A gimbal rejects low-frequency motion. It cannot reject high-frequency vibration, because its own control loop is not fast enough and the energy arrives through the structure rather than as a measurable attitude change. Everything above roughly 40 Hz is the mounting system’s problem, and a good gimbal on a bad mount produces bad images.
Key takeaways
- Vibration energy arrives in distinct frequency bands, and each band is handled by a different part of the system — isolators, gimbal servos, or nothing at all.
- Low-frequency airframe motion is the gimbal’s job; high-frequency rotor energy must be killed mechanically before it reaches the sensor.
- Isolators are chosen from payload mass and target natural frequency, not by feel — an under-loaded mount amplifies the band it was meant to absorb.
- Commission every build the same way, because identical aircraft with different payload masses do not share a damping solution.
On this page
Where the energy comes from
Three sources dominate on a multirotor. Propeller blade-pass frequency is rotor RPM multiplied by blade count — a two-blade prop at 6,000 RPM produces 200 Hz. Motor electrical and bearing frequencies sit higher. Airframe structural resonance sits lower, typically 15–60 Hz, and is the most dangerous because it can coincide with the gimbal control bandwidth.
On fixed-wing and VTOL the picture changes: engine or pusher-prop harmonics, plus aerodynamic buffet in the 5–30 Hz band that behaves more like turbulence than vibration. The isolation strategy that works on a quadcopter is often wrong on a VTOL.
Frequency bands and what handles them
| Band | Typical source | Handled by | Symptom if unhandled |
|---|---|---|---|
| 0–5 Hz | Aircraft manoeuvre, wind gust | Gimbal control loop | Horizon drift, slow wander |
| 5–20 Hz | Airframe flexure, buffet | Gimbal loop + isolator | Low-frequency image wobble |
| 20–60 Hz | Structural resonance | Isolator design (critical band) | Amplified motion, jello |
| 60–200 Hz | Blade pass, motor | Isolator + payload stiffness | Jello, rolling-shutter skew |
| 200–800 Hz | Motor electrical, bearings | Material damping, mass | Fine image softness |
| > 800 Hz | Acoustic, prop tip | Generally benign | Rarely visible |
Choosing isolators without guessing
A vibration isolator is a mass-spring system with its own natural frequency. Below that frequency it transmits everything; near it, it amplifies; above roughly 1.4 times it, it attenuates. The design rule is therefore simple to state and easy to get wrong: place the isolator natural frequency well below the lowest disturbance you need to reject, and well above nothing you care about.
For a typical multirotor with 150–250 Hz blade pass, an isolator natural frequency around 15–25 Hz works. Getting there requires matching isolator stiffness to the actual suspended mass — the same rubber mounts that work perfectly with a 400 g payload will be badly under-loaded with a 120 g one and over-loaded with a 1.5 kg one.
This is the most common integration error we see: correct isolators, wrong payload mass, resulting resonance sitting right on the blade-pass frequency and amplifying it.
Symptoms and their causes
- Jello (wavy vertical distortion). Rolling-shutter sensor sampling rows while the payload vibrates. Almost always an isolation problem, not a gimbal problem.
- Horizon tilt that appears only in flight. Vibration rectification biasing the accelerometer. The fusion filter believes gravity has moved.
- Sharp images at hover, soft in forward flight. Aerodynamic buffet in a band the isolator does not cover.
- Worse at one specific throttle setting. A resonance being excited at that RPM. Change prop, mass, or isolator stiffness — not the gimbal.
- Loose or aged isolators. Elastomer stiffens with age and cold. An installation that was correct two winters ago may not be now.
How this shows up in our payloads
Our stabilised gimbals such as AX-20D use a three-axis nonorthogonal mechanical structure that keeps the suspended mass compact and the structural modes high, which widens the usable gap between isolator resonance and disturbance frequencies. Integrated micro pods such as MV-4P bring their own damped mounting interface so the integrator is matching a known mass rather than guessing.
Related reading
Technology: payload mechanical integration and IMU, AHRS and attitude fusion.
Field practice: gimbal troubleshooting, payload integration checklist and gimbals on fixed-wing and VTOL.
A repeatable commissioning procedure
Vibration problems are diagnosable in about twenty minutes if you approach them systematically, and can consume weeks if you change one thing at a time by intuition. The procedure below is what we recommend to integrators commissioning a new airframe and payload combination.
The principle is to identify the frequency before changing the hardware. Almost every unsuccessful vibration fix we see is a correct remedy applied to the wrong band — new isolators fitted when the real problem was a bent propeller, or propeller balancing done when the real problem was an under-loaded mount resonating.
Record the results. A commissioned configuration that is written down can be reproduced across a fleet and audited after a change; one that lives in an engineer’s memory cannot.
- Bench test first: run the aircraft restrained at several throttle settings and record video at full resolution. Jello that appears at a specific RPM identifies the frequency band immediately.
- Check the mechanical basics before anything else: propeller balance, motor bearing condition, loose arms, cable strain on the payload.
- Confirm the isolator is loaded correctly — measure static sag and compare with the manufacturer’s design point for that payload mass.
- Change one variable at a time and re-run the same test at the same throttle settings.
- Test in forward flight as well as hover; buffet occupies a different band from propeller vibration.
- Document the final configuration: isolator part number, payload mass, propeller type and the throttle range tested.
FAQ
My footage has wavy distortion. Is the gimbal faulty?
Usually not. Wavy vertical distortion — jello — is a rolling-shutter sensor sampling image rows while the payload vibrates at high frequency. Gimbals cannot reject that band. Check isolator condition, isolator stiffness against actual payload mass, propeller balance and any hard contact between payload and airframe.
Can I use the isolators that came with a different payload?
Only if the suspended mass is similar. Isolator natural frequency depends on stiffness and mass together; reusing mounts sized for a heavier payload puts the resonance in the wrong place and can amplify vibration rather than attenuate it.
Why is the image fine in hover but soft in forward flight?
Forward flight adds aerodynamic buffet at lower frequencies than propeller vibration, often in a band the isolator does not attenuate. It also changes the rotor loading and therefore the blade-pass frequency. Test at cruise speed, not only in hover.