Discussing a system of regulating thrust through lighting/extinguishing engine pairs, I was met with a reply:

Really there is no need to treat engines as a pair in this context.

In fact, beyond ancient Soviet rockets I am unsure of any design that would do so. Engines can gimbal to compensate for a missing engine.

I'm aware the shuttle could compensate for one main engine malfunction through gimbaling, but I'm not aware of any other rockets where just one side engine could be switched off, and the rest would compensate the offset thrust through gimbaling, without disabling the engine directly opposite.

How common or uncommon is that solution? Can one, for example, just plausibly assume a rocket with 4 radially placed engines (central, 5th engine disabled for certain reasons) can continue flying straight on three engines with one of the side engines switched off (without knowing much more about the rocket)?

  • $\begingroup$ See Apollo 6: "A phenomenon known as pogo oscillation damaged some of the Rocketdyne J-2 engines in the second and third stages by rupturing internal fuel lines, causing two second-stage engines to shut down early. The vehicle's onboard guidance system was able to compensate by burning the second and third stages longer, though the resulting parking orbit was more elliptical than planned." $\endgroup$
    – Uwe
    Commented Sep 24, 2018 at 21:16
  • $\begingroup$ The posted answer covers this well. I just wanted to add that you should consider the viability of a solution that entails shutting off an engine after one has failed already. That sounds like a spectacularly bad idea to me. $\endgroup$ Commented Sep 24, 2018 at 22:48
  • $\begingroup$ You also have to draw a clear distinction when taking about "one side being switched off" between booster rockets like the shuttle SRB and the main engines. Both could be gimballed, but any SRB failure was non-recoverable. By contrast, gimballing of the shuttle main engines was sufficient to support two-engine out operations. (basically, the gimbal range of any engine is sufficient to point through the CG, either individually, or in combination with any of the other engines -- provided GNC does its job to prevent attitude excursions outside the recoverable envelope) $\endgroup$ Commented Sep 26, 2018 at 5:56

1 Answer 1


How common or uncommon is that solution? Can one, for example, just plausibly assume a rocket with 4 radially placed engines (central, 5th engine disabled for certain reasons) can continue flying straight on three engines with one of the side engines switched off (without knowing much more about the rocket)?

It's very common. I'd hesitate to say you can "just plausibly assume" it, but the ability to survive an engine-out is a very useful feature and it's almost free to engineer if you're assuming a cluster of 4 or more gimbaled engines. If the engines are not too widely mounted, you only need a small amount of additional gimbaling range (beyond what's needed for nominal steering) to compensate for the lost engine.

There are a few launchers with proven outboard-engine-out capability via gimbaling.

  • Saturn V: Apollo 6 lost not one but two outboard engines on the second stage and was able to continue ascent to a somewhat lopsided orbit on the remaining 3. The failed outboards (#2 and #3) were adjacent rather than opposed, so thrust was unbalanced. Engine #2 began leaking fuel and losing power from about 70 seconds into the second-stage burn, but struggled on for a few more minutes before failing completely at T+413, nearly four and a half minutes into the second-stage burn. An oxidizer shutoff valve for #3 was mistakenly wired to #2's controls, so when the engine management system tried to shut off propellant flow to the failed engine, it killed #3 as well. (Further exhaustive details in Rocketdyne's failure analysis document.) If both engines had been lost early in the burn, it would have been a loss-of-vehicle situation.

  • Falcon 9: SpaceX's CRS-1 flight lost one corner engine (on the old 3x3 Falcon 9 1.0 layout) and continued ascent.

  • Saturn I: Flights SA-4 and AS-101 flew successful missions with a single engine out late in first-stage flight -- SA-4 as a test, AS-101 accidentally.

One clustered-engine launcher I'm not sure of is the Proton rocket; it's had several launch failures on the first stage, but I don't think any of them were uncomplicated single engine-outs. The 6 RD-253/RD-275 first-stage engines each gimbal only in a single plane, so there are limits to how much they could contribute to compensation for a lost engine.

Soyuz might not be able to manage loss of a booster engine. Despite the way the booster bodies are angled, the engines themselves seem to be fixed parallel to the longitudinal axis of the rocket, rather than pointing through the CG; stability is maintained with small swiveling vernier rocket nozzles rather than by gimbaling the primary nozzles, so it may not have the control authority required to compensate for a lost outboard engine. In any case, the remaining fuel in a shut-down booster would go unused (there's no cross-feed facility) which would likely fail the mission even if the rocket maintained stability.

The liquid boosters on the Long March 3B are at least pointed through the CG (as discussed here), so losing one wouldn't destabilize the rocket severely; as with Soyuz, the fuel in that booster would be wasted.

By contrast, the only launcher I know of that was designed to handle an engine-out by shutting down the opposing engine was the N-1. Its first-stage engines weren't gimbaled; steering was achieved with differential throttle and grid fins. Losing an engine thus almost demanded that the opposing engine be shut down, but with 30 first-stage engines, the loss of two wouldn't be a problem after the first few seconds of flight.

  • $\begingroup$ Hand-drawn diagram in this Apollo 9 technical information summary shows 2 & 3 as adjacent rather than opposed: ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19700009516.pdf $\endgroup$ Commented Sep 24, 2018 at 22:13
  • 2
    $\begingroup$ Much more common; the only launcher I know of that was designed to shut off an opposing engine on loss was the N-1, with its 30 first-stage engines. $\endgroup$ Commented Sep 24, 2018 at 22:49
  • 1
    $\begingroup$ @MagicOctopusUrn: Are ion engines even used with gimbal and in number greater than 1? (other than the colloid gel RCS thrusters). You can vector the thrust a bit using electrostatic deflection, and extra thrusters only add mass and energy requirements, giving thrust but not specific impulse. $\endgroup$
    – SF.
    Commented Sep 25, 2018 at 14:02
  • 1
    $\begingroup$ Not gimbaled, but the Dawn spacecraft had 3 ion engines mounted at different angles, each pointed through the CG, with the expectation that it would switch from one to another as they wore out. Only one was used at a time. $\endgroup$ Commented Sep 25, 2018 at 14:24
  • 1
    $\begingroup$ @OrganicMarble I was really surprised to see how long the #2 kept firing after damage to both the injector and thrust chamber wall. $\endgroup$ Commented May 9, 2019 at 16:57

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.