Starship flight 2: it went higher, it went

Starship flew again. The second integrated flight test. Seven months after the first attempt ended in an explosion at four minutes.

This time: all 33 Raptor engines fired. The booster performed correctly. Hot staging (lighting the upper stage engines while still attached to the booster) worked. The stages separated.

Then the booster exploded shortly after separation. And the upper stage lost contact and the flight termination system activated near the edge of space.

Two explosions. More data. Progress.

What worked

Hot staging. This was the big question. SpaceX decided to ignite the Starship upper stage engines while still attached to the Super Heavy booster, using the exhaust to push the stages apart. It’s a technique used by some Russian rockets but never at this scale.

It worked. The stages separated. The venting shield on top of Super Heavy held up (mostly). The exhaust pressure successfully pushed the ship away from the booster.

The booster also reached its target altitude and initiated its boostback burn. The engines relit. The booster was heading home when it experienced a “rapid unscheduled disassembly” (SpaceX’s preferred euphemism for explosion).

The upper stage reached space. Actual space. Altitude above the Karman line. It was performing its coast phase when the flight termination system triggered. SpaceX hasn’t said exactly why, but the ship was off-trajectory.

What the data means

NASASpaceflight broke down the telemetry. Compared to flight 1:

Flight 1: multiple engine failures during ascent, no stage separation, flight termination at 4 minutes.

Flight 2: all engines worked, successful stage separation, booster boostback initiated, ship reached space. Total flight time approximately 8 minutes.

That’s four more minutes and several more milestones. In rocket development, that’s a lot of data in one flight.

The iteration speed

What makes SpaceX’s approach different from traditional aerospace is the speed. Seven months between flights. Each flight builds on the failures of the previous one. The pad that was destroyed in flight 1? Rebuilt with a water deluge system. The engine failures? Addressed in the new vehicles. The stage separation that never happened? Now it uses hot staging.

Most aerospace companies would spend years analyzing a flight failure before flying again. SpaceX analyzes, fixes, and flies. The next vehicle is already on the pad before the previous flight’s post-mortem is finished.

This approach is expensive in hardware. Each flight costs a vehicle. But it’s cheap in time, and in rocket development, time is the most expensive resource. A year of analysis costs more than a vehicle.

The pattern

I’ve been watching this pattern since SN8. Fly. Explode. Learn. Fix. Fly again. Every flight gets further. Every explosion teaches more than the last.

SN8: belly flop worked, landing burn failed. SN9: same result. SN10: landed, tipped over, exploded. SN11: exploded in fog. SN15: landed. Flight 1: all engines lit, pad survived (barely). Flight 2: stage separation, ship reached space.

The curve points up. Not smoothly. Not without setbacks. But unmistakably up.

It still exploded. It still broke apart. But it broke apart further along, higher up, faster, and after achieving more milestones than the previous attempt.

That’s progress. Measured in altitude and explosions. Both going up.

Related thinking: