Case Study: Auditing the All-Electric Communications Satellite

"The whole reason two of them fit on one rocket is a decision made in a trade study."

Executive Summary

In March 2015 a single Falcon 9 launched two commercial communications satellites at once — ABS-3A and Eutelsat 115 West B, both built on Boeing's 702SP bus. They were the first all-electric commercial comsats: they carried no chemical apogee engine at all, raising themselves from their drop-off orbit to geostationary orbit entirely on xenon-fueled electric thrusters over several months. This case study audits that mission-design decision with the tools of Chapter 29. We will reconstruct the trade study that chose electric over chemical, quantify the mass saving that let two satellites share one launch, and identify the driving requirement — launch cost per satellite — that made the whole architecture close. We are not designing anything here; we are checking someone else's design and asking why it is the shape it is.

Skills applied

• Reconstructing a delta-v budget and rolling it up (§29.3).
• Running a chemical-vs-electric trade study and sizing each option with the rocket equation (§29.2).
• Tracing the mass–power–cost coupling from a subsystem choice to launch cost (§29.5).
• Identifying the driving requirement behind a real architecture (§29.2, §29.5).

Background

A geostationary communications satellite is dropped by its launch vehicle into a transfer orbit (GTO) and must climb the rest of the way to GEO on its own, then hold station for its service life — the delta-v budget of §29.3, roughly 2.6 km/s ideal for a satellite launched from a mid-latitude site that folds its plane change into the climb. For fifty years the standard answer was a chemical apogee engine: high thrust, orbit-raising in a few days, but — because chemical exhaust is slow — a lot of propellant.

By 2015, electric propulsion (Chapter 20) had matured enough that operators dared to remove the chemical engine entirely. The bet: accept months of slow spiraling to GEO in exchange for carrying a small fraction of the propellant mass — and use the mass saved to launch two satellites where one used to go. Let us audit whether the numbers justify the bet.

Phase 1: Reconstruct the delta-v budget

The satellite's own propulsion (not the launch vehicle) must supply, approximately (Tier 2 — real values vary by orbit and operator):

Rolling this up (the roll_up_dv of §29.3): ideal $= 1{,}800 + 750 + 11 = 2{,}561\ \text{m/s}$; margined $= 1{,}800(1.05) + 750(1.10) + 11 = 2{,}726\ \text{m/s}$. This is the same budget for both architectures — the delta-v a satellite must supply does not care how it supplies it. What changes is the propellant mass, and that is set by the engine.

Phase 2: Size both architectures

Take a dry mass of 1,500 kg (bus plus payload, excluding propellant) and size the propulsion with size_vehicle (§29.5) for each engine choice:

import math
G0 = 9.80665
def size_vehicle(dv, isp, payload):
    ve = isp * G0
    m_wet = math.exp(dv / ve) * payload
    return m_wet - payload, m_wet

for name, isp in [("chemical", 320), ("electric", 1800)]:
    mp, mw = size_vehicle(2726, isp, 1500)
    print(f"{name:9s} Isp={isp:4d}s -> propellant {mp:6.0f} kg, wet {mw:6.0f} kg")
# Expected output:
# chemical  Isp= 320s -> propellant   2076 kg, wet   3576 kg
# electric  Isp=1800s -> propellant    250 kg, wet   1750 kg

The chemical satellite masses about 3,576 kg fueled; the electric one about 1,750 kg. Same mission, same delta-v, same dry mass — and the electric version is less than half the launch mass, because its six-times-faster exhaust drops the required propellant from over two tonnes of bipropellant to about 250 kg of xenon. That factor-of-two in wet mass is the entire story.

🔗 Connection: This is the mass ratio of §29.5 doing its work in reverse. The chemical satellite's mass ratio is $e^{2726/3138} = 2.38$; the electric one's is $e^{2726/17652} = 1.17$. The dry mass is the same 1,500 kg in both, but the chemical bird multiplies it by 2.38 and the electric bird by only 1.17. The rocket equation's exponential, so often the villain, is here the hero: raising $v_e$ collapses the mass ratio toward 1.

Phase 3: Trace the coupling to launch cost

Now follow the mass–power–cost loop (§29.5) to the bottom. A Falcon 9 can deliver roughly 5,000 kg to GTO (Tier 2). At 3,576 kg each, two chemical satellites (7,152 kg) do not fit — each needs its own launch. At 1,750 kg each, two electric satellites (3,500 kg) fit comfortably under one fairing, stacked. So:

Halving the launch cost per satellite is an enormous competitive advantage in a commercial market, and it comes entirely from a propulsion trade study — a single box in the funnel of §29.1. That is systems engineering: a subsystem choice, propagated through the coupling, that reshapes the mission's economics.

Phase 4: What the operator paid for it

The trade is not free, and a fair audit names the cost. Electric orbit-raising takes months, not days — roughly 6 to 8 months of slow spiraling (Tier 2). During that time the satellite earns no revenue and spends longer crossing the Van Allen radiation belts (Chapter 20 discusses the radiation dose electric transfers accumulate). For a satellite meant to earn for fifteen years, trading six months of early revenue for half the launch cost is usually a clear win — but for an operator who needed capacity now (a replacement for a failed satellite, say), the chemical architecture's speed to station could be worth its extra mass. The trade depends on the weighting, exactly as §29.2 warned.

🔧 Engineering Reality: All-electric is not the only answer the trade produced. Some operators chose a hybrid: a small chemical engine to do the fastest, most radiation-heavy part of the climb quickly, then electric thrusters for the rest and for station-keeping. Hybrid lands between the two rows above — more mass than all-electric, faster to station. That there are three viable answers (chemical, electric, hybrid), each best under a different weighting, is precisely why the decision is made in a documented trade study rather than by rule of thumb.

Phase 5: Identify the driving requirement

Which requirement drove this architecture? Trace the sensitivities. The satellites were designed to a tight launch-mass target because launch cost per satellite was the metric the business lived or died by — and launch mass is set, through the coupling, by the propellant mass, which is set by the choice of propulsion for the delta-v budget. The driving requirement was therefore effectively "minimize launch cost per unit of payload capacity", and it drove the design straight to the lowest-wet-mass propulsion option the schedule could tolerate. Everything else — the array size, the thermal design, even the choice to stack two satellites — followed from serving that one number.

⚠️ Common Misconception: "Electric propulsion is just better, so of course they used it." Electric is not universally better — it is better for this weighting of criteria. The same electric thruster that wins for a 15-year commercial comsat would lose badly for a mission that must reach its orbit in days, or one whose driving requirement is time-to-station rather than launch cost. The audit's lesson is not "use electric"; it is "the right answer is the output of the trade study, and the trade study is driven by the driving requirement."

Discussion Questions

1. The delta-v budget was identical for both architectures, yet the wet masses differed by a factor of two. Explain, in terms of the rocket equation, how the same delta-v produces such different masses.
2. Suppose xenon prices tripled. Would that change the trade study's outcome? Estimate the propellant cost as a fraction of a satellite's total cost before answering.
3. The audit called launch cost per satellite the driving requirement. Propose a different mission (still a comsat) whose driving requirement would be something else, and say what architecture it would favor.
4. Hybrid propulsion was mentioned as a third option. Sketch the row it would occupy in the Phase 3 table and argue when a customer would choose it.

Your Turn: Extensions

• Option A (analysis). Re-run Phase 2 for a satellite with a heavier payload — dry mass 2,000 kg instead of 1,500 kg. Do both wet masses scale by the same factor? Does the "two per Falcon 9" conclusion survive for the electric version? (Hand-trace size_vehicle; add # Expected output:.)
• Option B (computation). Extend the Phase 2 script to also print each architecture's mass ratio and the number of satellites that fit under a 5,000 kg GTO capacity. Reproduce the Phase 3 table.
• Option C (your mission). If your mission is Track A, decide chemical vs. electric vs. hybrid for your comsat and record the sized wet mass in your MDR. If it is another track, identify the analogous propulsion trade and its driving requirement.

Key Takeaways

1. The same delta-v budget sizes very different vehicles, because the engine's $I_{sp}$ sets the mass ratio: raising $v_e$ six-fold collapsed the mass ratio from 2.38 to 1.17 and halved the wet mass.
2. A subsystem trade can reshape the mission's economics. Choosing electric propulsion — one box in the funnel — is what let two satellites share one launch and halved the cost per satellite.
3. Every trade has a cost; a fair audit names it. Electric bought mass with time (months of spiraling and extra radiation), a price worth paying only under the operator's weighting.
4. Architectures serve the driving requirement. Launch cost per satellite drove this design; a different driving requirement (speed to station) would have justified a different, heavier answer.