You’ve tuned P-gains down to conservative levels, enabled RPM filtering, and your quad still buzzes at a specific throttle point. The problem isn’t your tune — it’s your frame. Every carbon fiber frame has resonant frequencies, and when motor vibrations hit them, no amount of PID adjustment fixes it. Here’s how to find and kill them.
Step-by-Step: Frame Resonance Diagnosis and Fix
Step 1: Motor Test — Find the Problem RPM
Remove props. Connect to Betaflight Configurator. Go to the Motors tab, enable the motor test safety toggle, and spin each motor individually from 0 to 100% in 5% increments. Watch the gyro graph on the Setup tab simultaneously.
A clean motor produces a smooth gyro trace with a gradual noise increase proportional to RPM. A frame resonance appears as a sharp spike at a specific motor output percentage — noise jumps from 0.05 to 0.50 deg/s between 45% and 50% throttle, then drops back down at 55%. That’s your resonance frequency.
Note the motor output percentage where the spike occurs. Multiply by your motor’s KV and battery voltage to estimate RPM. On a 6S setup with 1750KV motors, 50% output is approximately (25.2V × 1750KV × 0.50) = 22,050 RPM ÷ 60 = 367Hz. But frame resonance frequencies for 5-inch quads typically land between 80-180Hz — harmonics, not the fundamental motor frequency, are what excite the frame.
Repeat the test for all four motors individually. If the resonance appears on only one motor at a specific RPM, the problem is that motor — a bent shaft or unbalanced bell. If it appears on all four at the same RPM, it’s the frame.
Step 2: Blackbox Spectrogram — Confirm in Flight
Props on. Fly a pack with blackbox logging enabled at 2kHz. Do a slow throttle sweep from hover to full throttle over 10-15 seconds — no aggressive moves, just a straight climb. Land.
Open the log in Betaflight Blackbox Explorer. Switch to the Spectrogram view. Look for horizontal bands — constant-frequency noise that persists across the throttle sweep. A frame resonance shows up as a bright horizontal line at a fixed frequency (e.g., 130Hz) that intensifies as RPM passes through the matching motor harmonic and fades as RPM moves past it.
This confirms the resonance exists in flight, not just on the bench. The spectrogram also tells you the exact frequency, which matters because different frequencies have different fixes.
Step 3: Physical Inspection — Find the Weak Point
Frame resonance at 80-120Hz is typically a loose arm. Remove the arm screws, inspect the carbon for delamination (looks like white-ish layers separating at the edge), re-tighten with fresh Loctite. If the arm shows delamination, replace it — the carbon layers are no longer bonded.
Resonance at 130-180Hz is usually the stack or camera mount. The flight controller stack acts as a cantilevered mass on standoffs. If the stack screws are tight but the standoffs have any play, the whole assembly vibrates. Add an O-ring between the standoff and the frame, and between the standoff and the nut. This isolates the stack mass from the frame.
Resonance at 50-80Hz is often the battery strap or battery itself. A 200g battery on a velcro strap is a mass-spring system. Strap it tighter, add a second strap, or use a non-slip battery pad.
Below 50Hz: check the arms for cracks at the root (where the arm meets the body). A hairline crack in 4mm carbon flexes under load and creates low-frequency oscillation that worsens under G-loading.
Step 4: Apply Targeted Fixes
Once you’ve identified the frequency and source:
Loose arm (80-120Hz): Tighten screws to 2Nm with a torque driver. Over-tightening crushes the carbon fibers at the bolt hole and weakens the arm. If the arm is delaminated at the bolt hole, it’s done — replace it.
Stack resonance (130-180Hz): Install soft-mount grommets under and over the FC. Add an M3 nylon washer between the nut and the grommet to prevent the nut from cutting into the rubber. If your stack has a vibration damper plate (some Hobbywing and T-Motor stacks include one), make sure it’s installed between the ESC and the frame.
Battery shake (50-80Hz): Double strap with both straps routed through the frame slots, not just around the top plate. A battery that can shift 1mm under G-loading is a 200g pendulum — it’ll shake the whole frame. Use a silicone battery pad with grip texture.
Arm crack (sub-50Hz): Replace the arm. No amount of epoxy fixes a structural crack in a loaded carbon arm. The crack propagates from the root outward; what you see is 30% of the actual damage.
Frame Resonance Frequency Guide Table
| Frequency Range | Typical Source | Diagnosis Method | Fix | Severity |
|---|---|---|---|---|
| 30–50 Hz | Cracked arm root, loose frame bolt | Visual + flex test | Replace arm, torque bolts | Critical — do not fly |
| 50–80 Hz | Battery shake, loose top plate | Tighten all screws, double strap | Second battery strap, silicone pad | Moderate |
| 80–120 Hz | Loose arm, delaminated carbon | Motor test per arm, visual | Tighten or replace arm | High — worsens over time |
| 130–180 Hz | Stack/camera mount resonance | Motor test, O-ring isolation | Add grommets, nylon washers | Moderate — tuning can mask |
| 180–250 Hz | Motor bell imbalance | Swap motor to different arm | Replace or balance motor | Moderate — specific RPM |
| 250+ Hz | Prop imbalance, bearing wear | Swap props, spin by hand | Balance/replace prop, oil bearings | Low — filters handle it |
What Most Pilots Get Wrong
Mistake 1: Filtering instead of fixing. When the spectrogram shows a clean 130Hz resonance line, the instinct is to add a notch filter at 130Hz. This works — temporarily. But notch filters add latency (2-4ms per filter), and the resonance will migrate as the frame fatigues. Six months later you’ll have three notches, 12ms of latency, and a quad that feels disconnected. Fix the mechanical source and you can remove the filter entirely.
Mistake 2: Testing with props on, on the bench. This is how you lose fingers. Never spin motors with props attached on the bench. The motor test in Betaflight is designed for props-off testing. If you need loaded data, fly it with blackbox — that’s what logging is for.
Mistake 3: Assuming carbon fiber doesn’t fatigue. Carbon fiber has excellent fatigue resistance compared to aluminum, but it’s not infinite. Every hard crash introduces micro-cracks at stress concentrations — bolt holes, arm roots, sharp inside corners. After 50-100 hard crashes, a frame arm can lose 30-40% of its stiffness at the root. You’ll notice it as a new oscillation that wasn’t there on the fresh build. That’s the frame telling you the arm needs replacement.
Mistake 4: Overtightening frame screws to “fix” resonance. Carbon fiber under compression delaminates at the bolt hole. A 2Nm torque is sufficient for M3 bolts into aluminum standoffs. Beyond 2.5Nm, you’re crushing the epoxy matrix that holds the carbon layers together. The arm fails at the bolt hole — the exact stress concentration you were trying to eliminate.
⚠️ Regulatory Notice: The diagnostic test flights described in this article should be conducted in accordance with 2026 drone regulations. Resonance testing often involves sustained throttle climbs that may exceed standard altitude limits. Verify local altitude restrictions and flight zone requirements with the FAA (US), EASA (EU), CAA (UK), CAAC (China), or your national aviation authority before performing test flights.
Resonance diagnosis works hand-in-hand with proper filtering. Our gyro filtering guide covers the software side — which filters to use and when. For the stack isolation techniques that prevent resonance from reaching the gyro in the first place, see our FC soft mounting guide. And if you’re seeing high-frequency noise above 200Hz, our RPM filtering guide covers the dynamic notch approach.
A frame designed with resonance in mind uses arm geometry that pushes natural frequencies above motor excitation ranges. The TBS Source One V5 frame uses 6mm arms with strategic cutouts to raise the first bending mode above 200Hz — well above typical motor-induced resonances. Available as a kit at uavmodel.com.