Study Guide — Highly Elliptical Orbit (HEO) Satellite Systems

Unit Overview

Every orbit discussed so far is circular — but the most elegant satellite system ever flown is not. The Highly Elliptical Orbit places its apogee over the region of interest and exploits Kepler's second law: the satellite dawdles near apogee for most of its period, then whips through perigee over the hemisphere nobody cares about. Paired with the critical inclination of 63.4° — the one angle at which Earth's equatorial bulge cannot make the orbit precess — HEO delivers continuous, high-elevation coverage of the Arctic and high northern latitudes that geostationary satellites physically cannot reach. Born as the Soviet Molniya communications system and matured by Sirius XM radio, HEO remains the specialist's answer to the question GEO fails: who watches the poles?

Learning Outcomes

After studying this unit and completing the tutorials you should be able to:

Apply Kepler's second law (via the eccentric-anomaly equation) to quantify apogee dwell.
Design Molniya and Tundra orbits: semi-major axis, e, i, argument of perigee.
Explain the critical inclination 63.4° and the J₂ precession it suppresses.
Phase a 3-satellite Molniya constellation for continuous high-latitude coverage.
Build HEO link budgets including the apogee/perigee variation and pointing dynamics.
Analyse HEO spacecraft design constraints: radiation belts, thermal, power, tracking.
Compare HEO against GEO/MEO/LEO for high-latitude missions with numbers.
Explain modern HEO uses: early warning, PNT augmentation, space domain awareness.

Indicative Study Time

Activity	Time
Guided reading (Sections 1–8)	5 hours
Figures studied / examples reworked	2 hours
Tutorial problems (Section 9)	3 hours
Revision of key equations & glossary	1 hour

Prerequisite knowledge: Kepler's laws; the eccentric-anomaly time equation; orbital elements (a, e, i, Ω, ω); coverage/elevation geometry; link budgets. The companion guide "LEO, MEO, GEO and HEO Satellite Systems" provides the regime overview this unit deepens.

Section 1 — HEO Fundamentals

1.1 The Orbit and Its Numbers

Vis-viva: v = √[μ(2/r − 1/a)] → v_p/v_a = (1+e)/(1−e)

Parameter	Molniya	Tundra
Semi-major axis	26 560 km (T = 11 h 58 min = ½ sidereal day)	42 164 km (T = 24 h = 1 sidereal day)
Eccentricity	e = 0.7	e ≈ 0.27
Inclination	63.4° (critical)	63.4° (critical)
Perigee / apogee radius	7 968 / 45 152 km	30 782 / 53 546 km
Perigee / apogee altitude	1 597 / 38 781 km	24 411 / 47 175 km
Perigee / apogee velocity	9.2 / 1.6 km/s	4.5 / 2.6 km/s
Satellites for continuity	3 per plane (8 h apart)	2 per plane (12 h apart)

Example 1.1 — Molniya sizing. Require T = ½ sidereal day = 43 082 s: a = (μT²/4π²)^1/3 = (3.986×10⁵ × 43 082²/39.478)^1/3 ≈ 26 560 km. With e = 0.7: r_p = a(1−e) = 7 968 km (altitude 1 597 km — safely above the atmosphere), r_a = a(1+e) = 45 152 km (altitude 38 781 km — above GEO altitude). The orbit crosses GEO's radius twice per revolution: a defining design consequence (Section 5).

1.2 Apogee Dwell: Kepler's Second Law Quantified

t(ν) = (T/2π) · (E − e sin E), E = 2·atan2(√(1−e) sin(ν/2), √(1+e) cos(ν/2))

Fig. A — Time from perigee vs true anomaly. For e = 0.7 the satellite sweeps the half-sky around perigee (ν ∈ −90°…+90°) in just 9.4% of the period (~68 min), while the apogee half (ν ∈ 90°…270°) consumes 90.6% (~10.8 h).

Example 1.2 — The dwell dividend. At e = 0.7, the satellite is above the northern horizon at usable elevation for ≈ 10.8 h per orbit from 60°N (Fig. B shows 10.77 h above 5°). Three satellites phased 8 h apart tile the day exactly: each takes an 8-hour "watch" centred on its apogee dwell.

Check Your Understanding 1

Compute the fraction of a Molniya period spent between ν = −30° and +30° (around perigee).
Why does Tundra need e ≈ 0.27 rather than 0.7 to achieve its 12-h dwell per satellite?

Answers

ν = 30°: E = 2·atan(0.4201·tan15°) = 2·atan(0.1126) = 0.2243 rad; sin E = 0.2224; t/T = (0.2243 − 0.7×0.2224)/2π = 0.00865 → fraction = 2×0.865% ≈ 1.7% of the period (≈ 12.4 min) for a full 60° of perigee sky.
Tundra's day-long period means each satellite must carry roughly half the day (12 h) of coverage; with e = 0.27 the apogee-side half-orbit already occupies ≈ 67% of the day (ν ±90° spans 33%), so two sats phased 12 h apart overlap generously — e = 0.7 would over-dwell (92%) and waste perigee altitude margin.

Section 2 — The Critical Inclination

2.1 The Problem: Perigee Precession

Earth's equatorial bulge (J₂) torques orbital planes, making the argument of perigee ω drift:

dω/dt = (3/4)·J₂·(R_E/p)²·n·(5 cos²i − 1), p = a(1−e²), J₂ = 1.0826×10⁻³

For most inclinations this is nonzero: perigee wanders, and with it the apogee — the dwell point drifts out of the northern hemisphere within months.

2.2 The Fix: 5 cos²i − 1 = 0

i_crit = arccos(1/√5) = 63.43° (or 116.57° retrograde) ⇒ dω/dt = 0 exactly

At 63.4° the J₂ torque vanishes: the ellipse freezes in inertial space, perigee parked over the southern hemisphere, apogee anchored over the north — forever, with no propellant. This is the single most beautiful parameter choice in operational satellite engineering.

Example 2.1 — What would go wrong at i = 55°? At a = 26 560, e = 0.7: p = 13 274 km; (R_E/p)² = 0.2304; n = 2π/43 082 s = 1.459×10⁻⁴ rad/s. dω/dt = 0.75 × 1.0826×10⁻³ × 0.2304 × 1.459×10⁻⁴ × (5cos²55° − 1) = 2.732×10⁻⁸ × (5×0.3290 − 1) = 2.732×10⁻⁸ × 0.645 = 1.76×10⁻⁸ rad/s ≈ 31°/year. Within ~3 years apogee would wander out of the service hemisphere; at 63.4°, zero.

2.3 Historical Note

The resonance was known from celestial mechanics (it also freezes the argument of perigee of theMoon's orbit — i = 5.1° is far from critical, which is why the Moon's perigee precesses over 8.9 years). Soviet engineers turned the anomaly into an architecture in 1965: Molniya-1 inaugurated both the orbit and the practice of naming a system after its orbit.

Check Your Understanding 2

Show that i = 63.43° satisfies 5cos²i − 1 = 0.
Does the critical inclination remove ALL orbit perturbations? List two that remain.

Answers

cos²(63.43°) = (1/√5)² = 1/5 → 5×(1/5) − 1 = 0. ✓
No — Ω (RAAN) still precesses (different J₂ term), and lunisolar perturbations act on ω and i at higher order; Molniya sats perform periodic station-keeping, just vastly less than at other inclinations.

Section 3 — Coverage and the Three-Satellite Constellation

3.1 The Elevation Argument

Molniya apogee sub-point: (63.4°N, longitude = −RAAN); elevation: sin el = (cos γ − k_a)/√(1 + k_a² − 2k_acos γ), k_a = R_E/r_a = 0.1411

Fig. D — Elevation vs latitude: Molniya apogee dwell vs GEO. At 60°N the Molniya satellite stands ~86° overhead while a GEO satellite on the meridian struggles at ~22°. Low elevation is not a comfort issue — it costs rain margin, noise temperature, and terrain clearance.

3.2 Continuous Coverage by Phasing

Fig. C — Ground tracks of three Molniya satellites phased 8 h apart over one day. Near apogee each trace lingers at high northern latitude (slow in-track motion — the dwell), then races through perigee over the southern hemisphere. Together they hold the northern dwell region continuously.

Spacing: 3 satellites phased by T/3 ≈ 8 h: each is the "apogee satellite" for ~8 h while the previous one descends and the next ascends.
Handover: switch service between satellites at roughly equal elevation (~mid-dwell), a gentle event compared with LEO's minute-scale beam hops.
Coverage zone: the union of the three dwell footprints blankets latitudes ≳ 50°N continuously at high elevation — the exact region where GEO drops below 15–20°.

Check Your Understanding 3

From Fig. D, find the latitude above which Molniya apogee dwell beats GEO's elevation by more than 40°.
Why must the three satellites share the same orbital plane rather than three different planes?

Answers

≈ 45°N (GEO ~33°, Molniya ~73° there).
Apogee dwell must occur over the same longitude band for continuity; three RAANs would place three apogees at three different longitudes, fragmenting coverage into three zones with gaps between them. Phasing within one plane keeps a single seamless dwell region.

Section 4 — HEO Link Engineering

4.1 The Budget Varies Over the Orbit

Fig. B — Simulated Molniya orbit from 60°N: elevation and FSPL (1.5 GHz) vs time. The satellite is usable for ~10.8 h; path loss varies by ~10 dB between the apogee dwell and the low-elevation approach; a perigee gap (satellite invisible or at unusable elevation) recurs every 12 h — filled by the phased companions.

Example 4.1 — Two-budget design. L-band (1.5 GHz) downlink from apogee: d ≈ 39 800 km (slant from 60°N at apogee) → L_FS = 92.45 + 3.5 + 92.0 ≈ 188 dB. With satellite EIRP 30 dBW (Molniya-class), ground G/T 15 dB/K, misc 2 dB: C/N₀ = 30 + 15 − 188 − 2 + 228.6 = 83.6 dBHz — a comfortable voice/data link. Near perigee the range shrinks (~8 000 km → 10 dB less path loss) but elevation collapses: the link would be easy if it were usable — HEO's link problem is geometry, not power.

4.2 Dynamics the Budget Must Respect

Effect	Magnitude	Design response
Range variation	slant 8 000 → 45 000 km → ~15 dB FSPL swing	Design to apogee; perigee is free margin
Doppler	±1.6 km/s apogee radial at L-band ≈ ±8 kHz; perigee passes ≈ ±45 kHz bursts	Ephemeris-driven pre-compensation; brief perigee passes often just avoided
Pointing rate	up to ~0.5°/s near perigee; ~0.05°/s in dwell	Track at dwell rates; switch sats rather than track perigee
Delay	133 ms one-way at apogee	GEO-class PEP thinking for data links
Rain margin	high elevation → short atmospheric path → small fade	HEO's quiet advantage over low-elevation GEO at high latitudes

Check Your Understanding 4

Why is the perigee pass typically not used for service even though path loss is minimal?
Compute the apogee slant range from a station at 63.4°N directly under apogee, and the corresponding FSPL at 4 GHz.

Answers

Geometry: at 60°N the satellite is below the horizon or under 5° for most of the perigee half of the orbit (it is over the southern hemisphere); the visible sliver sits at minimum elevation with maximum atmospheric noise, shortest pass duration, and highest Doppler — poor economics all round.
γ ≈ 0 → d = h_a = 38 781 km; L_FS = 92.45 + 12 + 20log(38 781) = 92.45 + 12 + 91.8 = 196.2 dB.

Section 5 — HEO Spacecraft Design

5.1 The Radiation Tax, Twice Per Orbit

Perigee at 1 597 km: skims the inner Van Allen belt's lower edge — brief but intense passes.
Apogee at 38 781 km: sits inside the outer electron belt for hours of every orbit.
Consequence: total dose between LEO and MEO levels, but with the twist that charging environments alternate: electron-dominated at apogee (deep dielectric charging), proton contamination spikes at perigee. Rad-hard design is mandatory — HEO satellites cost GEO-class money.

5.2 Power, Thermal, and the GEO-Crossing Hazard

Solar array + battery sized for the apogee dwell (hours of continuous service) with eclipse seasons near equinoxes.
Thermal: apogee dwell = long sunlit periods at fixed attitude; perigee = rapid attitude re-acquisition — radiator and heater budgets must cover both.
Debris conjunctions at GEO radius: the orbit crosses the geostationary ring twice per revolution; conjunction screening against the GEO catalog is a unique HEO operational duty.

5.3 Longevity and Replacement

Molniya sats historically served 3–5 years (radiation + propellant for inclination/ω maintenance against lunisolar perturbations); modern designs with electric propulsion and better parts extend this. Replacement requires the full equatorial-plane launch (no direct-injection shortcuts), making on-orbit sparing attractive.

Check Your Understanding 5

Why does a HEO satellite face the outer electron belt for hours while an MEO GNSS satellite only traverses it on its way elsewhere?
Name two operational duties a HEO control centre performs that a GEO control centre does not.

Answers

MEO circular orbits reside at one radius — GNSS at 20 180 km sits inside the belt continuously too (Section 4 of the MEO guide), but a HEO satellite sweeps from belt to belt and back twice per orbit, experiencing both charging regimes every 12 h.
(1) Conjunction screening against the GEO catalog at the two ring crossings; (2) satellite switching choreography between the three phased satellites (GEO has no constellation handover).

Section 6 — The Ground Segment: Track or Switch?

6.1 Two Architectures

Architecture	Mechanism	Cost / behaviour
Tracking stations	One antenna tracks the dwell satellite (~0.05°/s), re-acquires the next	Higher capex; seamless links; suits continuous traffic (Molniya practice)
Fixed pointing + switch	Three antennas parked near the three apogee regions; traffic switches at handover	Cheaper antennas; a brief switch interruption per 8 h; Sirius XM practice with its Tundra trio

6.2 Handover Discipline

The 3-satellite switch is engineered to occur when the ascending satellite reaches the descending satellite's elevation — typically 30–40° elevation, both links healthy. At that moment both links are briefly active (make-before-break), and the ground network re-routes in milliseconds. Compare LEO's minute-scale beam hops: HEO handovers are civilized — one every 8 hours.

Check Your Understanding 6

Why is make-before-break switching easy at HEO handover but hard for LEO?
Estimate the tracking rate burden: apogee dwell apparent motion ≈ 0.05°/s vs a GEO satellite's 0.004°/s — what does this mean for antenna mount design?

Answers

At HEO handover both satellites sit at high elevation with hours-long dwells — receivers can lock both weeks in advance; LEO handover happens mid-pass with Doppler still changing on both links, leaving no calm window.
HEO tracking is ~12× GEO's rate — still trivial for modern motorized mounts (rates of 1–5°/s are standard); the real burden is the perigee pass, which is exactly why systems switch instead of tracking through it.

Section 7 — Systems and Case Studies

System	Orbit	Purpose	Lesson
Molniya-1…3 (1965– )	Molniya, 3-sat planes	National TV, telephony, military comms over USSR	Proved HEO economics: three satellites did what dozens of LEO sats could not — persistent high-latitude coverage with GEO-class simplicity
Sirius XM (2000– )	Tundra, e ≈ 0.27, 24 h, i = 63.4°, 3 sats	Satellite radio (CONUS)	HEO for a commercial consumer market: elevation is product quality — tall buildings block low-elevation GEO signals; 60–70° elevation fixed that. (Sirius sats used inclined elliptical orbits with continental apogee dwell)
Russian EKS/Kupol	Molniya	Missile early warning	The stare mission: apogee dwell over the northern horizon watches for depressed-trajectory launches that GEO-radars see late
Proposed HEO-PNT/SDA	Molniya variants	Polar navigation augmentation; space domain awareness	Modern revival: GNSS + HEO sats would close the polar gap for aviation and shipping
Interim 1960s US	—	Advent (GEO) vs Molniya observation	The US initially dismissed HEO; the 1968 White House ordered a Molniya-class study — the orbit's utility was never in doubt after that

The pattern: HEO succeeds wherever the customer values elevation (Sirius), northern persistence (Molniya, EKS), or both. It fails whenever a business model could use GEO or LEO instead — see the ICO/Odyssey lesson from the MEO guide applied here: orbit choice cannot rescue weak economics.

Section 8 — Trade-offs: HEO vs the Alternatives

Axis	HEO (Molniya)	GEO	LEO polar
High-latitude elevation	60–90° sustained	<30° above 55°N	variable, passes of minutes
Satellites for N-hemisphere continuity	3 (one plane)	impossible (geometry)	dozens (Iridium-class)
Ground antenna	track slowly or 3 fixed + switch	fixed dish	tracking/e-scan
Latency	~130 ms at apogee	~120 ms	~5–30 ms
Radiation	both belts, every orbit	outer belt edge (transfer only)	minimal
Debris exposure	crosses GEO ring 2×/orbit	resides in the ring	densest shell
Per-satellite cost	GEO-class (rad-hard)	GEO-class	low
Constellation ops	3 sats, one plane, calm	single sat, simplest	hundreds–thousands, intense
Best missions	Arctic comms, EW stare, polar augmentation	equatorial/mid-lat broadcast & trunk	global capacity, latency, imaging

The verdict: HEO is not a general-purpose orbit — it is the geometrically inevitable answer to one specific question: continuous high-elevation coverage of latitudes above ~55°. Where GEO cannot see and LEO cannot stay, three elliptical satellites in one plane watch forever.

Section 9 — Tutorial Problems

Constants: μ = 3.986×10⁵ km³/s², R_E = 6 371 km, sidereal day = 86 164 s, c = 3×10⁵ km/s, J₂ = 1.0826×10⁻³, FSPL = 92.45 + 20log f(GHz) + 20log d(km).

Derive the Molniya semi-major axis from T = ½ sidereal day, then compute r_p, r_a, and both velocities for e = 0.7.
Using the eccentric-anomaly equation, compute the fraction of the period for which a Molniya satellite is between ν = 120° and 240° (the core dwell window), and convert to hours.
Compute dω/dt at i = 63.4° and at i = 70° for a Molniya orbit, and comment on the perigee drift over a 5-year mission at i = 70°.
Design a Tundra orbit: a (24 h), e such that the apogee dwell half-orbit spans ~14 h; compute r_a, r_p, and dwell fraction between ν ±90°.
Compute elevation at 65°N for (a) Molniya at apogee (sub-point 63.4°N, same meridian) and (b) GEO on the meridian. Then compute the corresponding slant ranges and FSPL at 1.5 GHz, and quantify HEO's path advantage at this latitude.
A HEO link at 4 GHz uses a 4 m ground dish (G/T = 22 dB/K) and a satellite EIRP of 33 dBW at apogee. Budget C/N₀ for the apogee dwell from 60°N (slant 39 800 km), then estimate the margin the same station would enjoy against the same satellite at GEO altitude on the same meridian.
Three Molniya satellites must guarantee ≥20° elevation continuously from 55°N. Determine the maximum phasing error (in minutes) tolerable between satellites before the handover dip breaches 20°, using Fig. B's dwell shape as data (state your reading of the elevation at ±1 h around handover).
Radiation: estimate the number of belt transits per year for a Molniya satellite, and contrast with a GEO satellite's belt exposure.
Constellation economics: compare the 3-satellite Molniya option vs a 66-satellite polar LEO option for a national Arctic operator on (a) ground segment cost, (b) satellite unit cost, (c) handover complexity, (d) latency — in a table with a 1-line verdict.
Design essay (1 page): "HEO-PNT: a Molniya constellation as a GNSS back-up for polar aviation." Cover geometry (elevation advantage), continuity (3-sat phasing), the radiation/atomic-clock challenge at apogee dwell, and dual-frequency signal design for scintillating polar ionospheres.

Selected solutions

a = (3.986×10⁵ × 43 082²/39.478)^1/3 ≈ 26 560 km; r_p = 7 968 km (alt 1 597), r_a = 45 152 km (alt 38 781); v_p = √(μ(2/7968 − 1/26560)) ≈ 9.2 km/s; v_a ≈ 1.6 km/s.
ν = 120°: E = 2·atan(0.4201·tan60°) = 2·atan(0.7276) = 1.211 rad; sin E = 0.9360; t/T = (1.211 − 0.6552)/2π = 0.0888. ν = 240° → by symmetry 1 − 0.0888 = 0.9112. Fraction = 0.8224 ≈ 82% ≈ 9.84 h.
At 63.4°: 5cos²i − 1 = 0 → dω/dt = 0. At 70°: 5(0.1170) − 1 = −0.415; pre-factor 2.732×10⁻⁸ rad/s → dω/dt = −1.13×10⁻⁸ rad/s ≈ −20°/yr → over 5 years ≈ −100°: apogee leaves the service hemisphere mid-mission — the critical inclination is not optional.
a = 42 164 km. Require dwell (ν ±90°) ≈ 14/24 h = 58% → t/T(90°) = 0.21 → solving (E − e sin E)/2π = 0.21 gives e ≈ 0.13–0.14 (r_a ≈ 47 700 km, r_p ≈ 36 600 km). Practical Tundra designs choose e ≈ 0.25–0.27 for a ≈ 66% dwell (≈ 16 h: r_a ≈ 53 500 km, r_p ≈ 30 800 km), overlapping generously with 12-h phasing — the margin covers station-keeping and handover dips.
(a) γ = 1.6° → sin el = (0.9996 − 0.1411)/√(1.0199 − 0.2821) = 0.8585/0.8590 → el ≈ 88°; d ≈ 38 781 km. (b) γ = 65° → sin el = (0.4226 − 0.1511)/√(1.0228 − 0.1277) = 0.2715/0.9459 = 0.2870 → el ≈ 16.7°; d = √(6371²sin²16.7 + 35786² + 2·6371·35786) − 6371·sin16.7 ≈ 39 885 km. FSPL nearly equal (187.7 vs 188.0 dB) — HEO's advantage is elevation (rain margin, noise, terrain), not path length.
C/N₀ = 33 + 22 − (92.45+12+20log 39800=91.99 → 196.4) − 2 + 228.6 = 85.2 dBHz. GEO comparison: same budget, el 21.9° → atmospheric path longer: add ~1 dB gaseous + rain margin difference ~5 dB → HEO enjoys ≈ 6 dB margin advantage at 60°N.
From Fig. B, elevation at ±1 h from the dwell centre ≈ 35–40° and at ±2 h ≈ 15°. A 20° floor is crossed ≈ 1.6 h from centre; with 8-h watches the dip between sats stays above 20° if phasing error < ±20 min (state assumption: linear dwell shape near handover).
2 transits per 11.97 h orbit → 2 × 730 orbits/yr ≈ 1 460 belt transits/yr; GEO satellites skirt the belts only during transfer orbit — effectively zero operational exposure.
Table: ground — HEO: 1–3 tracking/switch antennas, simple; LEO: tracking network or e-scan fleet, complex. Sats — HEO: 3 rad-hard GEO-class units; LEO: 66 cheap sats + replacement churn. Handover — HEO: 1 calm switch/8 h; LEO: continuous beam hops. Latency — HEO ~130 ms; LEO ~30 ms. Verdict: for pure Arctic persistence, HEO wins on operations; add latency-sensitive broadband and LEO re-enters.
Model essay: elevation 60–90° at 65°N fixes polar aviation's GNSS-geometry problem; 3 sats phased 8 h give continuity; atomic clocks at apogee face outer-belt electrons — optical clocks/radiation-hardening budget; dual-frequency (L1/L5-like) counters high-latitude ionospheric scintillation; conclude HEO-PNT is a credible, under-used complement to MEO GNSS.

Section 10 — Glossary of Key Terms

Apogee dwell: The long, slow arc near apogee (Kepler's second law) during which the satellite serves its coverage zone — 90.6% of the period at e = 0.7.
Argument of perigee (ω): Angle from ascending node to perigee; ω = 270° parks perigee over the southern hemisphere.
Critical inclination: 63.43° (or 116.57° retrograde): the J₂ argument-of-perigee precession vanishes; the HEO signature parameter.
Eccentric anomaly (E): Auxiliary angle relating mean and true anomaly; drives the time equation t = (T/2π)(E − e sin E).
Graveyard consideration (HEO): HEO disposal orbits are chosen off the operational planes; dead Molniya sats are re-entered by lowering perigee into the atmosphere — an option GEO lacks.
Molniya orbit: a = 26 560 km, e = 0.7, i = 63.4°, T = ½ sidereal day; named for the Soviet system (1965) that inaugurated it.
Perigee gap: The hours around perigee when the satellite is at unusable elevation for the northern zone — filled by phased companions.
Phase (constellation): Time offset between satellites in the same plane; T/3 ≈ 8 h for three-satellite Molniya continuity.
True anomaly (ν): Angle from perigee to the satellite's current position, measured at Earth's centre.
Tundra orbit: 24-h period, e ≈ 0.27, i = 63.4°; apogee dwell ~2/3 of the day per satellite; Sirius XM heritage.

Section 11 — References and Further Reading

V. A. Chobotov (ed.), Orbital Mechanics, 3rd ed., AIAA, 2002 — perturbations and the critical inclination.
J. R. Wertz & W. J. Larson (eds.), Space Mission Analysis and Design, 3rd ed., Microcosm Press, 1999.
T. Pratt, J. Bostian & J. Allnutt, Satellite Communications, 2nd ed., Wiley, 2003.
W. L. Pritchard, H. G. Suyderhoud & R. A. Nelson, Satellite Communication Systems Engineering, Prentice-Hall, 1993 — HEO link practice.
Public mission data: NASA/NSSDC master catalog (Molniya, Sirius XM, EKS entries) — orbit parameters of record.

Exam guidance: HEO questions reward the chain: Kepler → time equation → dwell fraction → coverage/elevation → constellation phasing → budget. Quote the eccentric-anomaly equation before dwell numbers, and always show that 63.4° comes from 5cos²i − 1 = 0. The most common error: computing dwell as an angle ratio instead of a time ratio — Kepler's second law exists precisely because they differ.