Hover throttle is the number that decides everything

The number nobody checks

Ask a pilot what their PID values are and they will tell you to three digits. Ask what throttle they hover at and you will usually get a shrug, or a guess, or "about the middle, I suppose."

That is the wrong way round. The PID numbers are downstream of a physical fact about the aircraft, and hover throttle is the cheapest available proxy for that fact. It costs you nothing to read — you already have the information, on your OSD, in every log you have ever recorded — and it tells you more about whether the aircraft is any good than any gain in the CLI.

Almost every difficult problem on a marginal aircraft traces back to it. This article is an attempt to show you why.

What hover throttle tells you

Roughly speaking: if the aircraft holds a steady hover at about half throttle stick, you have somewhere near twice the thrust you need to stay airborne — a thrust-to-weight of about 2:1 — and something like half your control authority still in reserve. If it hovers at 75%, you have very little left in the tank.

Now the honest caveat, and it matters. Throttle stick position does not map linearly to thrust. Thrust rises roughly with the square of RPM, throttle maps to RPM through the ESC and the motor's loaded response, the battery sags as current climbs, and firmware applies its own throttle curve on top. A 50% stick does not produce exactly 50% of maximum thrust, and anyone who tells you it does is selling you a tidy story.

So treat hover throttle as an indicator, not a measurement. It is a rule of thumb of the same species as "if you can't hold the motor bell, you have a problem" — imprecise, directionally reliable, and vastly better than not looking. A build that hovers at 40% is in a fundamentally different regime from one that hovers at 70%, and you do not need a thrust bench to know which one you are flying. If you want the actual number, you measure it, and there is a section at the end on how.

Why control authority is the real currency

Here is the mechanism, because once you see it the rest of the article is obvious.

A flight controller flies the aircraft by making the motors differ from each other. That is its only lever. To roll right it speeds up the left-hand motors and slows the right-hand ones; the thrust imbalance produces a moment, and the aircraft rotates. Every correction the PID loop makes — every gust rejected, every stick input honoured, every wobble damped — is spent as a difference between motors, on top of whatever throttle is holding the aircraft up.

That difference needs headroom. If the motors are already sitting near maximum just to hover, there is nowhere to go up. The controller asks for more thrust on one side, the motor is already flat out, and nothing happens. The only remaining way to produce the moment is to pull the other side down — which works, sort of, but it means the total thrust falls and the aircraft loses altitude every time it corrects. Push it further and the correction it needs is simply not available in either direction.

That is motor saturation, and it is the single most important failure mode in this article. An under-powered aircraft is not merely slow. It is untunable, because the controller's commands stop producing the responses the controller expects. And past a certain point it is uncontrollable, because there is no longer enough authority left to hold attitude at all.

No PID value creates thrust the motors do not have. You cannot gain your way out of it, you cannot filter your way out of it, and every hour you spend trying is an hour of your life.

The consequences, one by one

It cannot be tuned. A saturated motor invalidates the log it appears in. The setpoint says one thing, the motor output is pinned at its ceiling, and the gyro trace that results tells you nothing about your gains — it tells you about your ceiling. Tune from that and you will chase phantom oscillations, add filtering you do not need, and lower gains that were never the problem. Check motor outputs for clipping before you interpret anything else; see reading Blackbox logs and, for what the terms are actually supposed to be doing, PID in plain English.

The motors cook. High hover throttle means high continuous current, and copper loss goes as the square of current. Hovering at 70% rather than 45% is not 50% more heat, it is a great deal more, and the motor has no rest between corrections. This is why hot motors so often turn out to be a hardware arithmetic problem wearing a tuning costume.

The pack sags harder and dies sooner. The same current that heats the windings runs through the pack's internal resistance, and the voltage you lose is I × R. A high-throttle build sags more, triggers its warnings earlier, gets less usable capacity out of every cell, and ages its packs faster. If your voltage falls off a cliff on every punch, start here — but check your hover throttle first, because it may be the cause of both.

There is no redundancy. This one is widely misunderstood. A hexacopter at 80% hover throttle cannot absorb a motor failure, and the extra arms do not save it — losing a motor means the survivors must produce the same total thrust from fewer of them, and if they were already near their limit, they cannot. Redundancy is not a property of arm count. It is a property of margin. See motor failure and redundancy.

Descents get dangerous. An aircraft with no margin cannot arrest a sink rate. Come down too fast and it settles into its own downwash — descending through disturbed, already-accelerated air, where the props produce less thrust exactly when you need more of it — and a marginal aircraft has nothing left to break the fall with. Some of what gets blamed on propwash is really this.

Failsafe and RTH become unreliable. Every automatic mode assumes the aircraft can climb to its return altitude, hold position against wind, and land under control. All three are authority. A build with no margin may not be able to do any of them, and will discover this at the worst possible moment. How failsafe works is worth reading with this in mind.

What good looks like

As a general target for a stable, controllable aircraft: roughly 2:1 thrust-to-weight, which puts your hover at or a little below half throttle.

That is a rule of thumb, not a law, and the sensible range is wide. Racers and freestyle quads run enormously more — 8:1 or beyond — because they are built for acceleration and the margin is the entire point. Heavy working aircraft routinely run much less, and a 5 kg platform that achieves a genuine 2:1 fully loaded is doing well; plenty of cinema and survey machines fly at less than that and are flown accordingly, by pilots who know it.

The condition that gets skipped: the target must be met at the weight you actually fly. Not the bare frame. With the payload fitted, the real battery you use, the mount, the ballast, the lens you swapped in. A build that is 2.2:1 on paper and 1.4:1 with the camera on it is a 1.4:1 aircraft, and it will behave like one.

Why it gets worse than you planned

Margin is not a constant. It erodes, and it erodes in the direction you would least like.

Payload creeps. Nobody ever removes anything. A GPS here, a bigger pack there, a second camera, and the aircraft that hovered at 45% on its maiden flight hovers at 60% a season later, and nobody noticed the day it crossed over.

Battery sag shrinks the margin as you fly. Lower state of charge means lower open-circuit voltage, lower voltage means less RPM for the same throttle, and less RPM means less thrust. Your throttle margin is therefore smallest at the end of the flight — which is precisely when you are furthest away, lowest on options and most tired. The aircraft you took off in is not the aircraft you land.

Heat and altitude take their cut. Thinner air means less thrust for the same RPM. A hot summer afternoon at a thousand metres is a materially different aircraft from a cold morning at sea level, and the difference is not small enough to ignore on a marginal build.

And any hardware change resets the sum. A heavier prop, a slightly different motor, a new ESC with a different limit — the whole calculation moves. The margin you measured on a fresh pack, at sea level, in winter, on the bare frame is not the margin you have in August at the end of a flight with the payload on.

How to actually measure it

In the air, informally: fly a stable, steady hover on a fresh-ish pack at your real all-up weight, and read the throttle. Not a guess afterwards — read it, or pull it from the log. Do it with the payload fitted and with the battery you actually use. It is a two-minute job and it is the most informative two minutes available to you.

Properly, statically: put the motor and prop on a thrust bench at the voltage you fly, and compare total thrust across all motors against the aircraft's all-up weight. That gives you a real thrust-to-weight rather than a proxy. Failing a bench, use the manufacturer's thrust table for your exact motor, prop and cell count.

Treat those manufacturer figures with suspicion. They are measured in ideal conditions — a bench with clean airflow, a fresh, stiff pack, no frame in the wash, often the most flattering prop in the range — and they are marketing as much as engineering. Discount them. If a build only reaches your target margin by trusting a manufacturer's best-case number exactly, it does not reach your target margin.

The number matters more than the precision of the number. An aircraft that hovers somewhere around 40% is a healthy machine you can tune. One that hovers somewhere around 75% is a hardware problem, and the only fixes are the expensive, physical ones: less prop pitch and more diameter, more motor, more voltage, or less aircraft. Find out which one you have before you open the tuning tab.