Case Study: Reconstructing the Merlin Engine and the Falcon 9 First Stage

"The best way to understand a machine is to try to rebuild it from its own numbers."

Executive Summary

A manufacturer publishes only a handful of numbers about a rocket engine — a thrust, a specific impulse, maybe a propellant load and a burn time. In this case study we show that those few numbers, run through this chapter's equations, are enough to reconstruct almost everything else: the effective exhaust velocity, the mass flow rate, the nozzle size, the thrust-to-weight at liftoff and at burnout, the total impulse of the stage, and the raw jet power pouring out of the engines. Our subject is the SpaceX Merlin 1D and the Falcon 9 first stage it powers nine at a time — the most-flown orbital hardware of our era. Every step is a chance to sanity-check one number against another, and the satisfying result is that they all lock together into a single consistent story, with only the honest two-percent seams you always find in real catalog data.

Skills applied

• Converting specific impulse to effective exhaust velocity and inverting $F = \dot m\, c$ for mass flow (§16.1, §16.2).
• Cross-checking a derived mass flow against an independent propellant-and-burn-time figure (§16.2, §16.6).
• Computing thrust-to-weight at liftoff and at burnout, and explaining throttling (§16.4).
• Computing a stage's total impulse and jet power, and checking them against thrust (§16.5, §16.6).

Background

The vehicle and its engine

Falcon 9's first stage carries nine Merlin 1D engines burning kerolox (RP-1 kerosene and liquid oxygen). The public, approximate numbers we will work from (Appendix H; Tier 2, version-dependent):

Everything else in this study is derived from these with the equations of Chapter 16. Treat every result as good to about two significant figures — the precision of the inputs.

Phase 1: From catalog numbers to exhaust velocity and mass flow

Specific impulse is a velocity in disguise; unmask it with $c = I_{sp}\, g_0$:

$$ c_{\text{SL}} = 282 \times 9.80665 = 2{,}765\ \text{m/s}, \qquad c_{\text{vac}} = 311 \times 9.80665 = 3{,}050\ \text{m/s}. $$

So each Merlin throws its exhaust at an effective $2.77\ \text{km/s}$ at the pad, rising to $3.05\ \text{km/s}$ in vacuum as the pressure term grows. Now invert the thrust relation $F = \dot m\, c$ to get the propellant flow through one engine (using the sea-level pairing, both measured at the pad):

$$ \dot m = \frac{F_{\text{SL}}}{c_{\text{SL}}} = \frac{845{,}000\ \text{N}}{2{,}765\ \text{m/s}} \approx 306\ \text{kg/s}. $$

A single Merlin gulps roughly 306 kilograms of propellant every second at full thrust — about a bathtub of kerosene and oxygen per second, per engine.

Phase 2: Does the mass flow survive an independent check?

We derived $\dot m$ from thrust and $I_{sp}$. Let us test it against completely different numbers — the propellant load and burn time — which we did not use above. Nine engines burning $395\ \text{t}$ in $162\ \text{s}$ implies an average total flow of

$$ \dot m_{\text{avg, total}} = \frac{395{,}000\ \text{kg}}{162\ \text{s}} \approx 2{,}438\ \text{kg/s}, \qquad \text{per engine } \frac{2{,}438}{9} \approx 271\ \text{kg/s}. $$

Our full-thrust figure was $306\ \text{kg/s}$; the flight average is $271\ \text{kg/s}$. These agree at the right level and, crucially, on the right side: the average is below full-thrust because Falcon 9 throttles its engines down through the region of maximum aerodynamic pressure and again near the end of the burn (both to limit loads and to reserve propellant for the booster's return). Two independent roads — "thrust and $I_{sp}$" versus "propellant and burn time" — arrive within about $13\%$ of each other, exactly the gap that throttling should produce. The reconstruction is trustworthy.

🔧 Engineering Reality: The two-percent (and here, thirteen-percent) seams are not errors — they are physics the single catalog number hides. A quoted thrust is a maximum; a quoted burn time reflects a real, throttled flight profile. An engineer who demands that every published number be perfectly mutually consistent will drive themselves mad; one who understands why they differ can read a whole flight profile out of the discrepancy.

Phase 3: Thrust-to-weight, at liftoff and at burnout

At liftoff, nine sea-level Merlins push with $F = 9 \times 845 = 7{,}605\ \text{kN}$ against a weight of

$$ W_0 = 549{,}000 \times 9.81 = 5{,}386\ \text{kN}, \qquad \frac{T}{W}\bigg|_{\text{liftoff}} = \frac{7{,}605}{5{,}386} \approx 1.41. $$

The vehicle leaves the pad at $a = g(T/W - 1) = 9.81 \times 0.41 \approx 4.0\ \text{m/s}^2$, the gentle $0.4\,g$ crawl you see in every launch video. Now jump to first-stage burnout. The stage has shed its $395\ \text{t}$ of propellant, so the remaining stacked mass is about $549 - 395 = 154\ \text{t}$, and the engines — now in near-vacuum — push harder, about $9 \times 914 = 8{,}226\ \text{kN}$:

$$ \frac{T}{W}\bigg|_{\text{burnout}} = \frac{8{,}226}{154{,}000 \times 9.81 / 1000} = \frac{8{,}226}{1{,}511} \approx 5.4. $$

Untamed, that would fling the vehicle forward at $a = 9.81 \times 4.4 \approx 43\ \text{m/s}^2 \approx 4.4\,g$ — punishing for structure and payload alike. This is why Falcon 9 throttles down before staging. The same lesson we drew abstractly in §16.4 — that $T/W$ rises through a burn as mass falls — is written into every Falcon 9 flight as a deliberate throttle-back.

🔗 Connection: That throttle margin does double duty. The same ability to throttle deep and hold propellant in reserve is what lets a Falcon 9 booster fly back — several engines for the boostback and entry burns, one throttled engine for the landing burn — and touch down intact for reuse. The propulsive landing and its economics are the subject of Chapter 22; the throttle range we are reading out of these numbers is where that capability is quietly bought.

Phase 4: Total impulse and jet power of the stage

The first stage's total impulse — its whole momentum punch — follows from $I_t = c\, m_p$ using a sea-level-to-vacuum average $c \approx 2{,}900\ \text{m/s}$ ($I_{sp} \approx 296\ \text{s}$):

$$ I_t = 2{,}900\ \text{m/s} \times 395{,}000\ \text{kg} \approx 1.15 \times 10^{9}\ \text{N·s} = 1.15\ \text{GN·s}. $$

Cross-check against thrust and time: $F_{\text{avg}} = I_t / t_b = 1.15\times10^{9} / 162 \approx 7.1\ \text{MN}$, just under the $7.6\ \text{MN}$ sea-level liftoff thrust — right where a throttled climb into vacuum should land the average. Finally, the jet power streaming out of the stage at liftoff:

$$ P = \tfrac{1}{2}\,\dot m_{\text{total}}\, c^2 = \tfrac12 (9 \times 306)(2{,}765)^2 \approx \tfrac12 (2{,}754)(7.65\times10^{6}) \approx 1.05\times10^{10}\ \text{W} \approx 10.5\ \text{GW}. $$

Ten gigawatts — roughly ten large ($1\ \text{GW}$) power stations' worth of kinetic energy, hurled into the exhaust every second by a stage the size of a grain silo. And yet, per §16.5, all that power buys only a "modest" $2.8\ \text{km/s}$ of exhaust velocity, because chemistry trades efficiency for the sheer thrust that lifts $549\ \text{t}$ off the Earth.

💡 Intuition: Ten gigawatts sounds impossible for a machine you could park in a warehouse — until you remember where the energy comes from. It is not generated on board, as a power station generates its gigawatt; it is released from chemical bonds already stored in the propellant, at whatever rate the turbopumps can force propellant into the chamber. The engine is not a generator; it is a valve on a bonfire. That is exactly why chemical rockets are power-rich and efficiency-poor (§16.5) — and why, the instant the tanks run dry, all ten gigawatts vanish.

Phase 5: The honest seams

One reconciliation remains. In Phase 1 we found $\dot m \approx 306\ \text{kg/s}$ from the sea-level figures. Carried to vacuum, that flow predicts $F_{\text{vac}} = \dot m\, c_{\text{vac}} = 306 \times 3{,}050 \approx 932\ \text{kN}$ — yet the widely reported vacuum thrust is $\sim 914\ \text{kN}$, about $2\%$ lower. Which is right? Both, to their precision. The public ${F_{\text{SL}}, F_{\text{vac}}, I_{sp,\text{SL}}, I_{sp,\text{vac}}}$ come from different tests and rounding, and they simply do not close to better than a couple of percent. A working engineer notes the seam, quotes results to two significant figures, and — when a mission actually depends on the third digit — goes to the primary test data rather than the catalog. The reconstruction is not less valuable for its seams; it is more honest.

There is a practical upshot for your own mission work. When you pull an engine's numbers from Appendix H or a manufacturer's press kit to size a stage, carry two significant figures and expect a few percent of slop between any two quantities you derive from them. If a design closes only when a third digit falls a particular way, it has not really closed — the margin is hiding inside the noise of the inputs, and you need either better data or a more forgiving design. Reconstruction is not a party trick; it is how you learn how much to trust each number before you bet a mission on it.

🔗 Connection: Notice that we never opened the engine. We deduced its nozzle size (§16.1: $A_e \approx 0.68\ \text{m}^2$), its propellant flow, its power, and its flight throttle profile from the outside, using only conservation of momentum and the definitions of this chapter. Chapter 17 finally opens the box — the injector, chamber, turbopump, and gas-generator cycle that produce the $306\ \text{kg/s}$ and $2.77\ \text{km/s}$ we treated here as given.

Discussion Questions

1. Our full-thrust mass flow ($306\ \text{kg/s}$) exceeded the flight-average flow ($271\ \text{kg/s}$). Explain physically why, and what a smaller gap would tell you about a flight's throttle profile.
2. Thrust-to-weight rose from $1.4$ at liftoff to a would-be $5.4$ at burnout. Trace exactly which quantities changed, and by how much, to produce that rise.
3. The stage's jet power (~$10.5\ \text{GW}$) dwarfs its "modest" exhaust velocity's reputation. Using $F = 2P/c$, explain how enormous power and modest exhaust velocity coexist in a chemical engine.
4. If SpaceX published only Merlin's vacuum thrust and vacuum $I_{sp}$, could you still reconstruct the sea-level thrust? What additional number would you need, and which equation would you use?

Your Turn: Extensions

• Option A (analysis). Redo Phase 3's burnout $T/W$ assuming the stage does not throttle. At what vehicle mass does the acceleration reach $4\,g$? (Set $g(T/W - 1) = 4g$ and solve for mass, using vacuum thrust.) How much propellant remains at that point?
• Option B (computation). Write a Python function reconstruct(F_sl, isp_sl, isp_vac) that returns $c_{\text{SL}}$, $c_{\text{vac}}$, $\dot m$, and predicted $F_{\text{vac}}$. Reproduce this study's numbers. (Do not run it — hand-trace and add # Expected output:.)
• Option C (design lens). Falcon Heavy straps three Falcon-9-class cores together (27 Merlins) with a liftoff mass near $1{,}420\ \text{t}$. Estimate its liftoff $T/W$ and initial acceleration. Is it higher or lower than Falcon 9's, and why?

Key Takeaways

1. A few catalog numbers determine the rest. From thrust and $I_{sp}$ alone we recovered exhaust velocity, mass flow, nozzle area, total impulse, and jet power — using only Chapter 16's equations.
2. Cross-checking is the discipline. A mass flow derived one way ($F/c$) must agree with one derived another way (propellant$/$burn time); the residual gap is the throttle profile.
3. Thrust-to-weight is a moving target within a flight, rising as propellant burns off — which is exactly why Falcon 9 throttles down before staging.
4. Catalog numbers carry honest two-percent seams. Quote to two significant figures, understand why sources differ (version, power level, sea level vs. vacuum), and go to primary data when a mission depends on the third digit.