How to Build a Satellite from Scratch: Orbital Mechanics, Materials, Power, RF, Propulsion, and Integration — The Complete A-to-Z Engineering Guide

A senior-engineer-level, equation-dense guide covering every subsystem of a satellite: Keplerian orbits and delta-v budgets, structural materials and mass…

2026-07-29 · FabFlow Team

How to Build a Satellite from Scratch: Orbital Mechanics, Materials, Power, RF, Propulsion, and Integration — The Complete A-to-Z Engineering Guide

Satellites are one of the hardest engineering challenges on (and off) Earth. You're building an autonomous vehicle that must survive launch vibration, operate in hard vacuum through thermal swings of ±150°C, point antennas and solar panels with arcsecond precision, and do all of this without a single maintenance visit for 3–15 years.

This guide covers every subsystem at the level of detail an actual builder needs. Equations, component selection criteria, failure modes, and real-world cost numbers are all included. No hand-waving.

1. Orbital Mechanics: The Math That Keeps You in Space

Before choosing a bolt or a battery, you need to know where you're going and how much energy it takes to get there. Orbital mechanics determines your launch vehicle selection, propulsion budget, thermal environment, eclipse duration, and ground station contact windows.

1.1 Keplerian Elements

Six parameters fully describe a satellite's orbit around a central body:

Where:

Element · Symbol · Meaning

Semi-major axis · a · Orbit size (km)

Eccentricity · e · Shape (0 = circular, <1 = elliptical)

Inclination · i · Tilt relative to equatorial plane (°)

RAAN · \Omega · Right ascension of ascending node (°)

Argument of perigee · \omega · Orientation of ellipse in orbital plane (°)

True anomaly · \nu · Position along orbit at epoch (°)

For a circular orbit (e = 0), orbital period comes from Kepler's Third Law:

Where \mu = GM is the gravitational parameter. For Earth:

\mu_\oplus = 3.986 \times 10^{14} \text{ m}^3/\text{s}^2
At 400 km altitude (ISS): a = 6778 \text{ km}, T \approx 92.7 \text{ min}
At 35,786 km (GEO): T = 23^\text{h}56^\text{m}4^\text{s} (sidereal day)

LEO velocity: At 400 km circular orbit:

1.2 Delta-v Budgets: The Currency of Spaceflight

Delta-v (\Delta v) is the total change in velocity your propulsion system must deliver. The Tsiolkovsky rocket equation governs what's possible:

Where I_{sp} is specific impulse (seconds), g_0 = 9.80665 \text{ m/s}^2, m_0 is wet mass, and m_f is dry mass.

Hohmann transfer (minimum-energy elliptical transfer between two circular orbits):

Transfer · \Delta v Total · Typical I_{sp} · Propellant Mass Fraction

LEO (200 km) → LEO (800 km) · ~0.32 km/s · 220 s (hydrazine monoprop) · 13.8%

LEO → GTO · ~2.5 km/s · 320 s (biprop MMH/NTO) · 54.9%

LEO → Lunar transfer · ~3.2 km/s · 320 s · 63.9%

GTO → GEO circularization · ~1.5 km/s · 320 s · 37.9%

LEO → Earth escape · ~3.2 km/s · 1500 s (ion thruster) · 19.6%

For a CubeSat: Most 1U–6U missions have \Delta v < 100 \text{ m/s} — just enough for orbit maintenance and deorbit at end-of-life. Launch vehicles deliver the hard 7.8 km/s.

1.3 Perturbations That Matter

J2 perturbation (Earth's oblateness): The dominant non-spherical term causes nodal precession:

Where J_2 = 1.08263 \times 10^{-3}, R_\oplus = 6378 \text{ km}, and n = \sqrt{\mu/a^3} is mean motion.

Sun-synchronous orbit: Choose inclination so \dot{\Omega}_{J2} \approx 0.9856^\circ/\text{day} (matches Earth's orbital rate around Sun):

At 600 km: i_{SSO} \approx 97.8^\circ. At 800 km: i_{SSO} \approx 98.6^\circ.

Atmospheric drag: Below 500 km, drag is significant:

Where \rho \approx 10^{-11} kg/m³ at 400 km (varies with solar activity), C_d \approx 2.2 for typical satellite geometries, A is cross-sectional area. A 3U CubeSat (10×10×34 cm, ~4 kg) at ISS altitude decays ~1 km/month without reboost.

2. Structural Design and Materials Selection

2.1 The Load Cases

Your structure must survive three regimes:

Launch: Quasi-static loads of 6–12g axial + 3–6g lateral, plus random vibration of 10–14 g<sub>RMS</sub> from 20–2000 Hz. Sinusoidal vibration at 5–100 Hz with amplitudes up to 10 mm.

On-orbit: Micro-vibration from reaction wheels (< 0.1g), thermal expansion mismatch.

Deployment: Shock loads from pyrotechnic release mechanisms (1000–5000g at >1 kHz, but only for microseconds).

2.2 Material Selection Matrix

Material · Density (g/cm³) · Young's Modulus (GPa) · Yield Strength (MPa) · CTE (μm/m·K) · Thermal Cond. (W/m·K) · Use Case

Al 6061-T6 · 2.70 · 69 · 276 · 23.6 · 167 · Primary structure, mounting panels

Al 7075-T6 · 2.81 · 71 · 503 · 23.6 · 130 · High-stress fittings, fastener inserts

Ti-6Al-4V · 4.43 · 114 · 880 · 8.6 · 6.7 · Propellant tanks, high-temp components

CFRP (M55J/cyanate) · 1.60 · 300+ · 800+ (fiber) · 0.1–1.0 · 50–150 · Solar array substrates, optical benches

Al honeycomb (5052, 1/8" cell) · 0.05–0.08 (core) · — · — · — · — · Sandwich panel cores

Invar 36 · 8.05 · 141 · 276 · 1.3 · 10.5 · Dimensionally stable optical mounts

The golden rule: Match CTE across bonded interfaces. A CFRP face sheet bonded to an aluminum honeycomb core with mismatched CTE will delaminate after 1000 thermal cycles (2–3 years in LEO). Use cyanate ester resin systems (CTE ≈ 1–3 μm/m·K) instead of epoxy (CTE ≈ 30–60 μm/m·K).

2.3 Mass Budget — First-Order

For a 100 kg LEO satellite (ISRO's IRS-class):

Subsystem · Mass (kg) · % of Total

Structure + mechanisms · 12–18 · 12–18%

Power (solar + battery + EPS) · 18–25 · 18–25%

ADCS (wheels, sensors, actuators) · 6–10 · 6–10%

Propulsion (wet) · 10–20 · 10–20%

RF communications · 4–8 · 4–8%

Avionics/OBC · 3–5 · 3–5%

Thermal (MLI, heaters, radiators) · 2–4 · 2–4%

Payload · 25–35 · 25–35%

Harness + misc · 4–6 · 4–6%

CubeSat mass: A 3U CubeSat design mass is 4.0 kg per the CubeSat Design Specification (Rev 14). Actual dry mass for well-built 3U birds: 3.2–3.8 kg, leaving 200–800g for payload integration margin.

3. Power Systems: Generation, Storage, and Distribution

3.1 Solar Array Sizing

The power generated by a solar panel is:

Where:

S = 1361 \text{ W/m}^2 (solar constant at 1 AU)
\eta_{cell} = cell efficiency: 28–32% for space-grade triple-junction GaInP/GaAs/Ge (AzurSpace 3G30C), 18–22% for Si
\cos\theta = sun incidence angle (degraded if not sun-pointing)
\eta_{path} = 0.85–0.92 (diode losses, wiring, string mismatch)
d = annual degradation: ~0.5–1.0%/year for GaAs, 2–3%/year in high-radiation orbits
t = mission year

Real example — 3U CubeSat: Two deployable panels (2U each, ~0.06 m² total per side) plus one body-mounted panel (~0.03 m²):

Orbit-average power with 60% sunlight fraction: P_{avg} \approx 32 \text{ W}.

3.2 Battery Sizing

Energy storage needed to survive eclipse:

LEO eclipse: Up to 35 minutes at 600 km SSO. GEO eclipse: Up to 72 minutes during equinox seasons.

For a 20W average load during 35-min LEO eclipse, with 30% depth-of-discharge (DoD) for 5000+ cycle life:

At 3.7V nominal (Li-ion), that's 44.5 / 3.7 = 12.0 \text{ Ah}. Standard 18650GA cells (3.5 Ah, 48g each): 4S4P (16 cells, 768g, 14 Ah, 207 Wh/kg pack level). Use Saft VES16 or ABSL 18650HC for flight heritage.

Key battery requirement: Cells must be vacuum-compatible (no liquid electrolyte venting). Standard commercial 18650s use vent mechanisms designed for 1 atm — in vacuum the electrolyte boils and the cell vents. Space-grade cells use hermetic glass-to-metal seals and low-vapor-pressure electrolytes (e.g., LiPF6 in EC:DMC with FEC additive).

3.3 EPS Architecture

A typical regulated bus uses peak-power tracking:

Direct Energy Transfer (DET): Solar arrays connected to bus through shunt regulator. Simple, heritage (ISRO's standard). Less efficient when P_{gen} > P_{load}.
Maximum Power Point Tracking (MPPT): DC-DC converter tracks array V_{mp} vs I_{mp} curve. 5–15% more energy capture. Uses perturb-and-observe or incremental conductance algorithm on a rad-hard MCU (MSP430FR or Vorago VA10820).

Battery charge regime: CC-CV with temperature-compensated termination voltage (-3.7 mV/°C/cell for Li-ion). Charge rate limited to 0.5C to prevent lithium plating at low temperatures (< 5°C).

4. Avionics and On-Board Computer (OBC)

4.1 Processor Selection: Why Your Raspberry Pi Won't Fly

The Van Allen belts deliver total ionizing dose (TID) of 10–100 krad(Si) over a 3-year LEO mission. A single heavy ion can cause:

Single Event Upset (SEU): Bit flip in memory. Mitigated by ECC memory + triple-modular redundancy (TMR).
Single Event Latchup (SEL): Parasitic SCR triggers in CMOS, drawing high current until power cycle. Destroys non-latchup-immune parts within seconds.
Total Ionizing Dose (TID): Gradual threshold voltage shift in MOSFETs, eventually preventing switching.

Component · TID Tolerance · SEL Immune · Use Case

Vorago VA10820 (ARM Cortex-M0) · 300 krad(Si) · Yes · Primary OBC

Microchip ATmegaS128 · 30 krad(Si) · Yes · Watchdog, low-criticality tasks

Xilinx Kintex UltraScale (XQRKU060) · 120 krad(Si) · Yes (rad-hard version) · Payload data processing

Raspberry Pi CM4 · < 5 krad(Si) · No · GROUND TESTING ONLY

Watchdog strategy: Dual-redundant external watchdog timers (MAX706R) with independent oscillators. OBC must pet within 1.6s or the satellite reboots. After 3 reboots in < 60 seconds, enter safe mode: deploy antennas, point solar panels at sun, beacon at 1200 bps on UHF.

4.2 Sensor Suite

Sensor · Accuracy · Update Rate · Mass · Used For

Sun sensor (CoSS, 6× coarse) · ±0.5° · 10 Hz · 120g total · Safe-mode sun acquisition, coarse pointing

Magnetometer (HMC1053 or fluxgate) · ±50 nT · 20 Hz · 30g · Magnetic field vector for detumbling + coarse attitude

Star tracker (Arcsec Sagitta or ST-200) · ±5 arcsec · 4 Hz · 350g · Precision attitude (quaternion output)

IMU (ADIS16490 or LN-200S) · Gyro bias: 0.005°/s · 100 Hz · 50g · Angular rate for propagation between star tracker updates

GPS receiver (NovAtel OEM719) · ±5m position, ±0.05 m/s · 1 Hz · 40g · Orbit determination (LEO only — GPS signal weakens above 3000 km)

4.3 Bus Architecture

CAN bus (1 Mbps, differential): Used by ISRO's I-2K bus, proven in radiation environments. Terminate with 120Ω at each end.
I²C (400 kHz, intra-board only): Sensor communication. Add I²C isolator (ADuM1250) to prevent a failed sensor from pulling down the entire bus.
SpaceWire (200 Mbps, LVDS): ESA standard for payload data. Router-based topology with redundant paths. Use the ESA SpaceWire RMAP protocol for register access.

5. Attitude Determination and Control System (ADCS)

5.1 Quaternion Math (Why Euler Angles Are Dangerous)

Euler angles suffer from gimbal lock — at 90° pitch, roll and yaw become indistinguishable. Quaternions don't:

With \|\mathbf{q}\| = 1. Rotation from body to inertial frame:

Quaternion propagation (using gyro angular velocity \boldsymbol{\omega}):

Attitude determination: Use a Multiplicative Extended Kalman Filter (MEKF). State vector: [\mathbf{q}_{error}, \Delta\mathbf{b}_{gyro}]^T. Measurements: star tracker quaternion + magnetometer vector. The MEKF estimates gyro bias in real-time, achieving 0.001°/hr stability on a Cold Atom Physics Laboratory-class instrument.

5.2 Actuator Selection

Actuator · Torque Range · Power · Precision · Best For

Reaction wheel (0.1–1.0 Nms) · 1–50 mNm · 1–10W at max torque · ±0.001° · Precision pointing, slewing

Magnetorquer (0.1–5 Am²) · 0.001–1 mNm (LEO) · 0.1–2W · ±1° · Momentum dumping, detumbling

Cold gas thruster (0.1–1N) · 0.01–1 Nm (lever arm) · — (propellant) · ±0.1° · Orbit maintenance + attitude

Control moment gyro (CMG) · 10–100 Nm · 50–200W · ±0.0001° · Agile spacecraft (WorldView, Pleiades)

CubeSat wheels: Sinclair Interplanetary 10 mNms reaction wheel — 185g, 0.4W at nominal, stores 0.015 Nms at 6500 RPM. Four in a tetrahedral configuration give full 3-axis control with single-fault tolerance.

Momentum dumping: Reaction wheels accumulate momentum from disturbance torques (drag, solar radiation pressure, gravity gradient). A 3U CubeSat absorbs ~10^{-7} Nm of cumulative disturbance per orbit. After ~500 orbits, wheels approach saturation. Magnetorquers dump this by interacting with Earth's magnetic field (B \approx 25\text{–}65\ \mu\text{T} at 400 km). Required magnetic moment:

For a 1 mNm torque at 40 μT field: m \approx 25 \text{ Am}^2 — achievable with 200 turns of AWG 28 around a 10×10 cm former.

6. RF Communications and Link Budget

6.1 The Friis Equation

The fundamental relationship for free-space path loss:

All terms in dB. For a 600 km LEO pass at UHF (437 MHz, \lambda = 0.686 m):

Parameter · Value

P_t (TX power) · +30 dBm (1W)

G_t (TX antenna gain) · +2 dBi (dipole)

G_r (RX antenna gain) · +14 dBi (crossed Yagi)

d (slant range at 10° elevation) · 2,674 km

Free-space path loss: 20\log_{10}(4\pi d/\lambda) · −153.9 dB

Atmospheric loss + polarization mismatch · −3 dB

Received power (P_r) · −110.9 dBm

Receiver sensitivity: A decent UHF receiver (FUNcube Dongle Pro+, RTL-SDR with LNA) achieves −124 dBm at 9600 bps BPSK with 10^{-5} BER. Link margin: −110.9 − (−124) = 13.1 dB — solid. For higher data rates, the SNR degrades:

Where k = 1.38 \times 10^{-23} J/K, T_{sys} \approx 300 K (Earth-looking), R_b is bit rate. At 1 Mbps with P_r = −110.9 dBm:

This is BELOW the ~9.6 dB required for QPSK at 10^{-5} BER. You cannot do 1 Mbps at UHF from LEO with 1W. Switch to S-band (2.4 GHz) with a high-gain patch array (G_t = +12 dBi) and a 3m ground dish (G_r = +35 dBi) for 10+ Mbps links.

6.2 Frequency Band Selection

Band · Uplink (MHz) · Downlink (MHz) · Max Data Rate (LEO, 5W TX) · Best For

VHF · 144–146 · 137–138 · 1200–9600 bps · Beacon, telemetry (AMSAT standard)

UHF · 435–438 · 401–402, 435–438 · 9.6–115.2 kbps · Command + low-rate data

S-band · 2025–2110 · 2200–2290 · 1–150 Mbps · Payload data, ISRO's standard TT&C

X-band · 7145–7235 · 8400–8500 · 50–600 Mbps · Earth observation downlink

Ka-band · 27.5–31.0 GHz · 17.7–21.2 GHz · 1+ Gbps · High-throughput (TerraSAR-X, KOMPSAT)

ITU coordination: Apply to your national administration (DoT/WPC in India) at least 2 years before launch. Frequency assignments for amateur bands (VHF/UHF) are simpler — coordinate through IARU/AMSAT.

6.3 Antenna Types

Type · Gain · Beamwidth · Deployment · Use Case

Monopole/Dipole (deployable tape-spring) · +2 dBi · Omnidirectional · Simple, 5–15g · UHF/VHF TT&C

Patch antenna (4×4 array) · +12 dBi · 30° · Fixed, body-mounted · S-band downlink

Parabolic dish (deployable) · +35 dBi · 2–5° · Complex, needs pointing · X/Ka-band high-rate

Helical (quadrifilar) · +6 dBi · 120° · Fixed · GPS L1/L2 reception

7. Thermal Control

7.1 The Thermal Balance Equation

At equilibrium, absorbed heat equals radiated heat:

Where:

\alpha = solar absorptance (0.14 for white paint, 0.92 for black Kapton)
\epsilon = infrared emittance (0.88 for white paint, 0.75 for black Kapton)
A_\odot = projected sun-facing area
Q_{internal} = internal heat dissipation (electronics)
\sigma = 5.67 \times 10^{-8} \text{ W/m}^2\text{K}^4
A_{rad} = radiating surface area

Passive control strategy: Use \alpha/\epsilon ratio coatings:

Radiator surfaces: white paint (Z93, \alpha/\epsilon \approx 0.17) — rejects heat
Cold areas: black Kapton (\alpha/\epsilon \approx 1.2) — absorbs sun, stays warm
Optical surface radiators (OSRs): quartz mirrors with \alpha/\epsilon \approx 0.08 — the gold standard for hot payloads

MLI (Multi-Layer Insulation): 10–20 alternating layers of aluminized Mylar (0.25 mil) and Dacron net spacer. Effective emittance of \epsilon_{eff} \approx 0.03. One layer of single-aluminized Kapton as outer cover (for atomic oxygen resistance). MLI can maintain a 100°C gradient across its thickness. Cost: ~₹15,000/m² for flight-grade.

7.2 Active Control

Heaters: Kapton film heaters (Minco) bonded to battery packs, propellant lines, and thrusters. Thermostat setpoints: battery ON at 0°C, OFF at 10°C; propellant ON at 5°C, OFF at 15°C.
Heat pipes: Constant Conductance Heat Pipes (CCHP) with ammonia working fluid. Grooved aluminum extrusion (\phi 8–12 mm, 0.5–2.0m length). Transport capacity: 50–500 W·m depending on diameter and tilt. Used to isothermalize large radiator panels.

8. Propulsion

8.1 Propulsion Technology Comparison

Type · I_{sp} (s) · Thrust · Power · Propellant · TRL

Cold gas (N₂, butane) · 60–80 · 0.01–10 N · 5–20W (valves) · N₂ at 200 bar · 9

Monopropellant hydrazine (N₂H₄) · 220–240 · 0.5–500 N · 5–20W (catalyst bed heater) · N₂H₄ (toxic) · 9

Bipropellant (MMH/NTO) · 310–330 · 10–40,000 N · Minimal · MMH + NTO (both toxic) · 9

Resistojet · 150–300 · 0.01–0.5 N · 10–500W · N₂, NH₃, H₂O · 8

Hall effect thruster (HET) · 1500–2500 · 10–200 mN · 200W–5kW · Xe, Kr · 9

Gridded ion thruster · 2500–4000 · 1–200 mN · 50W–7kW · Xe · 9

Green monopropellant (AF-M315E, LMP-103S) · 230–260 · 0.5–22 N · Minimal (pre-heated catalyst) · Hydroxylammonium nitrate (HAN) based · 7–8

Water electrolysis · 300–400 · 1–100 mN · 5–50W · H₂O (electrolysis → H₂ + O₂) · 5–6

For a 100 kg LEO satellite requiring 50 m/s of \Delta v for orbit maintenance:

With cold gas (I_{sp} = 70 s): propellant mass m_p = m_0(1 - e^{-\Delta v/(I_{sp}g_0)}) = 100(1 - e^{-50/(70 \times 9.81)}) = 7.0 kg.

With a miniaturized Hall thruster (I_{sp} = 1500 s): m_p = 100(1 - e^{-50/(1500 \times 9.81)}) = 0.34 kg. But you need 200W of power and 200+ hours of firing time.

8.2 CubeSat Propulsion Options

Product · I_{sp} (s) · Total Impulse · Mass (wet) · Size

Enpulsion IFM Nano Thruster · 300–800 · 2,500 Ns · 1.2 kg · 1U

Vacco MiPS (cold gas) · 46 · 99 Ns · 456g · 0.5U

Busek BIT-3 (RF ion) · 2,100 · 28,000 Ns · 1.5 kg · 1.5U

Aurora Resistojet · 150 · 1,200 Ns · 1.0 kg · 1U

The BIT-3 running on solid iodine propellant gives a 6U CubeSat roughly 300 m/s of \Delta v — enough for lunar transfer from GTO — with 60W input power and ~2 months of continuous thrusting.

9. Assembly, Integration, and Test (AIT)

9.1 The Vibration Campaign

Random vibration per NASA GEVS (GSFC-STD-7000B): 14.1 g<sub>RMS</sub> for 60 seconds per axis. Your satellite is mounted on an electrodynamic shaker. The test finds:

Loose fasteners (preload verification via torque audit before/after)
Resonant modes that couple with launch vehicle harmonics
Solder joint fatigue (especially BGA packages — use underfill for parts > 15×15mm)

Notching: If a component resonance exceeds qualification levels, the test input is "notched" down at that frequency to avoid over-testing. But you must justify this with coupled loads analysis.

9.2 Thermal Vacuum (TVAC)

8–12 thermal cycles from −25°C to +60°C at ≤10⁻⁵ Torr. Minimum 4-hour dwell at each extreme. Purpose:

Drive out volatile contaminants (outgassing requirement: TML < 1.0%, CVCM < 0.1% per ASTM E595)
Verify heater/thermostat operation in vacuum
Check for cold-start failures (electronics that work at 25°C but fail to boot at −20°C)

Contamination control: Total molecular contamination on sensitive surfaces (optical, solar arrays) must stay below 100 Å over mission life. Bake-out at 60°C for 48 hours under vacuum before installing optics.

9.3 EMI/EMC

MIL-STD-461G: CE102 (conducted emissions, 10 kHz–10 MHz), RE102 (radiated emissions, 10 kHz–18 GHz), RS103 (radiated susceptibility, 2–40 GHz at 20 V/m). Pay special attention to:

Switching power supply harmonics (100–500 kHz from DC-DC converters)
Reaction wheel PWM noise (20–50 kHz depending on motor poles)
RF transmitter desense to GPS L1 (1575.42 MHz) and star tracker electronics

10. Launch and Regulatory

10.1 Launch Costs

Launcher · LEO Capacity · Cost (₹) · Cost/kg (₹) · Rideshare 3U

ISRO PSLV (rideshare) · 1,750 kg to SSO · ~₹130 Cr · ~₹74,000/kg · ~₹15–25 Lakh

SpaceX Transporter (Falcon 9) · Shared · 1.1M for 200 kg · 5,500/kg · ~$30,000 (₹25 Lakh)

ISRO SSLV · 500 kg to 500 km · ~₹30 Cr · ~₹60,000/kg · Contact NSIL

Rocket Lab Electron · 300 kg · 7.5M · 25,000/kg · Dedicated 3U

10.2 Licensing in India

Frequency: WPC (Wireless Planning & Coordination) — license every transmitter. Form 1 + link budget analysis. Processing: 6–12 months.
Launch: IN-SPACe authorization required. Submit orbital debris mitigation plan (25-year rule: deorbit within 25 years of end-of-life or move to graveyard orbit).
Remote sensing: If your payload includes a camera with resolution better than 5m, you need additional clearance from the National Remote Sensing Centre (NRSC).
Export control: ITAR-free components only unless you have DSP-5 export license (12–18 month process). Most CubeSat components from ISISpace, AAC Clyde Space, EnduroSat are ITAR-free.

The Full Satellite Integration Sequence

Bench test each subsystem in air at ambient (1 month)
FlatSat test — all subsystems on a bench, connected with flight harness, run through full mission simulation (2 weeks)
Vibration test — 3-axis random vibration + sine burst (1 week)
Post-vibe functional — retest all subsystems, torque audit on all fasteners (3 days)
TVAC — 8 thermal cycles, hot and cold turn-on tests (2 weeks)
Post-TVAC functional — final comprehensive performance test (1 week)
Mass properties measurement — CG location within ±2 mm, moments of inertia within ±5%
Final closeout — torque all fasteners, witness marks applied, photography of every connector mated
Ship to launch site in nitrogen-purged container (humidity < 30%, temperature 15–25°C)
Launch site processing: Final battery charge, remove red-tag items, mate to deployer, wait.

Total AIT timeline: 6–15 months for a 100 kg-class satellite. 3–6 months for a well-organized CubeSat team of 5–10 engineers.

Building a satellite isn't rocket science — it IS rocket science, plus RF engineering, power electronics, control theory, thermal analysis, materials science, and project management spanning 2–5 years. But every component is documented, every equation is verified, and hundreds of student and startup teams have proven it's possible. The CubeSat revolution means you can now build flight hardware for ₹15–50 Lakhs and launch it for another ₹20–30 Lakhs. That's less than the cost of a typical seed-stage startup's engineering team for one month.

The barrier isn't cost or technology — it's the willingness to read datasheets, do the math, and accept that space has a zero-tolerance policy for oversight.