Two readings and an exponent
Thrust and power on a propeller are related by a power law: double the thrust and the power goes up by more than double. Momentum theory for an ideal rotor gives an exponent of 1.5, and real motor and propeller combinations land near it without landing on it. Fitting your own is one division of two logarithms.
The defaults — 400 g at 2.9 A and 900 g at 9.4 A on 22.2 V, so 64.4 W and 208.7 W — give an exponent of 1.4570. A 1450 g quad needs 362.5 g from each of four motors, which lands at 57.0 W a motor. At 21.6 V in flight that is 2.64 A each and 10.6 A from the pack.
Efficiency falls the whole way up
The same two readings say 6.21 grams per watt at the lower point and 4.31 at the upper. That is not a fault in the motor, it is what propellers do: thrust goes as the square of the airspeed through the disc and power as the cube, so the harder you push a given propeller the less you get per watt. The hover, at 362.5 g a motor, comes in at 6.36 g per watt.
This is the arithmetic behind slow-flying efficiency and behind large slow propellers on endurance machines. It is also why a heavier aircraft costs more than proportionally: the extra weight is bought at a worse exchange rate than the weight already there.
The payload trap
Put 300 g of gimbal on the default machine and the pack current goes from 10.6 A to 14.0 A. That is a 21 percent weight increase costing 32 percent more current, and the endurance falls accordingly on top of whatever the extra pack weight already cost. The page prints both percentages side by side because the gap between them is the entire point.
The same sum kills the idea that a bigger pack always means longer flights. The extra capacity arrives attached to extra grams, those grams are charged at the same steepening rate, and somewhere there is a crossover after which a bigger pack flies for less time than a smaller one.
Straddle the hover
A power law fitted to two points is trustworthy between them and speculative outside them. If the hover sits at 362 g a motor, take the readings at something like 250 and 550 rather than at 400 and 900 — the page says which case you are in and will tell you when it is extrapolating.
Two points also cannot see the shape of the real curve, which is not a single power law from zero throttle to full. The motor has an efficiency peak, the propeller has another, and they are in different places. Four readings plotted out show that immediately. Two readings give you a usable number near where they were taken and nothing more, which is honest enough as long as it is understood.
Questions people ask
How do I work out hover current without flying?
Measure thrust and current on a stand at two throttle settings, fit the exponent between them, and read off the power at the thrust one motor has to make to hold the aircraft up — all-up weight divided by the motor count. At the defaults that is 362.5 g a motor, 57.0 W, and 10.6 A from the pack across four motors. It is the figure the endurance calculators ask for and cannot derive themselves.
What exponent should the power curve have?
Momentum theory for an ideal rotor says 1.5, and a real motor and propeller land near it — the defaults here fit 1.4570. Much above 2 usually means the upper reading was taken with a sagging pack or a hot motor. Much below 1.2 usually means the two points are too close together and noise is dominating. The page fits yours rather than assuming a number, and prints the ideal beside it for comparison.
Why does adding a small payload cost so much current?
Because power rises faster than thrust. On the fitted curve here, adding 300 g to a 1450 g machine — a 21 percent weight increase — costs 32 percent more current. Every gram is bought at a worse exchange rate than the gram before it, which is the same reason a bigger battery stops paying for itself past a certain size.
Should I measure current at the pack or at the motor?
At the pack is simpler and more useful, because it already includes the speed controller losses that a motor-side measurement misses. If you did measure at the motor, put a loss percentage in the field for it. Enter zero when the clamp was on the pack lead — otherwise the losses are counted twice.
How far apart should the two bench readings be?
Far enough apart that the difference is much larger than the noise in your stand, and positioned so the hover thrust falls between them. Readings at 250 and 550 g bracket a hover at 362 g and give an interpolation; readings at 400 and 900 give an interpolation too but weight the fit towards throttle settings you rarely use. The page states which case you are in.