KTP900F Touch Calibration Failure: Hardware Diagnostic Procedure

1. Problem Overview

The SIMATIC KTP900F (Key and Touch Panel, 9-inch widescreen) is a member of the SIMATIC HMI Comfort Panel family that combines an analogue-resistive touch overlay with mechanical function keys. A recurring field failure mode is loss of touch calibration — touches are registered offset from the actual finger position, in some quadrants only, or fail to register at all.

The failure typically presents as one or more of the following user-visible symptoms:

• Touches land 5–40 mm away from the intended target, especially near the corners.
• Calibration drifts back to an incorrect state within minutes to hours after the on-device calibration routine completes.
• Calibration dialog itself is unreachable because a touch is required to enter Control Panel > OP > Touch.
• The 16 function keys (F1–F16) continue to operate normally even when the touch overlay is unusable.
• The panel otherwise boots, displays the project, and communicates over PROFINET/Ethernet without alarm.

When this behaviour survives a full factory reset, a WinCC image update (formerly OS update) and an engineering change of TIA Portal version from V16 to V17 or V18, the panel is no longer behaving like a configuration issue. It is behaving like a degraded touch sensor stack.

2. KTP900F Identification and Hardware Background

Before opening the panel or generating an RMA, confirm the exact article number and firmware state. The KTP900F exists in multiple hardware revisions that share a similar part-string prefix.

*Region-specific suffixes (-0BX0, -0CX0, etc.) cover pre-configured versions for selected markets and certification bundles. Mechanical construction, touch controller, and service procedure are identical.

The touch sensor stack consists of:

1. A polyester top layer (clear, anti-glare coated).
2. An ITO (indium tin oxide) top conductive layer.
3. A spacer dot matrix.
4. An ITO bottom conductive layer on the rear glass.
5. A 4-wire analogue-to-digital front-end connected to the panel motherboard via an FFC (flat flexible cable) at the rear of the display assembly.

Each layer is bonded under heat and pressure. The dominant end-of-life mechanism is ITO trace micro-fracture caused by repeated mechanical deflection of the front membrane — most often near the corners, where users press hardest. Once a corner trace develops a hot-spot of high resistance, the controller interprets the divided voltage as a touch centroid biased toward the dead corner, which is exactly what the operator observes as "drifted calibration".

3. Official Calibration Procedure (Baseline Reference)

Before declaring a hardware fault, the engineer must run the documented Siemens calibration routine and observe whether the result is stable. The procedure is the same for all Comfort Panels and is documented in the WinCC Comfort / TIA Portal help and in the panel operating instructions.

3.1 Entering the Control Panel on a Run-time Project Panel

1. Power on the panel and wait for the project start screen.
2. Press the Help button in the project (if configured) or navigate the project to a button linked to the system function OpenControlPanel / AcknowledgeAlarm trigger.
3. Alternative: use the loader at boot — hold the centre of the touch surface during power-up until the loader appears, then select Settings > OP > Touch.
4. If a USB mouse is available, plug it into the front USB port and use it to navigate to Control Panel > OP > Touch > Recalibrate.

3.2 Calibration Targets

The panel asks the operator to touch five crosshair targets: upper-left, upper-right, lower-left, lower-right, and centre. The system computes new X / Y scaling factors and stores them in the registry hive:\p>

[HKEY_LOCAL_MACHINE\Software\Siemens\HMI\Touch]
"CalibrationX1"=…  "CalibrationX2"=…
"CalibrationY1"=…  "CalibrationY2"=…

3.3 Stability Observation Window

After saving, observe the panel for at least 30 minutes of powered operation in ambient shop-floor temperature. Drift within that window is a strong indicator of a degraded ITO layer rather than a bad calibration constant.

4. Root Cause Classification

The failure has three families of cause. The diagnostic procedure is built to confirm which family applies before any hardware is ordered.

The source field experience — three independent projects, identical symptom, identical behaviour after a V16 → V17 upgrade, identical behaviour after factory reset, identical behaviour after OS update — drives the prior probability strongly toward family C. The diagnostic below is designed to rule out A and B with high confidence so that family C can be reported to Siemens support without ambiguity.

5. Structured Diagnostic Procedure

Work the procedure top to bottom. Do not skip steps. Record the result of each step in the maintenance log.

5.1 Step 1 — Confirm Article Number and Image Version

On the panel, navigate to Control Panel > System > Device > Device Version or read it from the loader start screen:

• Record Article No. (must match the plate on the rear of the unit).
• Record Image version (e.g. V16.0.0.6 or V17.0.0.12).
• Record Boot loader version.

If the article number on the plate does not match what the firmware reports, the panel has had its image swapped in service — re-flash the correct image and retest.

5.2 Step 2 — Test the Touch in the Loader

Reboot the panel and select Settings > OP > Touch > Touch Test. The test screen paints a 4 × 4 grid and a status field that updates the raw X / Y ADC counts in real time.

• Press each grid intersection with a plastic stylus.
• Look for dead rows / columns (the same raw X or Y value regardless of stylus position).
• Look for large jumps (>300 ADC counts between adjacent grid points).

Expected behaviour: smooth monotonic X / Y trace across the panel area. Any flat-line region, sudden step, or wrap-around is definitive hardware evidence.

5.3 Step 3 — External Mouse Bypass Test

Plug a USB mouse into the front USB-A port of the KTP900F. The resistive touch and the USB HID mouse are completely separate input paths. If the project is fully operable with the mouse, then:

• The display is good.
• The motherboard CPU, image, and project are good.
• The remaining suspect is the resistive touch stack, its FFC, and its analogue front-end.

This single test rules out the entire software/configuration family in under five minutes.

5.4 Step 4 — FFC Inspection (Power Off, ESD Safe)

1. Power down the panel and isolate from 24 V DC.
2. Wait 60 s for the bulk capacitors to discharge.
3. Remove the KTP900F from its cut-out using the integrated mounting clips.
4. Open the rear housing. The touch FFC is a short, gold-finger, 4-wire ribbon that exits the display bezel assembly and seats into a ZIF connector on the rear of the motherboard.
5. Re-seat the FFC: lift the ZIF latch, withdraw the ribbon, inspect the gold pads under 10× magnification for wear or contamination, reinsert with the cable straight, and close the latch.
6. Reassemble and retest.

If the calibration stabilises after re-seating, the fault was a high-resistance contact at the ZIF, which is recoverable. If the fault persists, the FFC itself is likely damaged or the ITO layer underneath is degraded.

5.5 Step 5 — Resistance Map of the Touch Stack

With the FFC exposed and the panel still de-energised, measure the resistance across the four bus bars on the ribbon with a calibrated bench DMM:

• Top layer X bus: 30 Ω – 80 Ω end-to-end, no opens.
• Bottom layer Y bus: 30 Ω – 80 Ω end-to-end, no opens.
• Top to bottom layer isolation: > 1 MΩ at 50 V.

An open circuit on either bus indicates a broken ITO trace (hardware). A low (<200 kΩ) isolation value indicates contamination or pressure damage shorting the two layers. Either reading confirms family C.

5.6 Step 6 — Cross-Verify with a Second Project Image

Transfer a clean WinCC project compiled against the same panel article number and run it for 30 minutes. A software-only fault would not persist across a project swap; a hardware fault would. This step is mostly documentary because the loader test and mouse bypass already isolate the issue, but it produces a clean evidence packet for the RMA request.

6. Repair and Replacement Procedure

Once family C is confirmed, there are three downstream options.

6.1 Option A — Replace the Touch / Display Assembly

Siemens sells the front-face kit (touch + display + bezel assembly) as a single service part for the Comfort Panel family. For the KTP900F this is:

• Spare part kit reference: 6AV2 124-1JC01-0AX0 (full unit exchange) or the equivalent display sub-assembly kit documented in the SIMATIC HMI Comfort Panels service manual.

Replacement time is approximately 25 minutes with a T8 Torx driver, an ESD mat, and the FFC re-seating procedure above. Re-fit the existing image from the SD card or via ProSave before commissioning.

6.2 Option B — Panel Exchange with Same Article Number

Order a replacement KTP900F with the same MLFB and the same image version pre-installed by the distributor. This is the lowest-risk field option: no software intervention beyond loading the project, and no commissioning curve beyond the standard panel start-up checklist.

6.3 Option C — Retrofit to a TP900 Comfort (Touch-Only)

If the function-key row is unused, a TP900 Comfort Panel (touch-only, same 9" widescreen footprint) is drop-in compatible at the cut-out. The project requires a re-target of the panel device in the TIA Portal project tree and a re-compile / re-transfer. This is the preferred retrofit in plants that have standardised on touch-only HMIs.

7. Verification After Repair

After any of the three options, run the full verification matrix:

1. Calibrate via Control Panel > OP > Touch.
2. Hold the panel under power for 2 hours at ambient shop-floor temperature. Recheck calibration drift.
3. Run a 24-hour soak test with the operator pressing each function key at least 50 times — this stresses the most common ITO failure zone near the function-key column.
4. Confirm PROFINET diagnostics are clean and the panel is reachable in the TIA Portal Online > Accessible Nodes.
5. Sign the panel off with the calibration constants, image version, and test duration recorded in the maintenance log.

8. Cross-Panel Application

The diagnostic and replacement procedure above applies unchanged to the wider SIMATIC Comfort Panel family. The table lists the article numbers and the same symptom set to be alert for in the field.

9. Documenting the Fault for Siemens Support

A well-prepared RMA request shortens the lead time considerably. The support engineer will request:

1. The MLFB and serial number from the rear plate.
2. The image version, bootloader version, and TIA Portal version.
3. A photo of the loader Touch Test screen showing the failed quadrant.
4. The result of the USB-mouse bypass test (should be "fully operable with mouse").
5. The result of the FFC resistance map (open or low isolation values).
6. The date of commissioning and the estimated hours of operation.

Submit the case through the Siemens Industry Online Support portal at support.industry.siemens.com with the support request type set to Hardware / Touch failure. Reference the SIMATIC HMI Comfort Panels operating instructions entry ID when quoting calibration behaviour.

10. Field-Proven Caveats

• USB mouse always works. Even on a panel with a totally dead touch stack, the front USB-A port continues to enumerate a HID mouse. Use this as the definitive diagnostic.
• Function keys are independent. The F1–F16 row is a separate matrix on the rear of the front bezel. A dead touch does not imply dead function keys.
• Image update does not fix touch hardware. A WinCC image update can repair OS-level corruption but cannot repair a broken ITO trace. Re-flashing repeatedly will only burn time.
• EMI can mimic the symptom. A panel installed within 200 mm of a Variable Frequency Drive or unshielded servo cable bundle can show calibration drift proportional to drive torque. Move the panel away from the noise source or reroute the cabling before concluding the touch is faulty.
• Stylus material matters. During calibration use a soft plastic stylus, not a finger. Finger contact area changes with pressure and skin moisture, which produces a slightly different centroid each pass and may make the test appear non-deterministic.
• Temperature gradient can mask hardware. A cold panel (5 °C) calibrated in an unheated cabinet may appear mis-calibrated once it reaches operating temperature (35 °C). Let the panel reach steady state before concluding.

11. FAQ