Size: the prop decides
Frame size is the diagonal distance between motor centres, in millimetres. In practice frames are named by the largest prop they take, so a "5-inch frame" is built for 5" props. Oscar Liang's frame guide gives these rough ranges:
| Diagonal motor distance | Max prop size |
|---|---|
| 280 mm and more | 7" |
| 220-250 mm | 6" |
| 180-220 mm | 5" |
| 150-180 mm | 4" |
| 120-150 mm | 3" |
| 90-120 mm | 2" |
He also suggests running the largest prop the frame supports. The motors sit at the arm tips, and the further they are from the centre, the more the quad resists changes in rotation. 4" props on a 5" frame will fly worse than the same props on a 4" frame.
Arm layout
Seen from above, the motors of almost every modern quad form a rectangle or a trapezoid. The common layouts:
- True X. All arms meet in the middle and every motor is the same distance from the centre. Pitch and roll respond the same way, which makes the tune symmetric and predictable. FPV Handbook calls it the safe default, and Oscar Liang notes it became the standard for 5" frames.
- Stretched X. The front and rear motors are further apart than the left and right ones. The idea is that the rear props get less disturbed air from the front props. FPV Handbook calls it a racing favourite that steadies fast straight flight, but warns that the uneven inertia couples roll into pitch during hard rolls. It lists picking stretched X for a first build "because racers use it" as a common mistake.
- Squashed X. The opposite: wider than it is long. Its main practical advantage is that the front props are easier to keep out of the camera view when you fly with a low camera angle.
- Deadcat. The front arms are spread wider than the rear ones, so the props stay out of the camera's view. That makes it a common choice for cinematic builds.
Oscar's example is the GEPRC Vapor: the D5 version has a wide X layout that keeps the props out of the camera's view, while the X5 is a true X, the symmetric layout preferred for freestyle.
What deadcat costs
An asymmetric layout means roll moves can come with a bit of yaw mixed in. According to Oscar Liang, the flight controller's mixer usually compensates, though not always perfectly. For smooth cinematic flying it's negligible, and he doesn't think most freestyle pilots will notice either. The second cost is physical: the protruding front puts the camera in the line of impact more often.
Betaflight does let you describe a non-standard motor layout. In custom mixing (mixer custom and one mmix line per motor) each motor gets its own share of throttle, roll, pitch and yaw. The documentation's example of a U-shaped hexacopter gives the two motors closer to the roll axis half the roll authority of the outer ones. For a typical deadcat you don't have to touch this, but it's good to know where that setting lives.
How much the layout really matters
Less than the arguments online suggest. In a Drones Stack Exchange thread, a frame designer and manufacturer estimated that the frame contributes around 10% of the final flight feel. He said changing motor KV by 10% and swapping props changes far more than changing the arm ratio, and recommended that new pilots start on a true X. Another user who logged the same move on pitch and roll on a stretched X couldn't find a difference in the logs. Treat these as experienced opinions, not measurements. They do point the same way: weight distribution, the drivetrain and the tune decide more than the shape seen from above.
For a first freestyle build, a true X gives you a reference point. If clean HD footage without props is your main goal, look at deadcat or a wide X.
Separate arms or a one-piece bottom plate
Some frames cut the arms and bottom plate from one carbon sheet (a unibody design). It is lighter and quicker to assemble, but a broken arm means replacing the whole plate and moving every component onto it. With separate arms you replace only the broken one. FPV Handbook calls this the first thing worth more than the marketing: you will break arms, and one arm costs a fraction of a whole plate.
Oscar Liang adds that, in his experience, frames with separate arms are usually stiffer than unibody designs, because the arms are sandwiched between two plates. Each joint between an arm and a plate also dissipates a little vibration energy as friction, which reduces how much reaches the flight controller.
Arm and plate thickness
Arms take most of the strain in a crash, so that's where thickness matters. Oscar Liang's guidance:
- 5": at least 5 mm, and 6 mm arms are increasingly common because of stronger motors and higher-speed impacts.
- Light 3" and 4": 4 mm or even 3 mm.
- 2": around 2.5 mm, possibly 2 mm.
Top, centre and bottom plates are commonly cut from thinner 2-3 mm carbon. Thicker carbon is stronger and stiffer, but heavier.
Shape matters too. Sharp inside corners on an arm are stress points that break first, while curves and round cut-outs hold up better. The direction of the cut also counts: carbon plate is stronger when cut along the weave than at 45 degrees to it. You can't judge carbon quality from photos, so reviews and the brand's reputation are your best guide.
Stiffness and why a 3D-printed frame is a bad idea
A frame that flexes in flight feeds vibrations to the flight controller. Oscar Liang lists the results: a quad that's hard to tune, jello in the FPV and HD footage, and in extreme cases overheating motors. FPV Handbook gives a way to tell frame resonance apart in a blackbox spectrogram: frame or arm resonance shows up as a horizontal line at a fixed frequency that doesn't change with throttle, while motor and prop imbalance shows as a diagonal ridge that climbs with throttle.
For the same reason Oscar advises against printed and plastic frames. They lack stiffness, so the frame's resonant frequencies drop into the range of real flight movement, and the filters can no longer separate noise from motion.
Weight
Oscar Liang quotes 60-90 g as typical for 5" race frames and 90-120 g for freestyle frames. FPV Handbook gives 110-150 g for a 5" freestyle frame, so treat both as rough guides rather than targets. Weight on its own doesn't make a frame tougher. A heavier quad carries more momentum into the same crash.
Check that your parts fit
Before buying, check:
- Stack mounting holes: 30.5×30.5 mm and 20×20 mm are standard, and smaller frames also use 25.5×25.5 mm and 16×16 mm.
- FPV camera size: micro or nano.
- Camera angle range: Oscar suggests 5-10 degrees to start, says 20-30 degrees is popular for freestyle, and that some racers go up to 60 degrees. A frame that lets you set anything from 0 to 45 degrees is a good choice.
- Mounts for your digital video system and for an HD camera if you fly one: some frames include printed TPU parts, others leave you to buy or print them.
- Room for the battery and a soft-mounted stack.
- Spare parts: if a frame has no spare arms or plates, a single crash can mean buying the whole frame again.
Hardware and finishing touches
Screws also add weight and cost. Steel is the cheapest and strongest but heaviest. Aluminium is the lightest, but soft: it bends in a crash and strips easily when overtightened. Titanium is light and strong, but expensive.
Carbon conducts electricity and blocks radio signals. A wire chafing against a sharp arm edge can short, and antennas should sit outside the frame. If the frame's edges aren't chamfered, Oscar suggests filing them yourself, under running water so you don't breathe carbon dust, and sealing them with CA glue to reduce delamination in hard crashes.
Battery position is a matter of style. For freestyle, Oscar mounts the battery on top: the centre of mass sits closer to the plane of the props, and you don't land on the pack. Race frames usually carry it underneath because there's no room on top.