Controller Circularity and Stick Gate Test
Stick drift tells you what your analog stick does when you leave it alone. Circularity tells you what it does when you push it all the way out and rotate it around the edge. Those are two completely different failures, and this test covers the second one. Roll the stick around its outer rim and the tool records how far it actually reached in every direction, then draws the real shape your controller is sending to the game. A healthy stick traces something close to a round ring. A worn or badly calibrated one traces a square, a lopsided oval, or a ring with a bite taken out of it, and that missing travel is exactly why your aim slows down at certain angles or your car refuses to hold a line through a long corner.
Run a Circularity Pass
Setup
Live trace
The thin outer ring is full travel at 1.00. The dashed ring is your outer travel threshold. The solid black outline is the shape your stick actually reached, rebuilt from the furthest point recorded in every direction.
Results
Reach by direction
Eight directions get measured separately, so a stick that is fine everywhere except one corner still gets caught.
Report
How to Run It Properly
- Plug the controller in with a cable if you can. Wireless works, but a wired pass gives you a cleaner baseline to compare against later.
- Press any button on the pad. Browsers hide gamepads until they see one input, so nothing appears until you do this.
- Pick the stick you want to check. Test one at a time. The left and right sticks wear at completely different rates because they get used for completely different things.
- Press Start, then push the stick hard against the outer edge and hold it there.
- Rotate slowly around the rim. Slow is the whole trick. A fast lap skips angles and leaves gaps that look like hardware faults but are really just missed samples.
- Keep going until the timer runs out. Two unhurried laps beat six frantic ones.
- Read the coverage number first. Below 85 percent, run it again before you trust anything else on the page.
If you have not checked the rest of the pad yet, start with the full controller tester so you know whether the gate is the only thing worth chasing, then come back here for the detailed outer travel reading.
What This Test Actually Measures
Every analog stick has a physical boundary, the point where the plastic stops moving because it has hit the edge of its housing. Push against that boundary and rotate, and the stick sends a stream of coordinate pairs to your computer. Plot those pairs and you get a shape. That shape is your gate, and it is what this test draws.
The reason the shape matters is that games do not read your stick as a picture. They read it as a distance from centre and a direction, and then they scale movement by that distance. If your stick reaches 1.00 when you push right but only 0.82 when you push down and to the left, then your character genuinely moves slower to the lower left than to the right, at exactly the same physical stick position. You will feel it as sluggishness in one corner of the screen without ever being able to say why.
The tool records the furthest distance reached in each of 72 narrow angle segments, then compares those distances to each other. Four numbers come out of that comparison.
Gate coverage
The share of angle segments that received at least one sample past your outer travel threshold. This is a measurement quality number, not a hardware number. Low coverage almost always means the lap was too fast, the stick was not pressed hard enough against the edge, or the pass ended early. Fix the coverage before you read anything else, because a partial lap makes a perfectly good stick look broken.
Circularity error
The average distance between each segment reading and the overall average, expressed as a percentage. Zero would be a mathematically perfect circle. Real hardware never gets there. Under 7 percent is a clean, even gate. Between 7 and 13 percent you have a shape that is measurably uneven but usually still playable. Past 13 percent the unevenness is large enough that most players notice it as inconsistent aim or steering.
Mean outer travel
The average distance the stick reached across all measured directions. This is a range number rather than a shape number. A stick can be beautifully round and still weak, and that combination shows up as low circularity error paired with a mean below about 0.90. Weak range usually points at a worn housing, a stick module that has lost travel, or software that is scaling your input down before the browser sees it.
Range spread
The gap between your furthest and shortest reach, relative to the average. Circularity error describes the general roughness of the whole ring. Spread describes the single worst offence. A stick with low error and high spread has one specific bad direction sitting in an otherwise decent gate, and that is a very different repair conversation from a stick that is uniformly rough everywhere.
Reading the Shape You Traced
The numbers narrow things down. The shape on the canvas tells you what is actually wrong. These are the patterns that show up over and over.
| What you see | What it usually means | What to try |
|---|---|---|
| A rounded square with the corners pushed out | A square gate. The housing lets the stick travel further diagonally than it does straight up, down, left or right. Very common on older pads and on plenty of cheap third party ones. | Nothing is broken. Turn on circular deadzone or radial scaling in the game if it offers one, or let Steam Input handle the squaring for you. |
| A ring with the diagonals sliced flat | Clipping. Both axes hit their limit at the same time, so the corners get cut off at a straight line instead of curving. | Usually a driver or remapper squaring the output. Close DS4Windows, reWASD or similar and run the pass again before blaming the hardware. The controller mapping tool shows you exactly which axes the browser is receiving. |
| A round ring with one dent in it | A worn spot. The sensor or the wiper has degraded at one angle, often the direction you hold most during play. | Clean around the stick base first, following the steps in the controller drift repair guide. If the dent survives cleaning, it is mechanical and only a module swap fixes it. |
| An oval, wider than it is tall or the reverse | One axis has lost more range than the other. Very typical of potentiometer wear, because the two potentiometers rarely wear evenly. | Recalibrate through the console or the vendor app, then confirm the result with the controller checker. If the oval survives calibration, the stick module is on its way out. |
| A small ring that never approaches the outer circle | Your travel is being scaled down. Either the housing has worn and the stick no longer reaches the stop, or something in software is capping the value. | Check for an active sensitivity or range setting in vendor software, then test on a second device to separate the pad from the PC. |
| A jagged, spiky outline | Electrical noise on top of your movement, not a shape problem at all. | Try a wired connection, a different USB port, and a different browser. Noise that only appears on Bluetooth is interference rather than wear, and the connection stability test will confirm it. |
| A ring that is offset from centre | The neutral point has moved. This is drift showing up inside a circularity test. | Capture the resting centre with the button in the setup panel, then rerun. If the offset is large, run the stick drift test instead. |
Circularity, Drift and Deadzone Are Three Different Problems
These get mixed up constantly, including by people selling replacement parts, so it is worth being precise.
Drift happens at the centre while nobody is touching the stick. The resting value sits away from zero and your character wanders. It is a rest position problem, and the stick drift test is the page that measures it.
Deadzone is the software fix for drift. It is a threshold below which input gets thrown away. It costs you precision near centre in exchange for ignoring the noise. It is a settings problem rather than a hardware one, and the joystick deadzone test finds the smallest value that still does the job.
Circularity happens at the outer edge while you are pushing as hard as the housing allows. It is a full travel problem, and no deadzone setting anywhere will touch it.
You can absolutely have all three at once, and they can each be at a different severity. A stick can rest perfectly at zero and still have a badly dented gate. A stick with terrible drift can have an almost flawless outer ring. Testing only for drift, which is what most controller tools do, means you miss half the picture. Running all three in one sitting, starting from the main gamepad tester, gives you the whole pad rather than a slice of it.
What Is Normal for Real Hardware
Potentiometer sticks
The carbon film design used in most stock controllers, including the standard Xbox and PlayStation pads. Physical contact means physical wear. A new one typically lands in the 3 to 6 percent range for circularity error. After a couple of years of heavy play, 8 to 12 percent is common and often still perfectly playable. Past 15 percent you are usually looking at a specific worn direction rather than general ageing.
Hall effect sticks
Magnetic sensing with no contact between the moving parts, so there is nothing to wear down electrically. These usually start cleaner and, more importantly, stay clean. The catch is that hall effect solves sensor wear, not mechanical wear. The plastic housing still degrades, so a hall effect stick can still develop a shape problem even though its readings stay noise free.
TMR sticks
A newer magnetic approach appearing in premium and enthusiast controllers. Similar durability story to hall effect, generally with lower power draw and finer resolution. The same caveat applies about the housing.
What a good result looks like
Coverage above 90 percent, circularity error under 7 percent, mean outer travel above 0.92, and no single direction sitting more than about 8 percent below the average. Hit all four and your gate is not what is holding you back. If the pad still feels wrong after a clean result here, the next two places to look are the latency test and the polling rate test.
Where the Uneven Shape Shows Up in Games
Circularity error is abstract until you connect it to something you have actually felt.
Shooters. Aim speed scales with how far the stick is pushed. A gate that is short toward the lower right means every flick in that direction arrives slower than the identical flick to the upper left. Because the difference is directional rather than constant, muscle memory never adapts to it. It just feels like you miss more often on one side.
Racing. Steering uses a narrow band of the stick and holds it for a long time. A dent in one direction turns into a car that will not hold a constant radius through a long corner, and drivers usually blame the game before the pad. Analog throttle and brake run on the same kind of scaling, so pair this with the trigger pressure test if braking feels inconsistent too.
Fighting games. Motion inputs pass through the diagonals. A gate that undershoots the corners means quarter circles and dragon punch motions register inconsistently, especially on faster inputs where the stick spends only a few frames in each direction. Drill the motion with the gamepad viewer open and you can watch each input land or fail to land.
Platformers and action games. Many games apply a walk threshold at roughly 0.7 travel. A stick that only reaches 0.85 in one direction sits uncomfortably close to that boundary, so your character occasionally walks when you asked it to run, and only when moving one particular way.
What Software Can Fix and What It Cannot
Fixable in software
- A square gate. Radial scaling maps the square shape onto a circle. Steam Input does this well, and many games include a circular deadzone option that achieves the same thing. Check what your current profile already does with the controller mapping tool before you stack another layer on top.
- Uniformly short travel. Range scaling multiplies your output up so a stick that tops out at 0.88 reports a full 1.00 at the edge.
- A small centre offset. Recalibration moves the neutral point back to zero, and the resting centre capture in this tool does the same thing for the duration of your test. For a permanent value to type into your game settings, use the joystick deadzone test.
- A modest oval. Independent per axis scaling can even out two axes with different ranges.
Not fixable in software
- A dent in one direction. Scaling stretches the whole shape, dent included. You cannot fill a hole by making the ring bigger.
- Jitter and noise on the reading. Filtering hides it and adds latency in exchange. It never removes the cause.
- Mechanical grinding, catching, or a stick that physically snags at one angle. That is friction inside the housing.
- Anything that keeps getting worse week over week. Software compensation applied to a failing stick just moves the goalposts until you run out of room.
When Replacement Is the Right Call
Replace or repair when circularity error passes roughly 15 percent and the shape has a clear defect rather than general roughness. Replace when one direction sits more than 15 percent below the others, because that is a localised failure and it spreads. Replace when the result gets measurably worse across passes taken weeks apart, which is why saving your report matters. Replace when the pad is still under warranty and any of the above applies, because you are giving away a free repair by living with it.
Hold off when coverage came in low, since the reading is not trustworthy yet. Hold off when the shape is a clean square, because that is a design characteristic and software handles it. Hold off when the pad is fine on a wired connection and only misbehaves wirelessly, because that is a link problem rather than a stick problem. Hold off when a remapping tool was running during the pass, until you have retested with it closed.
On most controllers the stick module itself is a cheap part. The cost is the labour, because it is soldered to the board. If you are not comfortable with a soldering iron, a repair shop is the sensible route, and on a mid range pad the repair bill often lands close enough to a replacement that it becomes a coin flip.
Getting a Measurement You Can Trust
Rotate slowly
This is the single biggest source of bad results. The browser samples roughly 60 times a second. A one second lap gives you about 60 samples spread across 360 degrees, which leaves visible gaps. A six second lap gives you 360, and the shape resolves properly.
Stay against the edge
Circularity is a measurement of the boundary. If your thumb drifts inward halfway round, you are measuring your thumb rather than the gate. Keep constant outward pressure the whole way.
Test one stick at a time
Combining sticks blends two sets of wear into one meaningless average.
Close your remapping software
DS4Windows, reWASD, Steam Input and vendor apps all sit between the controller and the browser, and any of them can reshape the output. Test with them off, then test again with them on if you want to see what they are doing. The controller mapping tool is the fastest way to spot something rewriting your axes.
Rule out the link before blaming the stick
A wireless pad on a weak connection produces gaps and spikes that look exactly like hardware faults. A dropout mid lap costs you coverage, and a low report rate costs you samples. Run the connection stability test and the polling rate test first, and a wired pass after that, before you decide a stick is worn.
Run it twice
Two passes that agree with each other are a measurement. One pass is a guess. If the numbers diverge noticeably, your technique changed between laps and neither result is reliable yet.
Save the report
The number today matters far less than the difference between today and six months from now. That comparison is what turns a snapshot into evidence, and it is what a warranty claim actually needs.
Browser and Platform Notes
Everything here runs through the browser Gamepad API, which shapes what is possible in a few specific ways.
- Chrome and Edge give the most consistent axis behaviour. Firefox works but occasionally maps axes differently on the same pad. Safari support is the least predictable.
- Sampling is tied to the browser animation loop, so your effective sample rate follows your display refresh rate. A 144Hz screen collects more points per lap than a 60Hz one. That affects coverage speed but not the measured shape.
- Background tabs get throttled hard. Keep this tab focused for the whole pass or the samples stop arriving.
- Some third party controllers only expose their axes properly in DirectInput mode. If the dot does not move, look for a mode switch on the pad, often a button combination involving Home.
- The axis mapping selector exists because a handful of controllers put the right stick on unusual axis numbers. Automatic covers almost everything, and the manual options cover the rest. If you are unsure which axes your pad uses, the controller mapping tool prints them.
- A controller low on charge reports less consistently over Bluetooth. If your numbers wander between passes, check the pack with the battery health test before drawing conclusions.
- Nothing you do here leaves your device. The tool reads the controller in the page and does the maths in the page. Exports are files your browser writes locally, and only you decide where they go.
Questions People Actually Ask
- My trace is square. Is my controller broken?
- Almost certainly not. A square gate is a housing design choice and plenty of controllers ship that way. It becomes worth acting on only when a game does not apply radial scaling, and even then the fix is a setting rather than a repair.
- Why does my coverage stop climbing near 90 percent?
- The last few segments are usually the awkward ones near a corner where your thumb naturally lifts slightly. Slow down through those directions specifically and the remaining segments fill in.
- Can circularity error cause stick drift?
- No, but both often come from the same underlying wear. A stick that is deteriorating mechanically tends to develop problems at both the centre and the edge, on its own schedule at each end. Run the stick drift test as well so you have both ends documented.
- My readings sit above 1.00. Is that a fault?
- It is a mapping quirk. A few controllers report raw values slightly beyond the normal range on the diagonals before anything clamps them. The tool caps the display at full travel. The shape is still valid.
- Should I test with the controller cold or warm?
- Both, if you are chasing something intermittent. Plastic and electronics behave differently at different temperatures, and a fault that only appears after an hour of play will not show up in a pass taken from cold.
- Does this work on a phone or tablet?
- Yes, with a controller connected over Bluetooth or through a USB adapter. The canvas is touch friendly and the layout collapses to one column. Expect slightly noisier readings on mobile than on a wired desktop setup.
- Why does my hall effect controller still show error?
- Because hall effect fixes the sensor, not the housing. Magnetic sensing removes electrical wear from the equation, but the physical boundary the stick pushes against is still plastic and still wears.
- Is a lower circularity error always better?
- Broadly yes, but chasing the last percent is not worth much. The gap between 4 percent and 6 percent is invisible in play. The gap between 6 percent and 16 percent is not.
Related Tests on Gpadtester
Circularity covers the outer edge of one stick. These pages cover everything else, and running them in sequence tells you far more than any single reading can.
Sticks and triggers
Whole controller
Connection and response
Read next
How to fix controller drift covers cleaning, recalibration and module replacement. It is the natural next step if your trace showed a dent rather than a square.
Other hardware
Away from the controller: keyboard test, mouse test, mic test, frame rate test and refresh rate test.