If the E-stop cuts the control power that feeds the PLC, and the operator panel draws from the PLC's supply, both restart together on release and the panel can stall at its welcome screen. Recover with a controlled power cycle once the machine is safe, then move the panel's feed out of the E-stop's path through a documented design review.
You press the E-stop, the machine stops as it should, you release it, the run button starts the machine again, and the operator screen sits on the welcome page. A full power cycle brings the screens back. That pattern points at power, not at the screen: the E-stop is cutting the control power that feeds the controller, the panel draws its power from the controller's supply, and both devices restart together on release. This article walks the safe recovery, the meter readings that confirm where the fault sits, the supply arithmetic that rules a sagging rail in or out, and the design change that stops the restart from happening at every stop.
Why does the screen stop at the welcome page after an E-stop?
Because the E-stop drops the 110 V control power that feeds the controller, and the panel draws its power from that same feed, so on release both devices restart together and the panel's startup stalls before it can load its screens.
The welcome screen is the boot stage the panel passes before it starts asking the controller for live values. A panel that boots clean after an ordinary power-down and power-up is telling you its hardware and its project are intact. What changes between a clean start and a hung one is the way power went away and came back, not a setting on the screen. The machine can be ready to run while the panel is still inside its own startup, which is why the run button works and the screens do not.
What is the safe way to get the screen back right now?
Once the machine is in a safe state, drop the feed, wait until the supply output reads 0 V on your meter, restore it, and let the controller finish booting before you judge the screen; a full controlled power cycle clears this hang in the arrangement described here.
Run the recovery the same way every time:
- Bring the machine to a safe state using its normal stop and reset procedure, and never bypass or jumper the E-stop to do these checks.
- Open the disconnect that drops the control feed, and lock it out under your site's procedure.
- Meter the DC output at the panel's terminals and wait for it to read zero, so every device finishes shutting down.
- Restore the feed and let the controller reach its normal running or idle state first.
- Confirm the panel passes the welcome screen and shows live values before anyone presses run.
- Log the date, the readings and the outcome. The log is what the controls engineer needs when the permanent fix gets planned.
What does each symptom point to?
Four causes cover this hang, and each leaves its own signature; check them in the order listed here: a feed that never returns, a rail that sags on the shared restart, a controller that is not ready in time, and a dirty restart with no voltage fault.
- The screen stays dark or keeps rebooting. The panel's feed never came back. Meter at the panel's terminals: a steady zero reading points to an open fuse, terminal or contact downstream of the E-stop relay. Correct the feed first, because no panel completes its startup without its supply.
- The panel lights, then hangs while the rail reads low during the ramp. The shared supply is sagging while the controller and the panel start together. Measure during the restart, not after it settles; the same supply drop that shows up elsewhere as a dim screen can stall a boot here, and the supply drop behind a dim screen is the same physics.
- The panel boots clean but shows stale values or communication errors. The panel came up before the controller did. Compare the controller's mode indication with its normal power-up state, and read the boot-order version of this failure in PLC COM timeout after power-up.
- Ordinary power-ups are clean, every E-stop restart hangs, and the rail holds its voltage through the ramp. The abrupt cut itself is the cause. That is the mechanism in the next section.
How do I measure the control supply through an E-stop?
Meter three points in one sequence: the 110 V AC control feed before the stop, the same feed while it is pressed, and the 24 V DC at the panel's terminals as power returns; that pattern separates a supply fault from a restart fault.
Use a meter rated for the voltages in the panel, and have a qualified person make the readings with the panel guards in place. The values below belong to the arrangement this article describes, a 110 V AC control circuit feeding a 24 V DC rail; take the points and voltages from your machine's drawings before probing.
- 110 V AC present before, 0 V while pressed, back on release: the E-stop is doing exactly what the wiring says, interrupting shared control power.
- A steady rated voltage before, 0 V while pressed, then a clean climb back and holding: the feed is healthy, and the hang is restart behavior.
- A rail that climbs but settles low, or dips when the controller pulls in: the rail is marginal on a shared restart, and the worked example below turns that into numbers.
After release, check the controller's mode indication before blaming the screen. A panel cannot show live values while the controller it reads is still coming up.
Why does a normal power-up work but the E-stop restart hang?
A normal power-down lets every device close its own work, so the next start is clean; an E-stop cut collapses the feed in tens of milliseconds, about one AC cycle at 60 Hz, and the panel restarts out of a brownout it never saw coming.
The milliseconds are fixed arithmetic: one cycle at 60 Hz is 1 divided by 60 seconds, which is 16.7 ms. A switch-mode supply is commonly specified to hold its output for one missing cycle, so a held E-stop, which is seconds long, collapses the rail and cuts everything on it mid-operation. On release the rail climbs from 0 V while the controller is starting too, and the panel's boot has to complete with both its supply and its data source in flux. That is the window a welcome screen sticks in.
An abrupt cut under load is also how a panel ends up worse than hung: some lose the project file and come back to a blank setup. If yours ever does, follow the recovery for a panel that loses its project after a power cycle.
Worked example: is a 60 W supply big enough for a shared restart?
Add the running currents, check the margin against the rating, then measure the rail during the ramp: a controller at 1.2 A plus a panel at 0.8 A is 2.0 A at 24 V, which is 48 W and leaves 12 W on a 60 W, 2.5 A supply.
Here is the arithmetic, with example ratings. Read the real ones off your devices' plates before you trust the result.
- Supply: 24 V DC at 2.5 A, which is 2.5 A × 24 V = 60 W, fed from the control circuit.
- Running loads: controller 1.2 A, panel 0.8 A, so 1.2 A + 0.8 A = 2.0 A, and 2.0 A × 24 V = 48 W.
- Margin at full running load: 60 W − 48 W = 12 W, which is 12 W ÷ 24 V = 0.5 A.
Twelve watts is thin for the moment both devices start together, because startup current runs above running current on each of them. So the measurement beats the paper number: meter the rail during the shared restart. If the rail dips while the controller pulls in, the sag is real and the supply needs resizing for the measured startup current, which is the method in sizing a separate 24 VDC supply for HMI inrush current. If the rail holds its rated voltage through the ramp and the panel still hangs, the supply passes and the cause is the dirty restart, not the watts.
Does the E-stop have to shut the controller down?
No: IEC 60204-1 sorts stop functions into three categories, and the stop has to remove power from the hazard regardless of the controller's state; nothing in that requires the controller's own supply to drop with it.
The categories describe how the motion is de-energized: a Category 0 stop removes power immediately, a Category 1 stop brings the motion to a controlled halt and then removes power, and a Category 2 stop leaves power available for a controlled stop. The risk assessment decides which one the machine needs, and the E-stop function has to do its job whatever the controller's program or network is doing.
The arrangement that avoids the restart is a familiar one: the E-stop contact drives a safety relay whose contacts drop the power to the outputs that cause motion, while the controller and its inputs stay energized. All inputs keep working, so the controller still sees the E-stop chain and can show the event on the panel; only the motion-causing outputs die. Motion energy is still removed the way the design requires, and nothing in the cabinet reboots.
What is the permanent fix?
Keep the E-stop's job of removing motion energy and take the controller's feed out of its path, and make that change only through a documented design review, because rewiring a safety circuit is not a maintenance task.
The correction in practice:
- Have the controls engineer review the stop design against the machine's risk assessment and set the stop category the machine needs.
- Rewire so the E-stop drops only the power that causes motion, and the controller, its inputs and the panel stay fed. Where the design keeps a cut of control power, give the controller and the panel a feed from a line the E-stop does not drop, with the supply mounted in the cabinet on a rail; a rail-mount DC supply is the usual mount for that.
- Update the drawings and the machine's safety documentation so the circuit on paper matches the circuit in the cabinet.
- Validate the full sequence before production: press the E-stop and confirm hazardous motion stops the way the design requires; release and reset by the machine's procedure; confirm the controller stayed up and the panel never left its screens; confirm that restarting still needs the intended operator action.
- Record the readings and device states at every step, so maintenance can tell a designed stop from an intermittent fault.
If the risk assessment does require the controller's supply to drop with the stop, the review ends with the restart as correct behavior, and the recovery in the second section above becomes the standing procedure for every stop.
Recommendation: treat the controlled power cycle as the recovery and the rewiring as the fix. Once the machine is safe, cycle the feed, watch the rail fall to 0 V, and let the controller boot fully before judging the screen; then have the stop design reviewed so the E-stop drops motion power only and the controller and panel stay energized through every stop. Until that review is done, expect the hang on each E-stop release and keep the log.
Questions
Can I keep clearing the screen with a power cycle and leave it there?
A controlled cycle is the recovery, not the repair. Drop the feed once the machine is safe, let the supply output fall to 0 V, restore it, and let the controller boot before judging the screen. Keep a log of date, stop count, readings and outcome; the hang returns on every E-stop release until the design changes, and that log is what the controls engineer needs to scope the rewiring.
If the panel had its own power feed, would the hang go away?
Only partly. A separate feed gets the panel booting clean, but the controller it draws its screens from would still restart on every E-stop release, so the pair still has to recover together before live values appear. The change that removes the symptom is keeping both the controller and the panel outside the feed the E-stop drops, while motion power still gets cut.
Who decides whether the controller stays powered through a stop?
The machine's risk assessment does, applied by the controls engineer in the design review. The review weighs the hazards, sets the stop category, and documents the rewiring before anyone touches the cabinet. Until that review is signed off, the operator's job is to run the standing recovery procedure and keep the log, never to rewire the circuit for convenience.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.