What Is Stick Drift and How to Fix It

Your character creeps left while your thumb is nowhere near the stick. The camera pans up on its own during a cutscene. That is stick drift, and you probably already know the symptom. This page is about the part that comes next: how to tell how bad it is, and which fix matches the number you get.

What the stick is telling the game

An analog stick is two sensors at right angles, one for left–right and one for up–down. Each reports a position from −1 to +1 with 0 in the middle. Drift means the sensor reports a non-zero value while the stick is physically resting at centre. The game cannot know your thumb is off the stick; it gets "a little to the left", so left you go. A game that ignores small inputs hides it, a game that trusts every unit shows it immediately. Same controller, different code.

Why potentiometer sticks wear out

Most first-party controllers still use potentiometer modules. iFixit's teardown of the Alps RKJXV series notes the same part family appears in PlayStation 3 through 5, Xbox One through Series X|S and Switch Pro controllers. Inside each of those modules sits a tiny variable resistor: a curved carbon track and a metal wiper that slides along it as you tilt the stick. The position of the wiper decides how much resistance the circuit sees, and that becomes the voltage the controller reads.

The catch is mechanical. The wiper spends most of its life parked at the centre of the track, and every return to centre is another rub on the same few millimetres of carbon. Over time the track thins unevenly, the contact gets noisier, and the "centre" voltage stops being the centre. Alps rates the potentiometer in its RKJXV datasheet at 2,000,000 direction cycles, which sounds like plenty until you count how many times a stick snaps back in one evening of a shooter.

Wear is not the only path, and iFixit's stick-drift page lists the rest: debris under the rubber boot, a spring that no longer pulls the shaft all the way home, a stretched plastic cage, and rarely a damaged trace on the board. Debris and spring problems give drift that changes from day to day. Track wear gives drift that only ever gets worse.

Magnetic sticks (Hall effect and TMR) have no track to wear, but they still drift for the mechanical reasons above; see Hall effect vs TMR vs potentiometer.

How much drift is normal

A brand-new potentiometer stick does not read a perfect 0,0 either; tolerance, spring preload and converter resolution add up to a small resting offset. Our stick drift test samples the stick every frame for 10 seconds with your hands off and reports the average offset (drift), the largest excursion (peak) and the wobble (noise), all in percent of full travel. So 5 % means the sensor thinks the stick is tilted one twentieth of the way to the edge.

Here is where we draw the lines, and why.

Under 2 %: normal. Tolerance plus electrical noise. No game reacts to it.

2 to 5 %: minor. You may see it in a menu or a very sensitive camera, but a 5 % deadzone hides it completely, and most games ship with more than that or let you set it. Microsoft's XInput documentation defines a default left-stick deadzone of 7,849 out of 32,767, roughly 24 % of travel; Unity's Input System defaults its stick deadzone minimum to 0.125, or 12.5 %. Competitive players turn these down, and 5 % is about as low as most settings menus go.

5 to 10 %: noticeable. Now the drift pokes through the deadzones people actually use. Sony's explanation of the DualSense Edge deadzone setting uses 10 % as its worked example, and that is about the top of a playable deadzone before aiming feels mushy. Here it is worth cleaning around the stick and recalibrating, because part of the offset may be debris or a stale calibration rather than wear.

Over 10 %: severe. A deadzone wide enough to hide this makes the stick unpleasant, and the trend only goes one way. If the test also shows high noise, the track is physically rough and no software value will smooth it. Replacement territory.

The percentages are ours, chosen from the deadzone ranges above and from how drift behaves over time. No manufacturer publishes a tolerance for a worn stick. If your game shows its own deadzone value, that is the real threshold for you.

Deadzone is a bandage, not a fix

A deadzone tells the game to treat any position inside a circle as exactly centre. Set it to 6 % and a 4 % drift vanishes. The stick has not moved; the game has agreed to stop listening to the first 6 % of travel.

The cost is precision: small movements now start 6 % further out, and some games do not rescale the rest, so you lose that slice off the top too. Our deadzone test shows raw and corrected position side by side so you can find the smallest value that covers your drift instead of guessing. Right call for minor drift and as a stopgap; wrong call when the number keeps climbing, because you will chase it with ever bigger circles.

When calibration helps and when it does not

Calibration means telling the controller or the software "this voltage is the new centre". A stick that reads 4 % left at rest gets 4 % subtracted from every reading, and rest becomes 0 again.

That works when the offset is stable: a settled spring, a module that left the factory slightly off, a stale calibration. Where it falls down is a worn track, because the problem there is not that the centre moved but that the reading is no longer repeatable. You calibrate today, the wiper finds a different patch of carbon tomorrow, and the drift is back with a different sign. Calibration also does nothing for noise, which is the sensor disagreeing with itself.

Where you calibrate depends on the controller: Windows has a wizard for DirectInput devices, Steam and DS4Windows have their own layers, the DualSense can be recentred in its own memory through a browser tool. Each is covered in How to calibrate a controller on Windows, Steam and DS4Windows. Consoles are stingier: the PS5 has no stick calibration for the standard DualSense, and Xbox added one to its Accessories app only in 2024; see PS5 and Xbox. The Switch has had a calibration screen in System Settings since launch, and the Steam Deck keeps its deadzone and recentring inside Steam itself, each with its own page: Switch and Joy-Con, Steam Deck.

Run the test again afterwards. Under 2 % and staying there: done. Under 2 % today and 6 % next week: the clearest sign you can get that the track is worn.

Cleaning: what it can and cannot reach

Dust and skin oil work their way under the rubber boot at the base of the stick. Debris holding the shaft a fraction off centre looks exactly like wear in a single test but behaves differently over time: it changes direction, and sometimes clears after a few hard rotations.

Cleaning without opening the controller means lifting the boot gently with the stick tilted and clearing the gap with a short burst of compressed air or a swab barely dampened with isopropyl alcohol, controller off, everything dry before you power it up again; iFixit's DualSense stick-drift page describes the same routine. Do not pour anything in: alcohol that reaches the track can strip its lubricant and make wear worse. If the number drops and stays down, debris was the cause. If it helps for a day, you were cleaning around a worn track.

When the only fix is a new module

Three things in your test history point to replacement rather than another round of cleaning: drift above 10 % that survived a clean and a recalibration; drift that returns within days of a successful recalibration, especially with a different sign; and high noise at rest, where the dot in the plot vibrates instead of sitting still off-centre.

On almost every first-party controller, replacing a module means desoldering the old one and soldering in a new one. iFixit rates the DualSense job as difficult, 1 to 4 hours, 14 solder joints per module, old module destroyed in the process. The DualSense Edge is the exception: its modules slide out on a latch in minutes. Replacements come as another potentiometer (cheapest, same wear curve), Hall effect or TMR, and the magnetic two remove the wear path entirely; the trade-offs are in the sensor type guide, the swap itself on the PS5 and Xbox pages.

If you will not solder and the warranty is gone, weigh a repair shop against a new controller; we are not going to tell you which pad to buy. Any potentiometer stick starts the same clock again, and the test shows where it is on that clock whenever you want to look.

Questions people ask

Can stick drift be fixed, or is it permanent? Debris and a stale calibration are fixable and the fix lasts. Worn carbon is permanent; the only fix for the stick itself is a new module, and a deadzone merely hides it.

Why does my drift show in one game and not another? Each game applies its own deadzone first. A 20 % deadzone never shows a 6 % drift; a 3 % deadzone shows it constantly.

Is drift the same as the stick not returning to centre? They look alike in one reading. A tired spring leaves the shaft off-centre and a healthy sensor reports that honestly; a worn sensor misreports a centred shaft. Nudge the stick into centre during the test: if the reading snaps to zero while you hold it there, the spring is the problem.