Both Starship and the N1 have more than 30 rocket engines. The N1 exploded after having a chain reaction from a failed engine.

N1 rocket engines:

enter image description here

Starship’s rocket engines:

enter image description here

Unlike the N1, SpaceX has done multiple static fire tests with Starship to make sure it doesn’t explode. What else is SpaceX doing to ensure the same incident does not happen again?

  • 25
    $\begingroup$ Computers for controls and simulation. Also: Starship does explodes :D $\endgroup$
    – Antzi
    Commented Jan 24, 2023 at 8:58
  • 8
    $\begingroup$ No no, Antzi, they 'spontaneously rapidly disassemble.' $\endgroup$
    – GdD
    Commented Jan 24, 2023 at 9:04
  • 5
    $\begingroup$ anyone else having deja vu from when people compared the Falcon 9 to the N1, and then the Falcon Heavy to the N1 ... $\endgroup$
    – Erin Anne
    Commented Jan 24, 2023 at 19:49
  • 7
    $\begingroup$ @GdD "rapid unscheduled disassembly", RUD is the technical term ;) $\endgroup$ Commented Jan 25, 2023 at 6:19
  • 3
    $\begingroup$ SpaceX: I need you to fly up and down, SN10: And then explode! SpaceX: No, I need you to land, SN10: ah okay, and then explode! $\endgroup$ Commented Jan 25, 2023 at 6:29

4 Answers 4


One of the problems the N-1 faced was a lack of testing. They never did full up static fires, since the engines (NK-15) could only fire once after which valves needed to be replaced (multiplied by 30+ engines). Thus no way to really static fire.

The NK-33 was the newer engine designed to help with that, which was never used on the N-1 but was refurbed and used on the Antares booster as AJ-26 engines till they blew one of those up as well. (Then Antares switched to RD-191's and those became unavailble due to the war in Ukraine. Those guys cannot catch a break man!)

The only way the N-1 had to test was a full up launch.

If you have been following, SpaceX has been test firing Raptor engines and iterating the design like crazy the last few years.

They have been doing lots of tests (cryo, pressure, load, static) and are iterating through lots of builds. (Booster 4 and 8 were basically discarded and never will fly, Starships as well, as the design is superseded before they finish testing).

Computers and simulations today are unbelievably better today.

  • 8
    $\begingroup$ Also we have way more experience today. A falcon heavy has a total of 27 engines which is not that much short of the 30 odd of superheavy and N1. $\endgroup$
    – TrySCE2AUX
    Commented Jan 24, 2023 at 14:09
  • 10
    $\begingroup$ an additional problem of the N1 was the control logic. If it detected a failed engine it switched of the one directly opposite to compensate thus loosing 2 engines for each one failed (becaue if used only differential steering and not gimbaling nozzles). $\endgroup$
    – TrySCE2AUX
    Commented Jan 24, 2023 at 14:12
  • 9
    $\begingroup$ Perhaps one other thing to note is that the N-1 wasn't even assembled until it reached Baikonur - it had to be shipped overland from the manufacturers, rather than by barge as everyone else does. So any kind of integration testing was impossible, especially since the engines could only be fired once. $\endgroup$ Commented Jan 24, 2023 at 15:14
  • 4
    $\begingroup$ The design / redesign, build a little, test a little, rinse & repeat, (and sometimes, design / redesign, build a little, test a lot, rinse & repeat) design cycle SpaceX uses is fairly new. Agile simply didn't exist when the N-1 was built. $\endgroup$ Commented Jan 24, 2023 at 16:11
  • 2
    $\begingroup$ To be fair, the N1 engines could be fired and tested repeatedly, but they had to replace the valves after each test. I don’t know why they never did a full test fire of all engines mounted to the rocket at the same time. Maybe time pressure or replacing the valves on all of them would have been to difficult. $\endgroup$
    – Michael
    Commented Jan 24, 2023 at 20:44

The design of the N-1 was constrained by available materials

There's nothing inherently wrong with having lots and lots of engines. In the N-1's case, this was not because they wanted lots of engines, but because they had to.

The design of the N-1 was not simply motivated by "the best space rocket we can make." The command economic system of the USSR and trade restrictions with Western economies placed severe constraints on what materials were available, so the consideration was "the best space rocket we can make from the components we have available."

One would think that the government could just say "produce the materials we need." But it is not so simple. The Soviet economy was not one monolithic entity, but rather a set of ministries responsible for their own slice. So for example, the ministry responsible for rocketry did not produce its own aviation-grade aluminum, but had to use what was available from the ministry of aviation. The ministry could not simply go to another vendor.

This specific constraint (thin aluminum) led to a spherical fuel tank design with an additional streamlined body, compared to a conventional rocket where the fuel tank walls are also the walls of the body. This added weight and complexity, and required additional compromises from the rest of the rocket. I'm sure it was not the only such trade-off. As a result, the rocket design was far from optimal.

In the meantime, national prestige as well as internal rivalry between various factions made it impossible for one rocket design to keep getting built and tested until they got it right, so the funding dried up and was redirected elsewhere (unlike SpaceX, which could do this until it ran out of money and credit).


The key feature that allows SpaceX to succeed where Soviet Russia failed is the advent of computer simulation. The sophistication of today’s advanced simulation software is outstanding and enables a vast range of possibilities and options to be modelled in very great detail. No such software was available in the 1960’s.

Those interested should check this video

They use a number of highly ingenious techniques for increasing the resolution of the simulation and decreasing the required number of data points (hint adaptive grid) see sim at 36.30

  • 9
    $\begingroup$ Simulation helps, but it's only as good as your assumptions. It must be validated by real, physical testing. The Soviets did neither on anything like the scale SpaceX does. $\endgroup$
    – John Doty
    Commented Jan 24, 2023 at 17:05
  • 1
    $\begingroup$ True physical testing is absolutely essential, although simulation makes it easier to home in on the optimal solution and explore a wide range of contingencies. $\endgroup$
    – Slarty
    Commented Jan 25, 2023 at 6:51

There are two other things that differ from the N-1 vs modern rockets

Poor on-board computers

The Apollo 11 mission, for instance, had its famous 1202 alarm

Through exhaustive testing, the team at MIT’s Instrumentation Lab designed the computer such that it would never be full at any point in a mission. There would always be space available for the next program, rules in place to interrupt a program if something needed to be run immediately, or space to schedule the program after whatever was currently being run through the computer. But when Apollo 11 was descending towards the lunar surface, the computer ran out of Core Sets. This is where the 1201 and 1202 program alarms come in.

In other words, the Apollo computers (which were bleeding edge at the time) just couldn't process everything that was going on. Whatever device you are reading this post on not only has vastly more processing power, it's also built on well-known and tested technologies.

The N-1 rocket had KORD as its central control computer and... let's just say that the Soviet computer asked the Apollo computers to hold its beer while it tried something

The KORD was found to have a number of serious design flaws and poorly programmed logic. One unforeseen flaw was that its operating frequency, 1000 Hz, happened to perfectly coincide with vibration generated by the propulsion system, and the shutdown of Engine #12 at liftoff was believed to have been caused by pyrotechnic devices opening a valve, which produced a high-frequency oscillation that went into adjacent wiring and was assumed by the KORD to be an overspeed condition in the engine's turbopump. The wiring in Engine #12 was believed to be particularly vulnerable to this effect due to its length; however, other engines had similar wiring and were unaffected. Also, the system's operating voltage increased to 25V instead of the nominal 15V. The control wiring was relocated and coated with asbestos for fireproofing and the operating frequency changed. The launch escape system was activated and did its job properly, saving the mockup of the spacecraft. All subsequent flights had freon fire extinguishers installed next to every engine. According to Sergei Afanasiev, the logic of the command to shut down the entire cluster of 30 engines in Block A was incorrect in that instance, as the subsequent investigation revealed.

In other words, KORD got confused and decided to turn the rocket off and lock it all together. Whoops! At least the empty crew capsule returned safely... KORD wouldn't do as poorly in subsequent launches, but mostly because there were other things (like exploding engines) that were beyond the control of KORD.

SpaceX computers have not really been the cause of any mission failures that I know of (premature shutdown on Falcon 9 flight #4 might have been human error). Indeed, SpaceX rockets are capable of doing something no 1960's rocket could dream of: landing again for reuse later. SpaceX even has their capsules dock automatically with the ISS. SpaceX is now regularly using Falcon 9 rockets, with 9 engines. Coordinating more engines should not be significantly more difficult.

The Soviets were in a race

Remember that N1-3L was launched on Feb 21, 1969. By that point, the Apollo Program had already sent and returned Apollo 8 from the Moon, and Apollo 9 was just over a week away. N1-3L would be roughly equivalent to Apollo 6 (last uncrewed Apollo mission). N1-5L was then "launched" on July 3, 1969 (it didn't make it very far), around two weeks before Apollo 11 launched.

SpaceX is not in a race here to get Starship working. Indeed, they have been taking their time with the fully-stacked Starship. The only real competitor it faces is NASA's Space Launch System, which took nearly a decade to get one rocket off the ground, and is (for now) exclusive to the Artemis program (21st century Apollo redux). The SLS is also far more expensive and takes a long time to build, as none of it can be reused.

  • 2
    $\begingroup$ "SpaceX even has their capsules dock automatically with the ISS." That's a weird note to add, given that not only other modern spacecraft (ESA, JAXA etc.) but even Soviet ones have been capable of that for decades. $\endgroup$
    – molnarm
    Commented Jan 25, 2023 at 14:41
  • 1
    $\begingroup$ @molnarm The difference there is that the capsules would fly to the ISS and then the ISS would use its arm to grab the capsule and manually dock it. Unmanned Dragon capsules are capable of fully automatic docking and Roscosmos was skeptical of that initially $\endgroup$
    – Machavity
    Commented Jan 25, 2023 at 14:56
  • 3
    $\begingroup$ @Machavity, as far as I can tell, Kurs (developed prior to 1985) was fully automatic. Igla (first deployed in 1967) was fully automatic, but required both spacecraft and station to maneuver. $\endgroup$
    – Mark
    Commented Jan 26, 2023 at 4:33
  • 3
    $\begingroup$ The 1201 and 1202 alarms weren't caused by an inadequate guidance computer. They were caused by the radar units generating interrupts far faster than they should have. A modern computer faced with what the AGC dealt with would have simply given up, leading to an aborted landing or a crash. The AGC recognized that it was being overloaded and did a peculiar sort of reset that dumped the load while picking up where it left off. The AGC managed a load far higher than it was designed for with no more visible difficulty than a few warnings that it hit its limits. $\endgroup$
    – JRE
    Commented Jan 26, 2023 at 10:34
  • 1
    $\begingroup$ @JRE Modern computers aren't designed for Moon landings, so calling them out for not being able to handle Moon landings isn't very useful. $\endgroup$
    – Ian Kemp
    Commented Jan 26, 2023 at 16:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.