CTRLCHK

Range envelope

See whether your stick actually draws a circle.

Circularity error is the missing twin of stick drift: the shape of full deflection, not the motion at rest.

Rotate each stick slowly around the rim during the range step. Average and per-quadrant error are computed from that motion.

Device

No controller detected

Waiting
Live controller input visualization

Connect a controller over USB or Bluetooth, then press any button to begin.

Standby—

Live readout

Every axis and button, measured in real time.

Nothing untouched is treated as a failure. This is a viewer, not a verdict — run a health check when you want scored measurements.

LX+0.00 
LY+0.00 
RX+0.00 
RY+0.00 
L radial0.0% 
R radial0.0% 
L2N/AUnmapped
R2N/AUnmapped
Buttons observed0 / 0

Why circularity is not a screenshot of a circle

A lot of “circularity testers” ask you to spin the stick and then overlay your path on a unit circle centered at the origin. That plot is pretty and often wrong. If the stick rests at (+0.12, −0.04), a mechanically round gimbal becomes an offset oval around (0, 0). You then “fail” a controller that is merely biased — which is a drift problem, not a circularity problem.

CTRLCHK subtracts the median rest from every sample, keeps only outer-reach points (radius ≥ 0.7), bins them around 360°, and compares each bin’s max reach to the mean reach of the envelope. Average error is the mean of those relative deviations. Quadrants (up-right, down-right, down-left, up-left) get their own mean reach and error so a smashed corner cannot hide in a global average.

How to run it

Start the health check, finish the hands-off centre so rest is known, then rotate each stick once around the outer edge. Slow is better than fast: we need samples in most angular bins, not a blur. Continue unlocks when range coverage is complete; circularity is computed from the same samples. Bands: under 5% excellent, 5–10% good, 10–15% moderate, over 15% poor. Poor circularity applies a penalty to the range contribution of the health score. It does not invent a drift failure.

If you only care about rest, stay on the stick drift test. If the complaint is “diagonals feel short” after a drop, this is the page. Snapback after a flick is a third axis — see the snapback test.

What the number cannot say

Circularity here is a 2D envelope in Gamepad units, not a CMM scan of the gimbal. Software gates, anti-deadzone, and the game’s own circularity correction sit downstream. A hall-effect module can still plot an oval if the firmware tilts the axes. Treat the measurement as evidence, then confirm in the game that actually matters to you.

FAQ

Questions people actually ask

What is circularity error?

Deviation of the stick’s outer envelope from a circle after subtracting the measured rest position. CTRLCHK reports average error and per-quadrant error. Under 5% is excellent; over 15% is poor.

Is circularity the same as drift?

No. Drift is motion at rest. Circularity is the shape you trace at full deflection. A pad can rest perfectly and still draw an oval.

Why subtract rest first?

A biased centre makes a perfect mechanical circle look like an egg if you measure from (0, 0). Rest-corrected reach is the honest shape of the gimbal.

Does an oval stick always need replacing?

Not always. Some cheap pads are oval from the factory and still playable. Above 15% error you will feel diagonal gating in games that expect a round gate.