Case Study: Reconstructing Perseverance's Seven Minutes of Terror

"Touchdown confirmed. Perseverance is safely on the surface of Mars." — NASA/JPL EDL commentary, 18 February 2021

Executive Summary

On 18 February 2021, NASA's Perseverance rover — a one-tonne, car-sized machine — went from the top of the Martian atmosphere to a wheels-down landing in Jezero Crater in about seven minutes, entirely on its own, while the engineers who built it sat in a control room watching a recording of events that had already finished. In this case study we take the EDL toolbox of §34.4 and audit the real landing: we compute the entry speed from the arrival geometry, estimate the deceleration the heat shield delivered, show with the chapter's terminal-velocity formula why the supersonic parachute could not finish the job, and account for the sky crane and the light-time delay that made the "terror" real. Every number will confirm the chapter's central claim — Mars is too thick to ignore and too thin to trust — and reveal why Perseverance sits right at the mass ceiling of parachute-based landing.

Skills applied

• Computing Mars entry speed from the arrival hyperbolic excess (§34.4, Ch. 11).
• Order-of-magnitude estimation of entry deceleration and energy dissipation (§34.4, Ch. 7).
• Applying the terminal-velocity formula to a real supersonic parachute (§34.4).
• Reasoning about the EDL chain — entry, parachute, powered descent, sky crane (§34.4).
• Accounting for communication light-time versus EDL duration, and hence autonomy (§34.3, Ch. 27).

Background

Perseverance launched on 30 July 2020 in that year's Mars window and cruised for about seven months on the Chapter-11 transfer. It arrived as the heaviest thing yet sent to the Martian surface, carried inside an aeroshell $4.5\ \text{m}$ across. Its landing reused — and pushed to the limit — the "sky crane" architecture first flown by Curiosity in 2012. The rough, widely reported EDL numbers (Tier 2):

Our job is not to admire these numbers but to reproduce them from the physics of the chapter, and see what they teach.

Phase 1: Arrival geometry — how fast does it hit the atmosphere?

Perseverance flew a direct-entry profile: no braking into orbit first, straight from the arrival hyperbola into the atmosphere. From Chapter 11, a minimum-energy Hohmann arrival has a hyperbolic excess relative to Mars of $v_{\infty,\text{Mars}} = 2.65\ \text{km/s}$. Falling into Mars's gravity well from the SOI edge to the atmospheric interface at $r_{\text{atm}} \approx 3{,}490\ \text{km}$ (about $100\ \text{km}$ altitude), energy conservation gives the entry speed:

$$ v_{\text{entry}} = \sqrt{v_{\infty,\text{Mars}}^2 + \frac{2\mu_{\text{Mars}}}{r_{\text{atm}}}} = \sqrt{(2.65)^2 + \frac{2(4.283\times10^4)}{3{,}490}} = \sqrt{7.0 + 24.5} = \sqrt{31.5} = 5.6\ \text{km/s}. $$

Our patched-conic estimate, $5.6\ \text{km/s}$, is within a few percent of the reported $\sim 5.4\ \text{km/s}$ — the small difference is the real 2020 arrival geometry, which was not exactly the idealized minimum-energy Hohmann (Tier 2/3). Sanity check: the entry speed is far below the spacecraft's $\sim 21.5\ \text{km/s}$ heliocentric speed at Mars's orbit — because, as Chapter 11 hammered, entry speed is measured relative to Mars, and Mars and the spacecraft were moving nearly together. Confusing the two is the classic error, and it would have you build a heat shield for four times the real energy.

Phase 2: The heat shield does most of the work

At $5.4\ \text{km/s}$ into even Mars's thin air, the vehicle drives a hypersonic shock and the heat shield must dissipate the bulk of the kinetic energy as heat — the physics of Chapter 7. The kinetic energy per kilogram at entry is

$$ \frac{1}{2}v_{\text{entry}}^2 = \frac{1}{2}(5{,}400\ \text{m/s})^2 = 1.46\times10^7\ \text{J/kg}, $$

so the $\sim 3{,}650\ \text{kg}$ vehicle arrives with about $5.3\times10^{10}\ \text{J}$ — roughly the energy of thirteen tonnes of TNT, nearly all of which the heat shield must turn into heat and a glowing wake in a couple of minutes.

We can bound the deceleration. The vehicle sheds most of its $\sim 5.4\ \text{km/s}$ during the minute or so of peak atmospheric braking. A rough average is

$$ \bar a \sim \frac{\Delta v}{\Delta t} \approx \frac{5{,}000\ \text{m/s}}{60\ \text{s}} \approx 83\ \text{m/s}^2 \approx 8.5\ g, $$

with the peak running higher, around $10$–$12\ g$ — exactly the reported figure. Sanity check: a $10\ g$ peak is survivable for hardware and even for a (braced) human, which is why aerodynamic entry, not propulsion, is used to remove the first and largest chunk of speed — it is free, paid for by the heat shield's mass rather than by propellant. The heat shield takes the vehicle from Mach 25 or so down to about Mach 2. It does the lion's share of the job.

Phase 3: The supersonic parachute — necessary but not sufficient

At about Mach $1.7$ and $11\ \text{km}$ altitude, Perseverance deployed a $21.5\ \text{m}$ disk-gap-band parachute into flow still moving faster than sound. Here is the crux of Mars EDL: even a chute this enormous cannot land the vehicle. Apply the chapter's terminal-velocity formula to the vehicle hanging under the chute (mass $\sim 2{,}270\ \text{kg}$ after heat-shield jettison; chute area $A = \pi(10.75)^2 \approx 363\ \text{m}^2$; $C_d \approx 0.5$; local density at $\sim 10\ \text{km}$, $\rho \approx 0.008\ \text{kg/m}^3$, Tier 3):

import math

def terminal_velocity(m, g, rho, Cd, A):
    """Steady-fall speed: v = sqrt(2 m g / (rho Cd A))."""
    return math.sqrt(2 * m * g / (rho * Cd * A))

A_chute = math.pi * (21.5 / 2) ** 2                 # ~363 m^2
v = terminal_velocity(2270, 3.71, 0.008, 0.5, A_chute)
print(f"terminal speed under the parachute: {v:.0f} m/s")
# Expected output:
# terminal speed under the parachute: 108 m/s

About $108\ \text{m/s}$ — roughly $390\ \text{km/h}$ — even under a $21.5$-metre canopy. Sanity check and meaning: that is highway-and-then-some speed, utterly unsurvivable as a landing, and it is the direct, quantitative reason Perseverance could not simply float down. The biggest parachute ever flown to Mars only slows the vehicle to a couple hundred kilometers per hour. The parachute is a step — it removes the transonic speed the heat shield left behind — not a solution. Something powered must take it the rest of the way.

Phase 4: Powered descent and the sky crane

Below the parachute's usefulness, Perseverance cut away the backshell and chute and fired up the eight Mars Landing Engines of its descent stage — supersonic-to-subsonic retropropulsion, the powered descent of §34.4. Using Terrain-Relative Navigation, it steered to a safe patch of Jezero Crater, slowing to a hover at about $20\ \text{m}$ altitude. Then came the maneuver that looks, on video, like science fiction: rather than land the rocket stage (whose plume would blast a crater and whose legs would have to be huge), the descent stage lowered the rover on $\sim 7.6\ \text{m}$ nylon cables — the sky crane — set it on its wheels at about $0.75\ \text{m/s}$, cut the cables, and flew away to crash at a safe distance.

Compare the touchdown speeds across the phases to see the whole relay:

Each method hands off to the next because none can do the whole job: the heat shield cannot land you, the parachute cannot land you, and the engines cannot fire through a heat shield at Mach 25. Mars EDL is a chain precisely because Mars's atmosphere gives you part — but only part — of the braking you need.

🔧 Engineering Reality: Perseverance sits at the mass wall. At about a tonne of rover, Perseverance is near the ceiling of what the heat-shield-parachute-sky-crane chain can deliver, because the $21.5\ \text{m}$ parachute is close to the largest that can be reliably deployed into Mars's supersonic flow. To land the tens of tonnes a crew needs (§34.4), there is no bigger chute — the physics does not scale — which is why human-scale landers must abandon parachutes entirely and ride engines down from much higher speed, the propulsive-descent bet of Starship. Perseverance is not a stepping stone to a crewed lander; it is the end of one technological road and the argument for taking a different one.

Phase 5: The terror was real — the accounting

Why "terror," and not merely "difficulty"? Because no one on Earth could do anything, or even know anything, in time. On 18 February 2021 Mars was about $2.05\times10^8\ \text{km}$ away, so the one-way light time was

$$ t_{\text{light}} = \frac{2.05\times10^8\ \text{km}}{2.998\times10^5\ \text{km/s}} = 684\ \text{s} \approx 11.4\ \text{minutes}. $$

The entire EDL sequence took about $7\ \text{minutes}$ ($\sim 420\ \text{s}$). So:

$$ t_{\text{light}} - t_{\text{EDL}} \approx 684 - 420 \approx 264\ \text{s} \approx 4.4\ \text{minutes}. $$

By the time the radio signal saying "we have begun entry" reached Earth, Perseverance had already been on the surface — alive or dead — for more than four minutes. Every one of the hundreds of EDL events fired on commands the rover gave itself, in real time, with no human in the loop and no possibility of one — the autonomy of Chapter 27 taken to its absolute conclusion. The control room's job during the seven minutes was not to fly the spacecraft. It was to wait, and to trust the machine and the physics. That is the terror: not danger you can fight, but danger already resolved, its outcome racing toward you at the speed of light.

Discussion Questions

1. Our patched-conic entry speed ($5.6\ \text{km/s}$) was a few percent above the reported value ($\sim 5.4\ \text{km/s}$). List two real effects that a patched-conic estimate omits, and say which way each would nudge the number.
2. The heat shield removed the vehicle's speed from $\sim 5{,}400$ to $\sim 420\ \text{m/s}$, but the engines removed only the last $\sim 160\ \text{m/s}$. Why is it right to spend "free" aerodynamic braking on the big chunk and expensive propellant only on the small one?
3. A $21.5\ \text{m}$ parachute still leaves the vehicle at $\sim 100\ \text{m/s}$. Explain, using the terminal-velocity formula, why building a parachute twice as wide would help far less than you might hope — and why it cannot be scaled to a crewed lander.
4. The sky crane exists to avoid landing the rocket stage itself. Give two engineering reasons landing a large descent engine directly on the surface is undesirable for a delicate rover.

Your Turn: Extensions

• Option A (analysis). Redo Phase 3 for Curiosity (2012), essentially the same architecture: a similar chute and $\sim 2{,}400\ \text{kg}$ under it. Recompute the terminal speed and confirm it is also "too fast to land." What does the near-identical result say about why the sky crane was reused unchanged?
• Option B (computation). Write light_time(distance_km) returning the one-way light delay in minutes, and edl_gap(distance_km, edl_seconds) returning how long after the rover is already down the "entry started" signal arrives. Hand-trace both for the landing-day distance and a $420\ \text{s}$ EDL. (Do not run it — add # Expected output:.)
• Option C (design). Sketch, in words and rough numbers, how a crewed lander's EDL must differ from Perseverance's. Which phases survive (entry, heat shield), which are discarded (parachute, sky crane), and what replaces them? Tie your answer to the mass wall of §34.4.

Key Takeaways

1. Entry speed is relative to Mars, and modest. From the arrival hyperbola, Perseverance hit the atmosphere at $\sim 5.6\ \text{km/s}$ — not its $21.5\ \text{km/s}$ heliocentric speed. Always ask "relative to which body?"
2. The heat shield does the heavy lifting, for free. It removes most of the $\sim 5.4\ \text{km/s}$ as heat at $10$–$12\ g$, leaving only the transonic remainder — because aerodynamic braking is paid for in shield mass, not propellant.
3. Even the biggest parachute cannot land you. A $21.5\ \text{m}$ chute leaves $\sim 100\ \text{m/s}$; powered descent and the sky crane finish the job. This is the quantitative face of the Mars mass wall.
4. The terror is the light-time. EDL (~7 min) finishes before the "we've begun" signal even reaches Earth (~11 min away), so the rover must fly the whole sequence autonomously — danger already resolved by the time anyone can watch.