Circularity Error Test

02 / Read the shape

Free Online Circularity Tester: Circularity Error Test

This free browser tool checks how evenly a thumbstick reaches its outer range. It reads both sticks in the same session. You can watch each path grow as you roll the cap around its rim. A round path means the stick sends near-even reach in each direction. A path with dents, bumps, or broad gaps needs a closer look. 🎮

The result helps you spot shape error before a game makes it hard to see. Aim may feel fine at the top, yet feel slow at the sides. A race car may turn well in one arc and then gain less steer in the next. A platform game may show a short hop when the cap hits a weak zone. The plot gives those small changes a clear form. It does not grade your skill. It shows the path sent by your pad during this run.

How to Use the Circularity Error Test

You do not need an app or a set-up file. Use a pad that your browser can read, then make one calm loop with each stick. Slow motion gives the page more time to save points. Fast flicks can skip parts of the rim and make the path look worse than it is. Keep your thumb on the cap with light, even force. Let the stick touch its full stop in each part of the loop.

1. Connect the pad. Use USB first when you want a clean check. Wireless input can work, but a link may sleep between moves.
2. Wake the signal. Press a pad button once. Wait for the top line to change from No Controller to Connected.
3. Reach the rim. Push the left stick to its edge. Keep the cap near that limit without forcing it past its stop.
4. Roll with care. Move through a full circle in two or three calm seconds. Cover the right, top, left, and bottom arcs.
5. Repeat on the right. The two panels use separate axis pairs, so each path can show a different shape.
6. Read the cards. Check the points, average error, and path. Clear the trace and run it again if your first pass was uneven.

A first pass can be a quick health check. A second pass gives you a better sense of repeat rate. Try the same motion with the same grip each time. Do not compare a slow full loop with a few hard taps. The sample count matters. A tiny set can look round by luck. A wider set shows more of the real edge. If one stick looks calm while the other has flat arcs, test each side again after a short rest.

“The trace is a map of reach, not a score for your play.”

Roundness is one part of pad health. If the cap rests off center, check the Stick Drift Test next. That page checks motion while your thumb is still. If the pad is a supported Sony model, the PS Controller Calibration page can help you review its center and edge set-up. For input timing, pair this shape check with the Controller Polling Rate Test. Each page checks a new part of the same control path.

03 / Follow the data 📈

How the Circularity Error Test Works

A gamepad sends stick data as small numbers. The browser reads those values on a pair of axes. The left stick is often axes 0 and 1. The right stick is often axes 2 and 3. This page keeps those pairs apart, then draws a live path for each one. The guide ring gives your eye a fixed shape. The path shows how close the real reach stays to that ring.

The names can vary by pad and driver. Some low-cost pads expose extra axes. Some maps report only one stick. This tool does not invent data for a missing pair. It shows an unavailable state instead. That choice keeps the readout honest when a browser map is not standard. You can check the raw API model in the MDN Gamepad API guide.

The Circularity Error Test keeps the raw axis pair visible beside each path. That small detail helps you spot a bad map. A path can look thin because the stick has wear, or because the browser never saw a full range. The axis readout gives you a second clue before you take the pad apart.

What the trace measures

Each frame gives the page an X value and a Y value. Together they form a point on a small plane. At rest, the point sits near the center. The page drops tiny center moves because thumbsticks can report small noise while still. Once the reach passes the edge threshold, the page saves the point. Its distance from center is the current radius. Its angle tells the page which part of the rim it came from.

The full edge is split into small angle buckets. When two points land in one bucket, the stronger reach stands in for that slice. This keeps the trace light and stops a long test from growing without limit. It also makes a slow loop easy to read. A full set has many points around the ring. A short loop has gaps. Those gaps do not prove a fault. They tell you to run a fuller pass before you judge the shape.

How the error value is formed

The page first finds the largest valid radius in the current trace. It then compares each saved radius with that high mark. If one slice reaches 0.96 while the high mark is 1.00, its shape gap is near four percent. The page takes the mean of all saved slice gaps and shows that value as average error. The math looks at roundness, not at how fast you move the cap. That is why a slow loop is the best way to collect a fair set.

The value is a browser estimate for the run on screen. It is not a lab grade, a factory spec, or a repair claim. A hand can press harder on one side. A shell can stop the cap in a slightly new place. A driver can scale one axis in a way that bends the path. Use the trend and the shape as clues. If the same dip shows in three calm runs, it deserves more care than one odd pass.

These bands are simple guides for play. They do not set a hard pass mark for every game. A slow story game may feel fine with a broad edge gap. A fast aim game may make a small gap clear. A racing game may show the issue only during a long turn. Your thumb pressure and the cap shape also affect the path. Read the graph with the game feel you notice.

Parts can shape the result too. A worn pot may lose reach in one arc. A tight shell can block the cap near a corner. A Hall sensor may keep a round path for longer, yet its map and shell still matter. Bluetooth and USB do not change the shape by rule, but a lag or drop can leave gaps in the sample. If one side looks weak, reconnect the pad and test again before you call it a part fault. 🧭