The superdracos in Dragon's launch escape system each produce $\color{black}{\texttt{71.2 kN}}$ (16,000 lbf) of thrust. There are eight superdracos, giving a combined thrust of $\color{black}{\texttt{570 kN}}$ (128,000 lbf).

If you assume a dry mass of $\color{black}{\texttt{9,570 kg}}$ (21,100 lbm) for Dragon, an empty mass of $\color{black}{\texttt{400 kg (880 lbm)}}$ for the payload trunk, a payload max of $\color{black}{\texttt{5,900 kg}}$ (13,000 lbm), and a combined MMH/NTO fuel capacity of $\color{black}{\texttt{2000 kg}}$, you get a total mass of $\color{black}{\texttt{17,870 kg}}$ (39,400 lbm).

Now consider the most demanding launch escape case: with Falcon at a peak acceleration of 3g and engines still running. At a minimum, you would need at least a hair over 3g in order to speed away, right?

Well, if you divide the superdracos' total thrust by the fully loaded Dragon's total mass, you get..

$$\frac{570,000 \texttt{ N}}{17,870 \texttt{ kg}} = 31.9 \frac{\texttt{m}}{\texttt{s}^2} = 3.25 \times \left( 9.8 \frac{\texttt{m}}{\texttt{s}^2} \right) = \textbf{3.25} \texttt{ g}$$

And this has me wondering: Does launch escape impose an upper bound on the payload capacity that a Falcon + Crew Dragon assembly can carry to the ISS?

I'd always imagined that capacity was entirely determined by orbit specs and stage 1/2 fuel capacity, so that if your orbit was less demanding (launching straight eastward to low altitude, say) then you could launch with less fuel and stuff the Dragon trunk with whatever fuel mass you didn't load...

But now that doesn't at all seem the case... Because even if you launched with 10,000 kg less in stage 1/2 fuel, say, you still wouldn't be able to carry more than 5,900 ish kg while ensuring the crew can positively accelerate away from a failing rocket on launch escape.

Can someone confirm if this is true---that even if you could lift more load to orbit, you might be forced not to, in order to satisfy your launch escape requirements?


3.25g is just a lower bound on the thrust acceleration you'd get from a Dragon + trunk + max payload + max fuel.

For an upper bound, you could assume a super light load approaching 0 kg. And in that extreme case, your superdracos would buy you ~50 m/s2, or just about 5 g's of acceleration.

And as your fuel load approaches 0 kg---at the end of your launch escape burn---your thrust acceleration would approach ~60 m/s2, or 6 g's.

So Dragon's launch escape acceleration would fall in the range [3.25, 6] g---though if you're maximizing load, you're likely to fall on the lower end of this range, and if your payload is over ~5900 kg, then you fall below your 3.25 g minimum and risk not having enough acceleration to positively speed away from a failing rocket in the worst-case scenario.


I ignored the slant of the superdracos, which I believe is 15 deg to the vertical. This would make their combined thrust a bit less: $570 \texttt{ kN} * \texttt{cos}(15 \texttt{ deg}) = 550 \texttt{ kN}$.

This would give a thrust acceleration of 3.15 g for a fully loaded Dragon, if I stick to my mass numbers.

But this article shared by @BrendanLuke in the comments (thanks!) quotes the NASA administrator saying 3.5 g was the max thrust acceleration during launch escape tests (which I believe were done for the worst-case scenario with the rocket going ~ 3gs?).

This would mean a lighter Dragon + trunk + payload + fuel. The actual mass would have to be closer to (3.15 g) / (3.5 g) = 0.9 x my mass estimate, or ~ 16,070 kg instead of 17,870 kg---some 10% lighter. The lighter mass would probably come from a lighter Dragon capsule or from a lighter payload, since the trunk is already a super light 400 kg, and my fuel mass estimate is halfway between the lightest and highest numbers I've seen (1350 kg and 2,500 kg, if I remember correctly)?

And your peak acceleration with slanted thrusters and lighter mass would then be 5.5 g instead of 5 g with zero payload, approaching 7 g as fuel tanks empty out.

This would put launch escape thrust acceleration in the range [3.5, 7] g. But the main point remains: that accelerations over 3.5 g are possible only at payload mass under 5,900 kg (now maybe less as mentioned above), so even if the Falcon launcher could support more payload mass, you wouldn't be able to carry it without jeopardizing the astronauts' ability to escape in the worst-case scenario.

  • $\begingroup$ Well - we can just turn around your argument and say the performance of Falcon 9 limited the (useful) performance of the escape system. You can't construct a causality from the fact that the two systems match in performance. $\endgroup$
    – asdfex
    Apr 29, 2021 at 17:48
  • 1
    $\begingroup$ Sure, the launch escape specs would come from your expected mission profiles, etc. But once you've selected your launch escape specs, they're fixed for all missions. There is no alternate set of launch escape thrusters you could use. It's just the 570 kN superdraco clusters. So once the launch escape system has been designed, it becomes a hard constraint on how much payload you can safely launch while retaining the ability to escape. $\endgroup$
    – user39728
    Apr 29, 2021 at 17:58
  • 1
    $\begingroup$ I think there's more nuance here, the engines are canted outwards at some angle limiting their effective 'vertical' thrust, but then this article from CollectSPACE quotes Jim Bridenstine saying peak g's of 3.5 during the inflight abort test of (a notedly stripped down) Dragon $\endgroup$ Apr 29, 2021 at 18:38
  • 1
    $\begingroup$ Ah! You're right, Brendan. Only some of that max thrust goes into your net thrust acceleration. Thanks for the quote---I actually hadn't read anything on how fast the Dragon would accelerate during launch escape. 3.5g seems very reasonable. All my masses came from various places on the web, so they're probably off by some. $\endgroup$
    – user39728
    Apr 29, 2021 at 18:48
  • 1
    $\begingroup$ A 15 deg slant to the vertical would mean a total thrust of 570 kN x cos(15 deg), or 550 kN. With my mass numbers, this would mean 3.15 g of thrust acceleration. So (3.15 g)/(3.5 g) would mean 0.9x the total mass I assumed... or ~1800 kg less. Total Dragon + trunk + payload + fuel would then be closer to 16,100 kg than my initial 17,870 kg. The difference would likely come either from a lighter Dragon or from a lighter payload (since my fuel mass is in the reported range, and since the trunk is already a very light 400 kg). $\endgroup$
    – user39728
    Apr 29, 2021 at 18:56

2 Answers 2


If your launch profile doesn't require you to hit 3G, you can increase the payload. Such a profile would "waste" Delta V, but would remain within the escape envelope.

Further, if you trigger the escape does it not also cut the main Falcon engines? So it should rapidly stop accelerating.

  • 2
    $\begingroup$ If the rocket is malfunctioning, you have no guarantee the engine controllers will be working right. They could well be failing too. So you have to assume the very worst possible case, because it could happen: that you're accelerating at a peak 3g, and that you cannot shut down the rocket engines before you fire the launch escape thrusters. $\endgroup$
    – user39728
    Apr 29, 2021 at 17:51
  • $\begingroup$ If you don't hit 3g, then you don't hit the upper bound imposed by launch escape (if in fact there is one). Doesn't mean the upper bound isn't there---it's just you never get close enough to it for it to matter. Just like a ceiling---it limits how high you can jump, and just because you can't jump high enough to hit it, doesn't mean it isn't there ;-) $\endgroup$
    – user39728
    Apr 29, 2021 at 17:54
  • 1
    $\begingroup$ @user39728 Interestingly, the flight termination system (FTS), slide 13 reliability is greater than the Commercial Crew requirement for abort system reliability Slide 29 "Commercial Crew Program: Key Driving Requirements Walkthrough". NASA. Archived from the original on March 28, 2012. so potentially the likelihood of the abort system failing is greater than the likelihood of the FTS failing $\endgroup$ Apr 29, 2021 at 19:05
  • 1
    $\begingroup$ @user39728 Another consequence of the FTS being more reliable than the (minimum required) abort system is that you potentially don't even have to consider (at least contractually) the possibility of aborting while still firing the first stage engines as it is outside of the odds you are required to achieve. This is (?perhaps) evidenced by the in-flight abort where they shutdown the engines while the abort was occurring. $\endgroup$ Apr 29, 2021 at 19:11
  • $\begingroup$ Interesting! This is great to know. So OK, it seems you could get away with accelerations smaller than 3.5 g. If you assume you can always shut down the launcher engines (if they didn't completely fail), then even a small acceleration would be enough. Although it still would be smart to ensure you have enough thrust acceleration for that super unlikely but still possible worst case where you need >3g to speed away from the rocket. So if SpaceX wanted to fill the Dragon trunk to the brim when the Falcon launcher allows it, they can... and they have to decide if they should : D $\endgroup$
    – user39728
    Apr 29, 2021 at 19:21

Some of the mass value assumptions in your question may be incorrect. For example, the value you found for the payload (max of 5,900 kg) does not agree with the information published in a report from the Office of the Inspector General which pegs it at 3307 kg. The dry mass value that I have in my notes for Crew Dragon is 7711 kg, including the trunk. See this FAA report

Dragon weighs approximately 17,000 pounds without cargo and is approximately 17 feet tall with a base width of 13 feet. Dragon-2 is composed of the capsule for pressurized crew and cargo, the unpressurized cargo module or “trunk,” and a nosecone.

Your question assumes a larger value of 9,570+400=9970 kg for Dragon and its trunk.

Later in the same report...

Dragon would contain approximately 5,650 pounds of hypergolic propellant, including approximately 3,500 pounds of dinitrogen tetroxide (NTO) and 2,150 pounds of monomethylhydrazine (MMH). Dragon would contain approximately 2,400 pounds of residual propellant after the abort test.

So after conversion to metric, the propellant numbers are 2,563 kg (start) and 1,089 kg (end). You assumed 2,000 kg.

If you use the numbers from the reports I cited, the total mass at the start of abort would be

$$m_{total} = 7700 + 3307 + 2563 = 13570 kg$$

and at the end it would be

$$m_{total} = 7700 + 3307 + 1089 = 12096 kg$$

If your thrust numbers are correct, then the initial acceleration would be

$$Acceleration = 550kN/13570kg = 40.53 m/s^2 = 4.14 g's$$

and the final acceleration would be

$$Acceleration = 550kN/12096kg = 45.47 m/s^2 = 4.64 g's$$

This should be more than enough to accelerate away from the rocket during an abort, and therefore the abort acceleration requirements probably are not directly limiting the rated payload to the ISS.


Your Answer

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