Calibration: what the software actually writes
An infrared frame is not bonded to your display. It clamps on from the outside, so the relationship between the beam grid and the pixels behind it is a physical fact that has to be measured once and stored. That measurement is calibration — and it lives in the controller, not in the operating system.
Infrared measures, capacitive tunes
Infrared needs calibration because the sensor and the display are two separate objects that are aligned by hand. Four corners of a frame can be 2 mm off from where the drawing says they are, and the beam grid has no way of knowing — so the tool asks you to touch known points and solves for the mapping.
A capacitive panel needs no calibration because the sensor is bonded to the glass and the glass is cut to the housing. The coordinates are where the electrodes are. What a capacitive panel sometimes needs is tuning: four firmware parameters that decide how sensitive it is, how fast it reports and how it behaves at the edges. Same tool family, different tab, different purpose.
Four points, nine, or twenty-five
The tool offers three densities, and the right one depends on how much accuracy the application needs and how patient the operator is.
| Mode | Time | Use it for | Result |
|---|---|---|---|
| 4-point | 20 seconds | Kiosks, signage, anything where a finger targets a 40 mm button | Solves origin, scale and rotation. Fastest, and adequate for 90 % of installs |
| 9-point | 45 seconds | Meeting boards and whiteboards where people draw | Adds per-axis linearity correction; removes the bow that shows up mid-edge |
| 25-point | 2 minutes | Large format above 86 inch, curved frames, anything with unusual glass | Maps local distortion across the surface, including the sag a long extrusion develops |
Recalibrate, or leave it alone
| Event | Recalibrate? | Why |
|---|---|---|
| First installation | Always | The frame clamps on the outside, so the mapping depends on how it sits on this particular display. |
| After replacing a controller or an LED tube | Always | The new board has no profile, or the old profile describes a different beam grid. |
| After the unit is moved, or the display is taken off the wall | Always | Removing and refitting changes the frame's position by a millimetre or two, which is 3–5 mm on the far corner. |
| After changing the cover glass thickness | Always | Glass refracts the beams; a 3 mm change shifts the far-corner mapping. |
| After rotating the frame 90° | Always, plus the orientation bit | Without the orientation bit the axes swap and the cursor moves sideways. |
| After an OS update or a driver change | Never | Calibration lives in the controller, not in the operating system. An OS update cannot lose it. |
| After cleaning the glass | No | Unless something was left on the surface during the original calibration. |
What we send, and what it does
Every order ships with the calibration tool and the profile file for the size you bought. It is not a download from a support portal that asks you to register — it comes attached to the order email, and again on request for as long as you own the unit.
- Windows tool — 32 and 64-bit builds for Windows 7 through 11. Runs from a folder, no installer, no admin rights needed for calibration.
- Linux command line — for kiosk images and Yocto builds where installing a GUI is not an option. Reads the same profile format.
- Profile export and import — .cal files, so a production line can clone one good unit into a batch, or a field engineer can email you a profile to reflash.
- Diagnostics tab — live beam map. A blocked or dead LED pair shows as a missing line before it becomes a customer complaint, which is how you tell a dirty filter strip from a failed tube.
- Firmware and descriptor tools — for reflashing a controller that enumerated as a mouse, or changing the touch-point count on a spare board.
If your integration needs the protocol rather than the GUI — for example, calibrating inside your own production test software — ask for the command reference. We send it to integrators who are building the frame into their own line.
Ten minutes, shared screen
For first-time installs we do the run together rather than describing it. It is faster than writing instructions, and it catches the two mistakes that make calibration fail: touching the frame instead of the glass, and calibrating before the frame is finally fixed.
Send the size and a time window
Plus your timezone. We reply with a slot and the tool build for your Windows version, if you do not already have it.
Connect the frame and open Device Manager
The device must read “HID-compliant touch screen”. If it reads “mouse”, we sort the descriptor first — calibration cannot fix that.
Screen share, then touch where we point
Four corners, then nine if the application draws. You will see the cursor land as each point is accepted.
Check the corners and the edges
A finger dragged 5 mm inside each border. This is where a frame that passed four points still fails.
Save, then power-cycle
The profile is written to EEPROM. Pull the USB, plug it back in, and confirm the corners still land — that proves the save, not just the session.
Four parameters, not a calibration
A capacitive panel does not have a calibration step. If a customer reports that theirs “needs calibrating”, the actual fix is almost always one of these four numbers, changed in the settings file and written to the controller.
| Parameter | What it does | When to change it |
|---|---|---|
| Touch threshold | How much capacitance change counts as a touch | Raise it for thick cover glass or a gloved hand; lower it if a panel misses light taps |
| Report rate | How often positions are sent, 60–200 Hz | Lower it for long or noisy cables, raise it for drawing and whiteboard use |
| Edge suppression | Dead band at the border, 1–5 mm | Increase it if the housing presses on the sensor; too much kills buttons at the edge |
| ESD recovery level | How the controller recovers after a discharge | Higher levels recover faster on outdoor units, at the cost of a brief touch dropout |
Be honest about the ceiling
| Symptom | Real cause | The actual fix |
|---|---|---|
| Dead stripe down one side | A failed LED tube, or a glass edge crossing the beam | Replace the tube or move the glass — calibration cannot create a beam that is not there |
| Cursor drifts a few millimetres after a reboot | The controller is not saving the profile, or the OS applies its own scaling | Check the profile is written to EEPROM; check display scaling is 100 % |
| Touch works but the position is inverted | X/Y swapped or the orientation bit is wrong | Set orientation in the tool, not in the OS |
| Jitter only in humid weather | Condensation on the filter strip | Clean and reseal; a gasket problem, not a calibration problem |
| Ghost touches with nothing touching | Grounding, or ambient IR from sunlight or a heater | Fix the ground first, then consider the high-immunity tube set |
Calibration questions
Do I need to calibrate a capacitive panel?
Where is the calibration stored?
Can I calibrate a batch on the production line?
Can you calibrate it for me remotely?
The cursor is offset by a constant amount everywhere. Is that calibration?
What if the tool does not see the controller?
Calibration problem? Send the beam map.
Connect the frame, open the diagnostics tab and screenshot the beam map. That image tells us in one look whether it is calibration, a tube, a cable or the ground.