DevicedialDoes the machine you just bought do what the box said?
The specification is a promise. This is the hour in which it is still cheap to check.
One gear check a month
MonthlyOne thing a month worth measuring before it fails properly โ a mic that reset itself, a stick that drifts, a switch sending two clicks. Two sentences and the test.
Free to use โ Pro is for keeping it
This tool, like every other one here, is free and needs no account. A free account keeps three setups and three email alerts. Pro takes the ads off every site we run and lifts that to 200 setups and 50 alerts, adds a public page for any result and lets you embed the tool without our badge.
Try also
What should I test on a new computer, monitor or headset?
The Devicedial new machine test plan puts a target beside every reading and counts how many of them the machine actually met. Each row carries the figure you were promised, the figure you measured, and which direction counts as better โ more hertz is good, fewer milliseconds and fewer dead pixels are good. That last switch is what lets one sheet hold hertz, decibels, frames and counts at the same time without adding numbers that have nothing to do with each other.
The six rows it opens with are the checks that most often disagree with the box: refresh rate, polling rate, microphone level and background noise, dead pixels and camera frame rate. Five of those six pass in the example and one does not, which is roughly what a real machine does. Run it in the first week. A monitor that delivers 60 Hz is an argument you win in the return window and lose a month later.
What to measure on a new machine, and a target that is not made up
Every figure below comes from a test on this site, and every target comes from the specification you were sold or from the scale the test prints.
| Check | Where the number comes from | A target that holds up |
|---|---|---|
| Screen refresh (Hz) | Refresh rate test | The rate printed on the box |
| Mouse polling (Hz) | Polling rate test | The rate printed on the box |
| Dead or stuck pixels | Dead pixel test | Zero, before you get attached to it |
| Mic peak while speaking (dBFS) | Mic test | โ12 dBFS or louder |
| Mic background noise (dBFS) | Mic test | Quieter than โ60 dBFS |
| Camera resolution and frame rate | Camera test | What the specification claims |
| Keys registering at once | Keyboard tester | Six, or every key on a board sold as NKRO |
| Stick drift score | Stick drift test | Five or lower on an untouched stick |
| Touch points at once | Touch screen test | The number the panel is sold with |
Targets are one-sided on purpose, because that is what a count can check. Microphone level is the exception worth knowing: the band is โ12 to โ6 dBFS, so a reading far above the target is its own problem and the mic test says so.
The order worth running them in
Start with the screen, because it is the check that most often fails and the hardest to argue about later: the refresh rate test for the rate actually delivered, then the dead pixel test on a clean panel, then screen resolution to confirm the two numbers a display has.
Then the things you talk through. The mic test gives you both figures this plan wants, peak level and background noise, and the camera test reports the resolution and frame rate the webcam is really sending rather than the one on its label. For a keyboard and mouse, the keyboard tester counts how many keys register at once and the polling rate test puts a number on how often the mouse reports. A controller gets the stick drift test before you have played enough to blame yourself.
What to do with the sheet afterwards
Keep it. The readings in it are the baseline that makes every later measurement mean something, and the two planners either side of this one both start from it โ the warranty claim log when a figure drifts out of spec, and the gear test schedule so the next reading happens on a date rather than after something has already gone wrong.
Frequently asked questions
Which of these can I get through before the return window closes?
All of them, in about twenty minutes. Every check in the sheet is a browser test on this site that needs no download and no account: the longest is the dead pixel sweep, because it is worth doing on a clean panel with the lights off, and the shortest are the rate tests at three seconds each. The return window is the only period in which a disappointing measurement is somebody else’s problem, so the plan is built to fit inside it.
The box says 144 Hz and the browser says 60. Is the machine faulty?
Usually not, and it is worth settling before you start a return. The common causes are the cable or port negotiating a lower mode, the display setting still sitting at its default in the operating system, hardware acceleration being off in the browser, or a laptop dropping the panel to save power. The guide on a monitor stuck at 60 Hz walks through them in the order worth checking, and only when all four are ruled out is the panel the suspect.
Is there any point testing a machine that seems fine?
That is the best reason to do it. A reading only means something twice: the second one tells you whether anything moved, and without a first one there is nothing to compare against. A stick that scores three now and nineteen in a year has a documented history rather than a suspicion, and the same sheet becomes the baseline for the warranty claim log if it ever comes to that.
How the test works
Sources: