Environmental testing — thermal vacuum (TVAC), vibration, acoustic, EMC - EMI: Question bank

Related vault topics worth having open: Spacecraft Thermal Control Systems, Structural Mechanics, Launch Vehicle Dynamics, Electromagnetic Wave Propagation, Reliability Engineering, Quality Assurance in Aerospace.

Definition The symbols used on this page

Read these once, then the traps below are self-contained.

$T_{s}$ = absolute temperature of the spacecraft surface in kelvin (K); $T_{\infty}$ = absolute temperature of the surrounding deep space ( $\approx 3$ K); a bare $T$ means "the temperature of whatever body we are discussing" when only one is present. $σ = 5.67 \times 1 0^{- 8} W/(m^{2} K^{4})$ = Stefan–Boltzmann constant; $ε$ = emissivity, a number from 0 (shiny mirror, radiates nothing) to 1 (perfect black radiator); $α$ = absorptivity, the fraction of incoming sunlight a surface soaks up (0 = reflects all, 1 = absorbs all); $A$ = radiating area; $S$ = solar flux, the sunlight power landing per unit area ( $\approx 1368 W/m^{2}$ near Earth).
$ω$ = the drive frequency the shaker pushes at (rad/s); $ω_{n} = k / m$ = the structure's natural frequency, the rate it likes to wobble on its own; $f_{n} = ω_{n} /2 π$ is the same thing in hertz.
$r = ω / ω_{n}$ = the frequency ratio — how the drive compares to the natural frequency. $r < 1$ = driving slower than natural, $r = 1$ = exactly on resonance, $r > 1$ = faster.
$ζ$ (zeta) = the damping ratio, a number telling how quickly free wobbles die out (0 = never, larger = quickly); spacecraft structures sit near $ζ = 0.02$ – $0.05$ .
$Q$ = the quality factor $= 1/ (2 ζ)$ = how many times bigger the shaking gets at resonance than the input. Small damping → big $Q$ → dangerous amplification.
$Tr$ = transmissibility = ratio of the shaking that comes out at the component to the shaking put in at the base (a pure number, not a temperature). Written $Tr$ throughout so it is never confused with the temperature $T_{s}$ .
dB (decibel) for sound is $20 lo g_{10} (p / p_{ref})$ where $p_{ref} = 20 μ Pa$ is the standard reference pressure (the quietest audible pressure); so 140–150 dB is a pressure ratio against that reference, not a force.

The picture below is the single most important plot for the vibration traps: how much the shaking is amplified as $r$ changes.

And this one anchors every thermal trap — where absorbed sunlight $α S A$ and radiated power cross to set the equilibrium temperature.

True or false — justify

A sealed (hermetic) box is unaffected by vacuum.

False — sealing stops the inside from seeing vacuum, but the box still loses its convective cooling path; heat generated inside can now only escape by conduction through the mounts and by radiation, so internal parts run hotter.

Space is roughly 3 K, so a spacecraft surface in sunlight will sit near 3 K.

False — the surface reaches an equilibrium where absorbed sunlight

α S A

equals radiated power (the crossing point in the second figure); that balance sits near room temperature or hotter, and the cold 3 K sky only sets the (tiny)

T_{\infty}^{4}

sink term.

Because radiated power goes as

T_{s}^{4}

, a hot object cools much faster per degree than a cold one.

True — the derivative

d q / d T_{s} = 4 σ ε T_{s}^{3}

grows with temperature, so each extra kelvin near 350 K sheds far more watts than near 250 K, which is why hot components self-limit more strongly.

Random vibration and sine vibration test the same thing, so running one is enough.

False — sine vibration hunts and characterises individual resonances one frequency at a time, while random vibration excites all modes simultaneously to reproduce fatigue; each reveals different failures.

A component with a high natural frequency

f_{n}

is automatically safer during launch.

True in spirit — pushing

f_{n}

above the launcher's strong low-frequency content ("frequency separation") avoids resonant amplification, which is exactly why designers stiffen structures to raise

f_{n}

If the input to the shaker is 1 g, the component can never see more than 1 g.

False — at resonance the transmissibility equals the quality factor

Q = 1/ (2 ζ)

, so with light damping (

ζ \approx 0.03

) a 1 g input becomes ~15 g at the part.

Acoustic testing and vibration testing are redundant.

False — a shaker drives load in through the mounting feet, while acoustic pressure loads the whole exposed surface area of large lightweight panels (solar arrays, antennas) that the shaker can't excite well.

EMC and EMI mean the same thing.

False — EMI is the unwanted emission or susceptibility (the problem); EMC is the goal of coexisting without interference (emissions low enough, immunity high enough).

Passing acceptance-level testing proves the design is qualified.

False — acceptance tests screen a specific flight unit for workmanship at flight levels; qualification tests prove the design survives higher (margined) levels, usually on a dedicated qual unit.

A sound level quoted as 150 dB means 150 units of force on the panel.

False — dB is a logarithmic pressure ratio against

p_{ref} = 20 μ Pa

; 150 dB corresponds to a pressure of

p = 20 \times 1 0^{- 6} \cdot 1 0^{150/20} \approx 632 Pa

, and only multiplying that by area gives a force.

Spot the error

"Outgassing only matters because it wastes a little material."

The mass loss is trivial; the real danger is that released vapour condenses on cold surfaces — optics, radiators, star trackers — degrading them, which is exactly what TVAC bake-out hunts for.

"Net radiated heat is

Q = A σ ε T_{s}^{4} - T_{\infty}

Wrong — you cannot subtract a bare temperature

T_{\infty}

from a power. Both sink terms must be fourth powers under the same factors:

Q = A σ ε (T_{s}^{4} - T_{\infty}^{4})

, where

A

= radiating area,

σ

= Stefan–Boltzmann constant,

ε

= emissivity,

T_{s}

= surface temperature,

T_{\infty}

= space temperature — every term now in watts.

"At resonance,

r = ω / ω_{n} = 0

Resonance is

r = 1

(drive frequency equal to natural frequency);

r = 0

is a static load with transmissibility

Tr \approx 1

"Above resonance the component is even more amplified."

The opposite — for

r > 2

transmissibility drops below 1, so the mount actually isolates the part (see the right tail of the first figure); amplification peaks only near

r = 1

"We can skip the low-level sine sweep and go straight to full random vibration."

The sweep first finds the resonances at low, safe level; jumping to full level blind risks overtesting or destroying the article before you know where its weak frequencies are.

"Higher damping ratio

ζ

raises the resonant peak

Q

Q = 1/ (2 ζ)

, so more damping lowers the peak; that is precisely why designers add dampers to survive resonance.

"The solar panel in the worked example radiates from one face, so we divide power by 1."

A thin panel radiates from both faces, so the radiating term carries a factor of 2; ignoring it overestimates the equilibrium temperature.

Why questions

Why is convection absent in orbit but present during ground TVAC setup?

Convection needs a fluid (air) to carry heat away; the chamber is pumped to

1 0^{- 5}

–

1 0^{- 7}

Torr so almost no gas remains, faithfully reproducing space where only conduction and radiation operate.

Why does TVAC use cycling rather than one hold at the hottest temperature?

Repeated hot–cold cycling drives thermal fatigue — expansion mismatches between materials crack solder joints and delaminate boards — which a single steady soak would never reveal.

Why keep every subsystem powered and functional during the thermal extremes?

Many failures only appear when both stress and operation coincide (a marginal oscillator drifting cold, a regulator sagging hot); a powered functional checkout at each extreme catches these intermittent faults.

Why is the launch acoustic environment quoted in decibels against

20 μ Pa

rather than in newtons?

Sound is a pressure field distributed over area, and dB is a logarithmic pressure ratio referenced to

p_{ref} = 20 μ Pa

; it compactly captures the huge broadband pressure range hitting large surfaces, which a single force number cannot.

Why must a spacecraft's first natural frequency usually exceed a launcher-specified minimum?

To keep the structure stiff enough that its resonances sit above the launcher's dominant low-frequency loads, preventing dangerous coupling — a frequency-separation requirement from Launch Vehicle Dynamics.

Why is EMC testing needed even if every box individually meets its emission limit?

Compatibility is a system property — many "compliant" emitters together, plus shared harnesses and grounding loops, can still couple enough to upset a sensitive receiver; only integrated testing proves coexistence.

Why does raising a component's temperature from 300 K to 350 K nearly double its radiated power?

Because power scales as

T_{s}^{4}

, the ratio is

(350/300)^{4}

; step by step

350/300 = 1.1667

, then

1.166 7^{2} = 1.361

, and

1.36 1^{2} = 1.85

, so a modest absolute-temperature rise nearly doubles the rejected power — the basis of passive thermal self-regulation.

Edge cases

If a spacecraft's natural frequency falls inside the random-vibration band (say 300 Hz), what happens?

That mode is driven at full spectral level and amplified by its quality factor

Q

, so any weak joint or fastener at that mode is shaken hardest and typically fails — which is the whole point of covering the band.

What if emissivity

ε \to 0

(a perfect mirror surface)?

The surface can radiate almost nothing, so absorbed heat can't leave by radiation and equilibrium temperature climbs; this is why radiators use high- $ε$ coatings, not shiny metal.

What happens to transmissibility when damping

ζ \to 0

exactly at resonance?

With

r = 1

, transmissibility becomes

Tr = Q = 1/ (2 ζ) \to \infty

, so an ideal undamped structure would amplify without bound — real structures survive only because some damping always exists.

At the frequency ratio

r = ω / ω_{n} = 2

, what is the transmissibility?

Using the force-transmissibility form

Tr = \frac{1 + ( 2 ζ r ) ^{2}}{( 1 - r ^{2} ) ^{2} + ( 2 ζ r ) ^{2}}

, plug

r^{2} = 2

: the numerator is

1 + 8 ζ^{2}

and the denominator

(1 - 2)^{2} + 8 ζ^{2} = 1 + 8 ζ^{2}

— they are equal, so

Tr = 1

exactly, independent of damping. That is why

r = 2

is the universal crossover from amplification to isolation.

What if the chamber can't reach the specified vacuum before temperature cycling?

Residual gas restores a convection path, so parts run cooler than in true space and outgassing is suppressed — the test then under-stresses the article and can falsely pass a design that would overheat in orbit.

What does a shock (pyrotechnic separation) test capture that random vibration misses?

A shock is a very short, very high-g transient (hundreds to thousands of g for milliseconds) that excites brittle/fastener failures through its sharp high-frequency content, not the sustained fatigue that random vibration produces.

Does radiative exchange depend only on the two surfaces' temperatures, or also on geometry?

Also on geometry — the view factor (the fraction of one surface's radiation that actually lands on another) scales the exchange, so a component facing deep space rejects far more than one hemmed in by warm neighbours; multi-surface layouts need view-factor bookkeeping, not the simple two-surface formula.

For a spacecraft where absorbed solar power stays fixed, what happens to equilibrium temperature if you double the radiating area?

Equilibrium

T_{s}

falls, since

T_{s}^{4} \propto 1/ A

for fixed input, giving

T_{s} \to T_{s} / 2^{1/4}

— larger radiators run cooler, the core lever of passive Spacecraft Thermal Control Systems.

Recall Quick self-check

Give the transmissibility at resonance in terms of damping ::: $Tr = Q = 1/ (2 ζ)$ , so lighter damping means a taller, more dangerous peak. Name the two heat paths that survive in vacuum ::: Conduction through the structure and radiation to space — convection is gone. State the difference between EMI and EMC in one phrase ::: EMI is the interference problem; EMC is the achieved state of not interfering. What does $r$ stand for ::: The frequency ratio $r = ω / ω_{n}$ , drive frequency over natural frequency.