Exciting times, an amazing bargain but not unexpected.
I'd thought he was waiting for the COTS2 and possibly 3 flights (several 100 million $) before they could start this. $70m is a snip for this task. Testing it with a real Dragon capsule launch will cost a bit more of course. You'd pretty much have to dedicate a full f9 launch to do it.
After all how does a rocket with no nose fly? Badly.
BTW anyone who saw his 1 hr talk before the National Press Club on the Spacex website might have known quite a bit of this. Not the landing *on* another planet but that the Dragon heat shield is spec'd for re-entry velocities up *to* a return from Mars to Earth. This may be easier than it sounds.
Once you've decided to have the TPS handle a re-entry from *lunar* orbit Mars may not be *that* much heavier. Note also the PicaX (Spacex's tweaked version of the NASA Ames Pica ablator) is 50-60% lighter than the Apollo Avcoat stuff (and they probably followed Max Fagets advice to only put it on where it's *necessary* not all over as the *very* conservative Apollo CM builders did)
The thrusters are *way* trickier. Here's some perspective.
Opening up Sutton 4th Ed (pg 18) shows the Apollo CM had 12 NTO/MMH each putting out 93lb, obviously in opposing pairs to stop any motion you'd started. Good for 230 secs (not sure if that's total across *all* thrusters or each).
OTOH the escape system main solid put out 147000Lb (for 8 sec) with steering by a 2 nozzle pitch control rocket (1 nozzle oversize, 1 undersize to steer the CM away from the exploding Saturn with *no* moving parts) putting out 2400Lb for 0.5sec. A 3rd one ejects the tower once its not needed. That puts out 31500Lb for 1 sec.
If *half* the 12 attitude thrusters could run for 230sec and point in a useful direction their total impulse (thrust x burn time) would *still* be < 1/9 the required impulse put out by the 2 solids which fire in an emergency.
Obviously Dragon is not Apollo but if the scale of response times, and hence forces needed are *roughly* similar you're looking at a system with normal mode at 1 and emergency mode at 10 units of thrust.
I'm aware of 3 precedents for this.
The Stentor engine driving the Blue Steel stand off missile had 2 separate chambers with a 4:1 thrust ratio, presumably driven by the same turbo machinery pumping Kerosene/Hydrogen Peroxide.
The old faithful RL10 (pumped LO2/LH2) has been throttled *down* to c 2% of main thrust
Perhaps more relevant the Lance missile achieved 10:1 thrust levels by using an ablatively cooled chamber-within-chamber design (both running at launch) using UDMH/IRFA forced out of their tanks by pistons driven by a solid propellant gas generator with sustainer throttling and steering by pulsed UDMH injection. chamber pressure was 950psi.
Bottom line. *Currently* wide throttling range *is* possible for pumped design or ones with high chamber pressures. Most thrusters are pressure fed with propellants and gas in seperate tanks (EG 5000psi for He, 200psi for the propellants leading to a 150psi chamber)
In the thruster world the state of the art *seems* to be the Boeing Canoga Park (AKA Rocketdyne) "Multi Use Thruster" throttling from 500-1100 Lb thrust using (allegedly) a 500psi chamber pressure to deliver a very small thruster with a claimed T/W of 290 but using the nearly unknown "Mixed-oxides-of-Nitrogen"
It's also likely to be a bit over the cutoff point that John whitehead at LLNL worked out for using reciprocating positive displacement pumps (although they do start and stop really well and would be fine at the attitude control level of thrust). Likewise Steve Harrington's team who confine the high pressure to their pump and a specific high pressure tank.
I hope Spacex don't chicken out and go with a solid solution, which they'd probably have to buy in. If they stick to a top mounted dual (triple if you include terminal maneuvers) use design this will move the state of the art *significantly* forward.
Exciting times.