the usenet group sci.space.policy. I’ve stopped posting there (except for on a rare occasion) due to the fact that the S/N ratio is so horrendously lows, but I do poke my head in from time to time. Mostly I’ll look for posts from the few people there who typically have clueful things to say (a short list consisting mostly of Henry Spencer, Paul Dietz, Monte Davis, Derek Lyons, Jorge Frank, and a few others). Anyhow, earlier this morning, Monte forwarded me some comments that Jorge wrote that I thought was so spot-on that it deserved further discussion:
The conversation started with a discussion by someone else about the flaws of the shuttle:
Leopold Stotch wrote:
You sir go to the head of the class. The tandem design of the STS is fatally flawed. The LOX/LH2 propellants will of course cause condensation and ice formation. Chunks of ice falling off the ET and hitting the vehicle is a bad thing. This can be ameliorated somewhat by adding insulation to the outside of the tank, but now we have the issue of chunks of insulation (and some ice too) hitting the vehicle. Unless someone can solve the issue of ice formation *without* adding insulation than can shed and damage the vehicle, tandem stacking of launcher and vehicle is a bad idea.
In reply, here’s what Jorge said (my emphasis):
Parallel staging, or “tandem stacking” as you put it, is not the only dimension of the problem here. The ice formation/foam shedding only happens because of the use of cryogenic fuels in the tank, and it is only a hazard because the orbiter TPS is fragile. And the orbiter TPS is fragile because it is optimized for heat capacity, and it must be optimized for heat capacity because the orbiter is quite dense at re-entry… *because* it carries its ascent propellant externally.
Thus the astute designer sees that this is a multi-variable problem, and that parallel staging is a variable that cannot be analyzed in isolation, because it is dependent on so many other choices. Choice of propellant, internal versus external propellant tanks, etc. Had the orbiter been equipped with internal propellant tanks, the peak entry temperatures would have been lower and a more robust TPS could have been employed, negating the debris problem. Had the external tank used non-cryogenic propellants, propellant temperatures would have been higher and ice formation would not have been as much of a problem, allowing less insulation (or none) on the tank, thus eliminating the debris problem at the source.
Of course, there are tradeoffs with all these alternative approaches. Non-cryogenic propellants would have had lower Isp, necessitating more propellant mass (though not necessarily larger tanks, due to higher propellant density). Internal propellant tanks would have made vehicle controllability during landing a more difficult issue. We have never had the chance to learn the devils in the details of all these alternative designs, because we have never had the funding to investigate them. In the absence of real-world examples, the wishful thinking of armchair designers leads to a “grass is always greener” mentality: *every* alternative must be better than what has been tried already, because we know all too well the disadvantages of what we have tried, but of the road not taken, we can see only the advantages.
And *that* was the result of the decision that the space shuttle must be an “operational” vehicle, capable of carrying payloads. Because, the Mathematica analysis went, in order to be economical, that operational vehicle would need to be capable of 50-60 flights per year. And since the entire US launch market at the time was only 50-60 flights per year, that meant that vehicle needed to be able to carry every existing payload – civilian and military – to meet that flight rate and therefore be economical. But the size of vehicle necessary to carry all those payloads introduced design difficulties that made it impossible to meet the flight rate. Therein lies the paradox.
The real mistake of the space shuttle was not that of attempting a reusable vehicle, nor a winged vehicle, nor a parallel-staged vehicle. The real mistake is that we attempted to build an “operational” vehicle before we had any real idea of what “operability” means in a space vehicle. The alternative – the real “road not taken” – would have been to build small experimental vehicles, starting from suborbital and working our way up, that explore all the different “corners” of the design trade space resulting from this multi-variable problem, and learning, one painful step at a time, what works and what doesn’t. Since these experimental vehicles would neither have carried payloads nor flown operational missions, there would be no attachment to them; they would have flown for a few years each and then retired and replaced with the next X-vehicle, just as happened with all the previous X-vehicles up to and including the X-15.
That approach may or may not have resulted in a truly economical launch vehicle by 2007, but it would surely by now have given us a better picture of what works and what doesn’t than the road we chose. By attempting an “operational” reusable vehicle that by definition would have to replace all the existing “operational” expendable vehicles, we locked ourselves into a path that was difficult to reverse and was expensive enough that we could not afford to replace it in parallel with flying it, necessitating another long and painful gap in our experience base.
And because that one vehicle represents the whole of our operational experience for the last generation, its failure has led many to overgeneralize. The space shuttle is a (partially) reusable, winged vehicle with parallel staging using a cryogenic propellant tank. And it failed to meet its cost, schedule, and reliability goals. Therefore, the reasoning goes, all reusable vehicles are bad, all winged vehicles are bad, all parallel-staged vehicles are bad, all cryogenically fueled vehicles are bad. This is nonsense. Were the emotionally charged names to be replaced with faceless variable names, any competent mathematics professor would reject this logic as faulty, and rightly so.
It’s interesting how truly enlightening building and testing hardware really is. Without real tests, it’s impossible to truly know what will work and what won’t. That’s one of the big things I’ve learned over the past three years at Masten Space Systems–you never know how much you really don’t know until you try to find out. When we first were discussing starting up the company, we all had several assumptions. Assumptions about what would be easy and what would be hard, assumptions about how long things would take, how hard they would be to get working, assumptions about where the markets would be and how to best approach the development of these capabilities from a business standpoint.
Quite frankly, we were wrong on a lot of the assumptions. The good news is that at least on some of the assumptions we were wrong in a good way (overly conservative–overlooking possibilities, etc). I think this has been a common challenge for the emerging entrepreneurial space industry. All of us have tried to be cautious about our expectations, and even then have still come across as being overly optimistic. If you had told me back when I was still on my mission, and had just been told about XCOR’s EZ-Rocket flight, that they wouldn’t fly their 2nd generation vehicle for another 6 years, I would’ve laughed you to scorn.
However, one thing I’ve also learned is that while the path has been a lot rockier, longer, steeper, and daunting than it appeared from the beginning, that it also looks like its making the end result a lot better than it would’ve been otherwise. Take XCOR for instance. Over the past six years, they’ve gained a lot of experience. They’ve built almost a dozen engines, running on several propellant combinations in a wide variety of sizes. They’ve gained experience running a space business that is actually cashflow positive on a regular basis. They’ve built a solid base of experience, and have really refined a lot of their operations and techniques. Had someone given them the ~$10M they needed to build Xerus way back in 2001, I doubt they could’ve made a vehicle half as good as they can now.
So the real challenge as Monte likes pointing out, is how to get “there” from here? If we accept the proposition that we really don’t know as much about how to build a reliable, affordable orbital transportation system as we’d like to think, where does that leave us? How do we develop these intermediate prototypes that we’ll need to learn what we don’t yet know? How will we overcome the fact that the incentives inherent in government programs make X-programs very difficult to get funding for? How will we overcome the fact that most of the big aerospace companies are publicly held, and thus by legal necessity have to act in a risk averse way? How will we overcome the fact that us plucky upstarts are on a whole woefully undercapitalized, and how will we create a business case that actually closes that involves doing the learning we still need to do before we can really deliver on an reusable and operable orbital transportation system?
Those are some of the questions I’d like to discuss over the coming months.