HMI screen won't wake from screensaver: two checks first

What each symptom points to

The pattern of the fault narrows the cause before you open the cabinet:

• Wakes sometimes, fails sometimes, no pattern. A race between the saver timer and a popup. The more often the project opens message boxes, the more often the panel fails to wake.
• Screen dark, CPU LED solid green. The panel is running and healthy. Touch is being refused by a held dialog, or the backlight is dimmed and the wake event never arrives.
• Screen dark, CPU LED off or blinking. Power first. Measure the DC terminals before you touch anything else.
• Terminal voltage below the panel's printed backlight cutoff during a blank. The supply or its wiring turned the backlight off, and the screensaver takes the blame it does not deserve.
• Voltage fine, LED green, no popup in the project, fault survives an update. The panel itself. Assemble the evidence and hand it to support.

Slow lighting after a successful wake is a different fault with a different fix; HMI screen slow to light after screensaver wake covers that one.

Read the rear LEDs before you tap again

Walk to the back of the panel during a blackout. The guidance for the panels in this example states two facts: the CPU LED shows solid green when everything is okay, and the backlight turns off below 19.2 V DC at the panel terminals. Every HMI manual prints its own cutoff and LED table in the power and indicators section, so read yours and substitute your panel's printed value everywhere a comparison appears below.

Photograph the LEDs while the screen is dark, and leave the panel in the faulted state until you have them. Every extra tap on dead glass tells you nothing new, and a power cycle erases the exact state you need to record.

Measure DC volts at the panel with the screen dark

Put the meter on the panel's own DC terminals, not on the supply's output terminals, so the wiring is inside the reading. Hold the probes through the blank and watch for drift: a rail that starts healthy and slides as the panel warms is the fault behind a lot of "random" no-wake behavior.

Commit the reading against the printed cutoff. A terminal reading within about a volt of the supply's setpoint clears the supply, and you go hunting the popup. A reading at or under the cutoff, or sliding toward it, puts the fault in the supply and its wiring. A rail that dips only while the panel is polling the PLC is the same family of fault; HMI backlight flicker when polling: check the power supply walks through that one.

Some panels also take a remote backlight-wake input, so a contact from the PLC can lift the backlight with no touch involved; external backlight wake control wiring documents the input on one such panel. It doubles as a test: if the contact raises the backlight and touch still does nothing, the panel is awake and the input lock is the problem.

The hidden popup that holds the lockout

Here is the documented behavior behind the random cases. A message box opens in the same instant the screensaver fires. The runtime locks touch input to that dialog, the saver dims or blacks the backlight, and the OK button the runtime is waiting for is invisible. Every touch outside the dialog lands nowhere, so the panel looks dead. The panel in the reported case sat on a release two versions behind, where this saver and popup interaction was known, and the update to the current release through the maker's documented update procedure ended it. Read the version screen on the panel itself after the transfer; the version installed on the engineering PC is a separate fact, and the panel's own version screen is the one that counts.

Reproduce the race before you fix it, so you know you fixed the right thing. Open the suspect popup on purpose, let the saver time out, then press the spot where the OK button sits. A screen that comes back at that tap shows the lockout and clears the touch hardware. The permanent fix lives in the project: close the popup automatically on a timer or an event, or keep it from opening in the seconds before the saver fires. Setting the HMI backlight saver timeout on a PLC panel covers where that setting lives, and turning an idle HMI screen off after 30 seconds and waking it by touch covers the wake-by-touch side. While you reproduce the fault, some terminals can be woken on a schedule through their own command interface, which keeps a lit screen on the line with no touch needed; waking a terminal from its screensaver on a schedule documents one such method.

If the vendor's AC adapter supply is in the circuit

One reported cluster of these no-wake faults traced to the panel vendor's own AC-to-DC adapter installation: grounding was insufficient and electrical noise was high. The test costs one spare supply and ten minutes. Substitute a different DC supply rated to the panel's voltage and current, change nothing else, and run the panel through three saver cycles. If the fault disappears with the substitute, the repair is grounding and wiring, done permanently; a temporary supply left in as the permanent answer is how the fault comes back. If the fault stays, refit the original supply and work the popup and version checks.

Example: two readings, two different faults

Scenario: a 10-inch panel on a 24 V DC rail blacks out at the saver timer and ignores touch. Its printed backlight cutoff is 19.2 V DC, and its manual says the CPU LED is solid green when the panel is healthy.

Branch one, power is innocent. During the blank, the CPU LED is solid green and the meter on the panel terminals reads 23.6 V DC. Margin over the cutoff: 23.6 - 19.2 = 4.4 V. The backlight has every volt it needs and the CPU is running, so the dark glass and the dead touch are one problem: the input lock. The project check finds a batch-complete message box that can open seconds before the saver fires. A forced popup, a timed-out saver and a blind press on the OK position bring the screen back. The fix: close that popup on an event in the project, and bring the panel's release up to date while the project is open.

Branch two, power is the fault. During the blank, the panel terminals read 18.4 V DC while the supply's own output reads 23.8 V DC. That is below the cutoff by 19.2 - 18.4 = 0.8 V, so the backlight is legitimately off. The missing volts are in the wiring: 23.8 - 18.4 = 5.4 V dropped between supply and panel. Say the panel draws 1.0 A; your manual's power figure replaces that number in your own arithmetic. Loop resistance is 5.4 V / 1.0 A = 5.4 ohms, and that connection dissipates 5.4 V x 1.0 A = 5.4 W as heat. A terminal carrying 5.4 W runs hot, works loose over time, and loose terminals are a recognized fire cause on DC wiring, so isolate the circuit before you re-terminate it. Remake the bad lug or crimp, then re-measure at the panel with the backlight lit: expect a reading within a few tenths of a volt of the supply setpoint.

The fix in order, and the check

1. Photograph the rear LEDs with the screen dark.
2. Measure DC at the panel terminals with the screen dark, and compare the reading with the printed cutoff.
3. If power clears, inventory the project's popups, reproduce the race with a forced popup and a blind OK press, then close the popup automatically or keep it away from the saver window.
4. Update the panel runtime through the maker's documented update procedure, then read the version back on the panel.
5. If power points at the supply, substitute a correctly rated supply for three saver cycles, then repair grounding and wiring permanently.
6. Re-enable the saver and blank the panel three times in a row, waking it by touch each time. Three clean cycles on the settings that produced the fault is the check.
7. Only after all six: call the panel defective, and send support the LED photos, both voltage readings, the version numbers and the project file.

A panel you already condemned is worth a second look before it becomes a spare; why an HMI backlight fails again after replacement explains how a healthy panel takes the blame for a supply fault.

Recommendation: Treat every won't-wake callout as the two-reading job above. Photograph the rear LEDs and read DC volts at the panel's own terminals while the screen is dark. A solid green CPU LED and a reading above the printed cutoff, 19.2 V DC on the panels in this example, sends you into the project after a hidden popup and a release update. A reading at or below the cutoff sends you into the supply run and its terminations. Replace touch hardware last; on this fault pattern it is the part that comes back from the bench with the same fault.

Questions