A gyroscope is a spinning rotor mounted in gimbals so its spin axis can hold or measure orientation independently of its support. The physics is conservation of angular momentum — a fast-spinning mass resists changes to its spin-axis direction, and any applied torque produces a perpendicular precession rather than a tilt. Engineers use this to build attitude references, stabilisers, and rate sensors that don't depend on external signals like GPS. You'll find gyroscopes in everything from a Honeywell HG1700 IMU on a guided munition to the MEMS rate gyro inside an iPhone.
Gyroscope Interactive Calculator
Vary rotor inertia, spin speed, and applied torque to see angular momentum and gyroscopic precession rate update in the diagram.
Equation Used
The calculator first converts spin speed from rpm to rad/s, then calculates angular momentum as H = I omega. For a torque applied perpendicular to the spin axis, the gyroscope precession rate is Omega = tau / H, so higher rotor inertia or spin speed reduces precession for the same torque.
- Applied torque is perpendicular to the spin angular momentum vector.
- Rotor behaves as a rigid body with constant moment of inertia.
- Steady precession is assumed, with bearing friction and gimbal losses ignored.
How the Gyroscope Actually Works
Spin a heavy disc fast enough and it develops angular momentum — the product of its moment of inertia and its spin rate. That angular momentum vector points along the spin axis, and Newton's laws say it won't change direction unless an external torque acts on it. Mount that disc in a pair of gimbals — rings that let the spin axis pivot freely on two perpendicular axes — and the rotor will hold its orientation in inertial space even as the frame around it tumbles. This is the rigidity-in-space property, and it's the foundation of every mechanical attitude reference ever built, from the Sperry artificial horizon of the 1920s to the gyrocompasses on a Nimitz-class carrier.
Apply a torque to the spin axis and something counterintuitive happens — the rotor doesn't tilt in the direction you pushed. It precesses 90° away, perpendicular to both the spin axis and the applied torque. The precession rate Ω equals the torque divided by the angular momentum (Ω = τ / H). This is why a spinning top doesn't fall over — gravity tries to tip it, but the result is a slow circular wobble around the vertical instead. Rate gyroscopes flip that physics on its head: instead of applying a torque to cause precession, they constrain the gimbal with a spring and measure the reaction torque the rotor produces when the housing rotates. That reaction is directly proportional to the input angular rate.
Get the bearings wrong and the whole thing drifts. A mechanical gyro with 0.01°/hour drift — navigation grade — needs gas bearings or magnetic suspension because rolling-element bearing friction alone produces enough parasitic torque to walk the spin axis off true by several degrees per hour. Gimbal lock is the other classic failure mode — when two gimbal axes align, you lose a degree of freedom and the rotor can no longer track motion in that plane. The Apollo 11 IMU famously had a 4th gimbal added specifically to defeat this geometry. MEMS gyros sidestep mechanical bearings entirely by using the Coriolis effect on a vibrating proof mass, but they trade drift performance — a typical consumer MEMS rate gyro drifts 10-50°/hour, fine for a phone screen rotation but useless for unaided inertial navigation.
Key Components
- Rotor: The spinning mass that stores angular momentum. Higher mass concentrated at the rim gives more moment of inertia per gram — that's why precision rotors are typically machined from beryllium or tungsten alloy with a rim-loaded geometry. Spin rates run 12,000-24,000 RPM in mechanical navigation gyros.
- Gimbals: Concentric rings that decouple the rotor's orientation from the housing. A 2-axis gimbal allows pitch and roll freedom; a 3-axis stack adds yaw. Bearing friction in the gimbal pivots is the dominant drift source — typical jewelled pivots run 0.001-0.01 N·mm of static friction, and that torque directly walks the spin axis.
- Pickoff (sensor): Measures gimbal angle relative to the housing. Synchros and resolvers were standard until the 1980s with 0.01° resolution; modern systems use optical encoders or capacitive pickoffs reaching 0.001°.
- Torquer: Applies a controlled torque to the gimbal to either re-erect the rotor or, in a strapdown rebalance loop, hold the rotor axis fixed and measure the current required. Torquer scale-factor stability sets the long-term accuracy of a rate-integrating gyro.
- Bearings or suspension: Ball bearings for industrial-grade units, gas bearings for inertial-navigation grade, electrostatic or magnetic suspension for strategic-grade. Each step down in friction roughly halves the drift rate.
- Proof mass (MEMS variants): In place of a rotor, a MEMS gyro uses a silicon mass driven into oscillation at 10-30 kHz. When the chip rotates, Coriolis force deflects the mass perpendicular to its drive direction, and capacitive plates pick up the deflection.
Real-World Applications of the Gyroscope
Gyroscopes show up wherever you need to know orientation without an external reference. The reason they're chosen over alternatives like magnetometers or GPS-based heading is twofold — they work indoors, underwater, and in jamming environments, and they respond instantly to angular rate rather than waiting for a position fix. The trade is drift: every gyro accumulates error over time, so the engineering question is always whether the drift rate is acceptable over the mission duration, and what aiding sensors will reset it.
- Aerospace navigation: Honeywell HG1700 ring laser gyro IMU used on the JDAM guided bomb and many tactical UAVs — 1°/hour drift class.
- Consumer electronics: InvenSense MPU-6050 6-axis MEMS gyro and accelerometer in the iPhone 4 and countless quadcopters like the DJI Mavic for image stabilisation and flight control.
- Marine stabilisation: Seakeeper 9 control moment gyroscope on 50-foot motor yachts — a 700 kg rotor spinning at 9,750 RPM that cancels up to 95% of roll.
- Spacecraft attitude control: Reaction wheel and CMG arrays on the International Space Station — four 100 kg double-gimbal CMGs that hold attitude without firing thrusters.
- Land vehicle navigation: Northrop Grumman LN-270 fibre-optic gyro INS in the M1A2 Abrams turret stabilisation and pointing system.
- Survey and drilling: Gyrocompass survey tools from Gyrodata used in directional oil-well drilling, where magnetic compasses fail near steel casing.
The Formula Behind the Gyroscope
The precession equation tells you how fast a gyroscope's spin axis will drift around when you apply a torque — and equivalently, how big a reaction torque the rotor produces when the housing rotates at a known angular rate. At the low end of the typical range, with a small rotor and a big input torque, precession is fast enough that you've effectively lost rigidity. At the high end, with a heavy fast-spinning rotor and a small disturbance torque, precession crawls and the gyro looks rock-solid in inertial space. The design sweet spot for a rate gyro sits where the spring-restraint torque produces a useful pickoff angle (typically 1-5��) at the maximum input rate the device must measure.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Ω | Precession rate of the spin axis | rad/s | deg/s |
| τ | Applied torque perpendicular to spin axis | N·m | lbf·ft |
| I | Moment of inertia of the rotor about spin axis | kg·m<sup>2</sup> | lb·ft<sup>2</sup> |
| ωspin | Rotor spin rate | rad/s | RPM |
| H | Angular momentum (= I × ω<sub>spin</sub>) | kg·m<sup>2</sup>/s | slug·ft<sup>2</sup>/s |
Worked Example: Gyroscope in a hobbyist boat-stabiliser CMG build
A small electric-launch builder in Maine is sizing a control moment gyroscope to damp roll on a 6 m wooden picnic boat. The rotor is a 12 kg steel disc, 200 mm diameter, spinning at a nominal 8,000 RPM. The builder wants to know how much counter-torque the unit will deliver against wave-induced roll, and where the design starts to fall apart at the edges of its operating range.
Given
- mrotor = 12 kg
- Drotor = 0.200 m
- ωspin nominal = 8000 RPM
- Ωgimbal (commanded precession rate to react against waves) = 0.5 rad/s
Solution
Step 1 — compute the rotor's moment of inertia, treating it as a solid disc (I = ½ × m × r2):
Step 2 — convert spin rate to rad/s and compute angular momentum at nominal 8,000 RPM:
Step 3 — output torque the gyro applies to the boat hull when the gimbal precesses at the commanded rate is τ = H × Ωgimbal:
That's a useful chunk of righting torque for a 6 m hull — not enough to flatten a yacht like a Seakeeper but enough to take the edge off short-period chop. Now the operating range. At the low end, drop spin to 4,000 RPM (half nominal) — perhaps the builder undervolts the motor for quieter dockside running:
The stabilising effect is roughly halved — passengers will still notice some damping but the boat will roll visibly in beam seas. At the high end, push the rotor to 12,000 RPM:
On paper this is a 50% gain in stabilising authority, but in practice you've now passed the safe rim speed for a hobby-grade steel rotor. At 12,000 RPM the rim is moving at 126 m/s — well into the regime where balance grade G1.0 and a containment housing become non-negotiable. The sweet spot for this build sits at 8,000-9,000 RPM, where the rotor delivers meaningful torque without crossing into industrial-balance territory.
Result
Nominal output torque is 25. 1 N·m at 8,000 RPM rotor spin and 0.5 rad/s gimbal precession. In hull terms, that's enough to noticeably damp a 6 m boat in 0.3 m chop — passengers feel the roll soften rather than disappear. The range tells the real story: 12.6 N·m at 4,000 RPM is barely worth the battery drain, 25 N·m at nominal is the design target, and 37.7 N·m at 12,000 RPM looks great until rotor balance and containment become safety issues. If the builder measures only 18 N·m on a hull-mounted load cell instead of the predicted 25, the usual suspects are: (1) gimbal-bearing static friction stealing torque before it reaches the hull — check for grease drag in the precession bearing, (2) rotor-balance imbalance creating a counter-precession that bleeds energy from the H vector, or (3) flexure of the gimbal yoke under load, which lets the spin axis tilt slightly toward the input torque direction instead of producing clean perpendicular reaction.
When to Use a Gyroscope and When Not To
A gyroscope isn't always the right answer for orientation sensing. The choice between a spinning-mass mechanical gyro, a fibre-optic or ring-laser gyro, and a MEMS gyro comes down to drift rate, cost, power, and how long the system has to coast without external aiding. Magnetometers and GPS heading sensors look attractive on cost but die in the wrong environment.
| Property | Spinning-mass gyroscope | MEMS rate gyroscope | Fibre-optic gyroscope (FOG) |
|---|---|---|---|
| Drift rate (typical) | 0.01-1°/hour (nav grade) | 10-300°/hour | 0.1-10°/hour |
| Unit cost | $5,000-$100,000+ | $1-$50 | $2,000-$30,000 |
| Spin-up / startup time | 3-10 minutes for rotor run-up | Milliseconds | Seconds |
| Moving parts | Yes — rotor, gimbals, bearings | Vibrating proof mass only, no rotation | None (light path only) |
| Service life | 5,000-20,000 hours bearing-limited | 10+ years, no wearing parts | 20+ years, optical-source limited |
| Power draw | 10-50 W | <5 mW | 1-10 W |
| Best application fit | Legacy aircraft AHRS, marine CMG stabilisers | Phones, drones, automotive ESC | Tactical missiles, AUVs, modern aircraft INS |
Frequently Asked Questions About Gyroscope
Datasheet drift figures are quoted at constant temperature with the bias term calibrated out. In a real build the dominant error you're seeing isn't long-term random walk — it's bias instability driven by self-heating and PCB temperature gradients. A 1°C change across the chip in the first minute after power-on can shift bias by 0.05-0.2°/s, which integrates into several degrees of heading error very fast.
Run a 30-60 second stationary calibration after power-on, average the bias, and subtract it. If the drift still won't settle, check that the IMU isn't mounted next to a switching regulator or a motor driver — magnetic and thermal coupling from those components is the usual culprit.
Reaction wheels produce torque by accelerating a flywheel — torque is limited by motor capability, typically 0.1-1 N·m. CMGs produce torque by precessing an already-spinning rotor, which gives massive torque amplification: a small gimbal motor controlling a high-H rotor can deliver 100+ N·m. So CMGs win wherever you need high agility on a large vehicle, like the ISS or an imaging satellite that has to slew quickly between targets.
The catch is the singularity problem. CMG arrays have geometric configurations where commanded torque can't be produced regardless of gimbal rate, and steering laws to avoid them are non-trivial. If your slew-rate requirements fit inside what reaction wheels can deliver, take the simpler hardware every time.
Counterintuitive but true. A gimballed gyro physically isolates the rotor from vehicle motion, so the rotor coasts in inertial space and bearing friction directly walks the spin axis off true. The drift you see is the integral of all parasitic torques over the mission.
A strapdown system bolts the gyro to the vehicle frame and uses a torque-rebalance loop to keep the rotor axis fixed relative to the housing. The rotor never actually drifts — the loop measures the current required to hold it in place, and that current is the angular rate signal. Modern electronics can null bias in this loop more accurately than a gimbal can mechanically follow, which is why every aircraft INS built since the 1980s is strapdown.
Nutation is the small high-frequency oscillation superimposed on steady precession. It shows up when you apply a torque suddenly rather than ramping it on, because the rotor's angular momentum vector overshoots the new equilibrium and rings down at the nutation frequency (which depends on the ratio of spin to transverse moments of inertia).
Two practical fixes: ramp the torque input over a time long compared to the nutation period, or add nutation damping — a fluid-filled damper ring on the gimbal, or in modern systems, an electronic damping term in the gimbal control loop. If your damper is sized correctly, nutation rings down in 2-3 cycles. If it never settles, the damper time constant is too long or the input torque is being applied as a square pulse.
You can spin the rotor with one, but expect short bearing life. The radial bearings in typical hobby outrunners are sized for a propeller load — a few hundred grams of axial-symmetric thrust — not for the gyroscopic loads a precessing rotor produces. When the gimbal precesses, the rotor bearings see a transverse moment proportional to H × Ωgimbal, and that's what kills them.
Build the rotor on its own pair of angular-contact bearings sized for the gyroscopic moment, and use the BLDC purely as a tangential drive through a belt or direct-drive stator wound around the rotor. Hobby-grade rotor builds typically get 200-500 hours before bearing noise and runout become unacceptable.
A gyrocompass finds true north by sensing Earth's rotation rate (15.04°/hour at the equator) projected onto the local horizontal plane. The signal it has to detect is tiny — fractions of a degree per hour — and it has to filter out vehicle motion, bearing friction torques, and thermal drift to extract it.
The settling time is the period the Schuler-tuned loop needs to damp out initial misalignment. Faster settling requires either a higher-quality gyro (lower noise floor, so the true-north signal stands out sooner) or external aiding from a coarse north reference. At high latitudes the horizontal component of Earth rate gets smaller and settling time stretches further — north of 80° latitude most gyrocompasses simply don't work, which is why polar surveys use FOG-based north seekers with active two-position carouseling.
References & Further Reading
- Wikipedia contributors. Gyroscope. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
