Motor thrust is a structural load, not only a performance number

The thrust number in your propulsion spreadsheet answers one question: can this aircraft hover with margin. The same number in your FEA setup has to answer a different one: what does the motor do to the mount it’s bolted to. Treating the two as the same can lead to sizing a motor mount for a load that isn’t the one that governs.

The motor is the Holybro 2216 920KV that ships with the X500 V2, on the stock 1045 prop. Holybro doesn’t publish a thrust table for it, so the bench data below is from the datasheet of the OEM-equivalent LigPower AIR2216II, which is the same motor under a different label. The datasheet states that its numbers are for reference only.

What the spreadsheet gives you

A static thrust figure at a throttle setting and a voltage. From the LigPower table at 16 V (4S at 4.0 V per cell), on the T1045 prop:

Three corrections before it goes near a structural model.

Voltage. The table is at 16 V. A freshly charged 4S pack sits at 16.8 V, so full-throttle thrust on a fresh pack is above 1,332 g.

Density. Static thrust drops with altitude and temperature. For structure you want the sea-level cold-day number, which is the highest.

Transients. A step throttle input can briefly overshoot steady thrust while the prop spins up. The manned-aircraft codes handle this with a factor of safety on the maximum operational load rather than by modeling the transient. NATO AEP-83, the Light Unmanned Aircraft Systems Airworthiness Requirements published under STANAG 4703, asks for a limit factor of at least 1.0 on maximum operational loads and an ultimate factor of at least 1.5 on structure whose failure would be hazardous (UL.2). We apply the 1.5 at the end.

The number that leaves this section is 13.1 N per motor, the highest steady thrust the motor makes at the datasheet voltage.

What the mount sees

Four things arrive at the bolt pattern.

Thrust, T. Along the motor axis, 13.1 N, pulling the motor up and away from the mount. The four bolts share it in tension, 3.3 N each.

Torque reaction, Q. The motor spins the prop one way and the stator pushes the mount the other. Q = P / ω. The datasheet gives 260 W electrical at 9,857 rpm (1,032 rad/s), and shaft power is lower by the motor efficiency, so 260 / 1032 = 0.25 N·m is an upper bound.

Gyroscopic moment, M_g. From this post, 0.25 N·m is when the aircraft pitches at 200 deg/s, using a rod approximation for the 13 g prop’s inertia. This acts about an axis in the plane of the mount, perpendicular to both the spin axis and the pitch axis, so it tries to tip the motor off its mount.

In-plane hub force. In forward flight the advancing and retreating blades see different airspeeds and the prop makes a small sideways force at the hub. On a 10 inch prop at 15 m/s it’s a few percent of thrust. The 2216 sits low on the X500’s arm-end mount, so the tipping moment from it is small.

Worked example: bolt loads on the 2216

The Holybro 2216 has a four-bolt pattern on a 16 mm × 16 mm square, M3. Bolt radius from the motor axis: r = √(8² + 8²) = 11.3 mm.

Gyroscopic moment: take the worst orientation, M_g about an axis parallel to one side of the square. The two bolts on one side go into tension and the two on the other side press the motor base into the mount. Treating the far edge of the pattern as the pivot, the two tension bolts sit 16 mm from it, so 2 × F × 0.016 = 0.25 and F = 7.8 N per bolt. Treating it as a pure couple with the bolts at ±8 mm gives the same 7.8 N.

Adding the tension components: 3.3 N from thrust plus 7.8 N from the gyroscopic moment, about 11 N on each of the two loaded bolts, with 5.5 N shear from torque on all four.

So on this motor the moment puts more than twice as much tension into a bolt as the thrust does. Thrust alone is 3.3 N per bolt. One hard pitch input puts 11 N into two of them, and it swaps between the pairs as the aircraft rolls and pitches, which makes it a fatigue problem as much as a static one. An M3 bolt has ample margin at 11 N. The material it threads into may not. On the X500 the motor bolts into a fiber-reinforced nylon arm-end mount, and a nylon thread under 11 N of alternating tension is a different question from a steel bolt under the same load.

Putting it into the model

The setup that produces the most misleading result is 13 N as a point load at the center of the mount. A point load on a shell or solid mesh is a stress singularity: the peak grows without bound as you refine the mesh, and it grows at a spot (the center) where no load actually enters the part. Here is how to avoid it.

Apply the loads where they physically enter the part, which is the bolt holes and the motor’s bearing face. Two ways to do that in CalculiX (keywords from the 2.22 manual).

Option A, direct bolt loads. Apply the per-bolt tension and shear from the hand calc to the cylindrical faces of the four holes, plus a matching compressive pressure on the contact patch under the motor base for the moment couple. More setup work, but every number traces back to the hand calc.

Option B, reference point coupling. Put a reference node at the motor’s center of mass, above the mount, and tie it to the four bolt hole surfaces with *COUPLING followed by *DISTRIBUTING (not *KINEMATIC, which would make the holes rigid and put artificial stiffness into the mount). Apply T, Q, and M_g at the reference node. The coupling spreads them to the holes in a statically equivalent way and the solver computes the individual bolt loads. Faster, and it gets the moment right automatically, but the stress right at the coupled holes isn’t meaningful; read it a few elements away. (CalculiX also has a separate element-based *DISTRIBUTING COUPLING keyword; it’s a different mechanism from the *COUPLING pair and easy to confuse.)

Either way the reaction at the bolt holes in the results should match the hand calc to within the coupling’s distribution error. If it doesn’t, one of them is wrong.

Combinations to run

Thrust alone is the most commonly run case and the least informative. Two that matter more.

Max thrust plus gyroscopic moment plus torque and every aggressive maneuver at full power. There are the peak static cases for the mount.

Hover thrust plus a reversing gyroscopic moment. The moment alternates every time the aircraft changes attitude, thousands of times per flight, at a load level nowhere near static failure.

What to do with this

Take the highest steady thrust from your bench data, not hover. Get torque from power and speed, and say whether the power is electrical or shaft. Get the gyroscopic moment from prop inertia and your controller’s maximum attitude rate. Resolve all three onto the bolt pattern by hand before you open the solver so you have something to check the model against. Then apply the loads at the holes, never at the center.

Sources

1. Holybro X500 V2 documentation and 2216 920KV motor listing, for the bolt pattern and mass.

   https://docs.holybro.com/drone-development-kit/px4-development-kit-x500v2

2. LigPower AIR2216II KV920 datasheet with T1045 thrust table, the OEM equivalent.

   https://cdn.robotshop.com/media/T/Tmo/RB-Tmo-266/pdf/ligpower_airgear_450ii_combo_set_multi_rotor_uav_p_datasheet.pdf

3. NATO AEP-83 (STANAG 4703), Light Unmanned Aircraft Systems Airworthiness Requirements, Edition A, 2014. UL.2 for the limit and ultimate factors, UL.5 for including propulsion loads in flight load cases.

   https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/391827/20140916-STANAG-4703_AEP-83_A__1_.pdf

4. CalculiX CrunchiX User’s Manual, version 2.22 (Dhondt, 2024), keyword entries for *COUPLING, *KINEMATIC, *DISTRIBUTING, and *DISTRIBUTING COUPLING.

   http://www.dhondt.de/ccx_2.22.pdf