A machine finishes the first two operations of its automatic cycle and then remains on Step 30. The HMI says Waiting for transfer complete. No controller fault appears. The motor drive is available and the pneumatic cylinder appears to be in position.
An operator concludes that the PLC has frozen. Another person suspects a timer. A third suggests replacing the cylinder position switch.
The controller may in fact be executing its program normally. What is missing is the combination of conditions required for the next state transition.
This article explains how PLC scan behavior, state machines, interlocks, timers, edge detection, input feedback and diagnostic timing can create the appearance of a stuck sequence. The aim is to distinguish a controller execution problem from a deliberate, but unsatisfied, transition.
1. A PLC sequence is a controlled state machine
A machine sequence can be described as a set of steps or states and the conditions that permit movement between them.
An illustrative transfer mechanism might follow:
Step 10 — Ready → Step 20 — Position product → Step 30 — Extend actuator → Step 40 — Confirm transfer → Step 50 — Retract actuator → Step 10 — Ready.
Different PLC programming styles express these ideas in ladder logic, function blocks, structured text or Sequential Function Charts (SFC).
In each case, three questions matter:
- Which step is currently active?
- What does the controller expect to happen in that step?
- What condition or event enables the next transition?
A step remaining active does not automatically mean that the CPU has stopped running. Often the program is waiting for a transition that has not become true.
2. Separate a frozen CPU from a waiting machine sequence
The phrase PLC stopped can refer to very different conditions.
| Observation | What it may indicate | What it does not establish |
|---|---|---|
| Controller reports STOP, FAULT or another non-running mode | PLC execution status needs review | Which machine component caused the mode change |
| Controller remains in RUN; sequence state does not advance | Current transition may not be satisfied | CPU has frozen |
| HMI screen does not update | HMI communication or display issue is possible | Controller stopped executing |
| An actuator does not move | Command, power, valve, load or process condition may be missing | Program logic is defective |
| Process moves but confirmation is missing | Feedback, timing or sequence interpretation may be involved | The physical movement failed |
The investigation should begin with the status of the actual controller and the actual sequence, not the wording of a general operator complaint.
3. Understand the scan cycle before interpreting a signal
In a typical cyclic PLC application, a controller updates its process image, executes program logic and makes output values available according to its execution and I/O configuration.
Siemens S7-1200 documentation explains that, for automatically updated process-image I/O, physical inputs are copied to the process image before execution of the user program, and output values are updated through the process-image mechanism.
The exact sequence and timing depend on the PLC platform, task organization, interrupt events, module configuration and communication architecture.
That means a field signal does not necessarily become visible to every instruction at the same instant it changes physically.
It also means an engineer should not assume that a momentary input was available for a full scan. Some short events may be missed or handled differently depending on the input filter and task design.
Engineering principle: The controller can be functioning correctly while the logic does not observe the event that the machine sequence requires.
4. The transition condition is the central diagnostic question
A sequence transition is commonly expressed as a logical combination of conditions.
For example, a transfer step might be completed only when:
Transfer Complete = Actuator Extended AND Part Present AND Transfer Permissive AND No Process Fault.
This expression is illustrative, not a real machine program.
All four terms must meet the expected logic for the transition to occur. If one term remains false, the state may remain active while every other subsystem appears normal.
| Condition | Function in the transition | Possible diagnostic ambiguity |
|---|---|---|
| Actuator extended | Confirms required mechanical state | Sensor state may not match actual movement |
| Part present | Confirms process condition | Detection can vary with product position |
| Transfer permissive | Confirms prerequisite from another station | An upstream sequence may be waiting |
| No process fault | Prevents progression during a defined abnormal state | A latched fault may remain active after the visible symptom clears |
The useful question is not merely What is stopping the machine? but Which exact transition term is not satisfied, and what evidence supports that state?
5. Interlocks and permissives are related but different
Industrial control systems use interlocks and permissives to coordinate equipment and prevent unwanted or conflicting operation.
A permissive generally represents a condition required before an action can start or continue. An interlock represents logic that restricts an action when a defined condition exists.
The precise names and implementations vary by machine.
Examples include a transfer station waiting for space at the next conveyor, a pneumatic sequence waiting for a completed retract position, or a drive not receiving its normal operation command because another process condition remains invalid.
A machine may be waiting because the logic is doing exactly what it was designed to do.
Never treat an unsatisfied interlock as a reason to defeat it. The relevant task is to understand which condition it represents and why the expected state has not been established.
Safety-related protective functions are separate from ordinary process interlocks and require their own validated safety architecture.
6. Why timers can keep a sequence in one state
Timers are often used to delay a transition, confirm a condition remains stable or detect that a process has exceeded an expected duration.
A timer's behavior depends on the instruction type, the controller platform and the program's execution path.
The Rockwell Logix 5000 instruction documentation distinguishes timer enable, timing and done states, and the relationship between a preset and accumulated value.
An important conceptual point is that some non-retentive timers reset their accumulated time when their enabling conditions become false. A condition that repeatedly disappears before completion may keep the timer from reaching its preset.
| Example observation | Possible explanation |
|---|---|
| Timer repeatedly starts but never finishes | Enabling condition is unstable or repeatedly false |
| Timer completes but state does not advance | Another transition condition may still be false |
| Timeout occurs despite physical movement | Feedback arrives late, intermittently or is interpreted differently |
| Timer appears inactive | The program may not be calling or enabling the relevant instruction |
| Similar sequence timing differs by mode | The machine may use different parameters or program paths |
A timer is evidence about the logic's timing behavior; it is not automatically the root cause of a delayed process.
7. A short pulse is not the same as a maintained condition
Some programs expect an input level to be true for a period. Others expect a rising edge, meaning a transition from false to true.
An edge-triggered condition normally responds to an event, rather than to a signal that simply remains true.
For example, a device may send a short completion pulse while the sequence is in a different state. When the sequence reaches the step that expects the pulse, the signal is already inactive.
The equipment then appears to have completed its action, but the controller's transition condition was never satisfied.
The interpretation depends on the actual program:
- Was the logic expecting a maintained level?
- Was it expecting a new edge after the step became active?
- Was the event recorded or latched appropriately for the application?
- Could input filtering, task timing or communication timing affect its observation?
These are design and evidence questions, not a recommendation to change the program without an approved review.
8. The difference between a momentary signal and a latched condition
A momentary signal is present only while an event or physical condition exists. A latched state can remain set after the original event has disappeared.
Latching is useful for preserving events or controlling sequences, but can also make interpretation confusing if the set and reset conditions are unclear.
A process may wait for a reset that has not occurred. A completion bit may remain active from an earlier cycle. Or an alarm acknowledgement may clear the displayed message while a separate process-state condition remains latched.
A useful engineering report identifies both the event that sets a state and the documented condition that clears it.
Never assume that an acknowledged HMI alarm and an internal program state refer to the same memory or have identical behavior.
9. Program execution order can affect the observed result
PLC code often executes in a defined task or organizational order, although hardware interrupts and event tasks can change execution timing.
In a cyclic implementation, two instructions writing the same variable or state may interact in ways that are not obvious from a single network or block viewed alone.
A state assignment performed in one part of the program can be overwritten elsewhere during the same task cycle.
Similarly, the conditions for resetting a sequence may be processed after the conditions for advancing it, depending on the chosen architecture.
This is a reason to review the complete documented ownership of a state variable, not just the ladder rung where it appears to be set.
A correctly structured sequence should make state ownership, transition conditions and reset behavior understandable to engineers maintaining the system.
10. Multiple tasks and event routines can complicate timing
Not every PLC application is a simple, single-routine cyclic loop.
Controllers may include periodic tasks, event tasks, organization blocks and hardware interrupts. I/O update timing can also differ between local modules and communication-based remote I/O.
Siemens documentation describes process-image partitions and event-associated I/O handling. Rockwell documentation similarly explains execution and instruction behavior within controller tasks.
Consequently, the same physical event might be interpreted differently depending on where it is handled.
The root question is not whether the scan is "fast enough" in the abstract. It is whether the configured execution model matches the timing and sequencing requirements of the application.
Timing analysis should be based on platform-specific documentation and actual engineering records.
11. The HMI state label may not reveal the full cause
An HMI might show Waiting for Cylinder Extended even when the underlying transition also requires an upstream station to be ready.
Operator labels are designed for readability. They do not necessarily enumerate every boolean term in a program's transition condition.
A generic alarm can therefore be correct at the machine level but incomplete as a diagnostic description.
A well-designed diagnostic interface may provide separate information for the active sequence step, the expected transition and the specific unmet prerequisite.
The engineer should use approved program documentation and event records to establish which conditions the controller actually uses.
12. Reconstructed factory case: the step that never completed
Consider an illustrative automatic assembly station.
Step 20 advances a workpiece into position. Step 30 extends a pneumatic cylinder. Step 40 confirms the workpiece has been transferred and permits the next motion.
After a production change, the machine intermittently remains at Step 30. Operators can see that the cylinder reached its mechanical end position. The HMI still indicates that transfer completion is pending.
The initial diagnosis is a faulty cylinder or solenoid valve.
However, the reconstructed sequence records reveal that the cylinder-position signal became valid, while the product-presence input briefly dropped before the controller evaluated the full transition.
The transition required both conditions simultaneously. The cylinder's physical movement was not the missing element; the unstable product-detection condition was.
Lesson: The active sequence step identifies where progression stopped, not necessarily which mechanical component has failed.
This case is a deliberately reconstructed teaching example, not a report from an identifiable factory.
13. Event ordering explains why a machine looked frozen
The following timeline is hypothetical and uses simplified values to explain the logic.
| Relative event | Recorded or assumed state | Engineering interpretation |
|---|---|---|
| Step 30 becomes active | Extend command requested | Sequence is progressing |
| Cylinder completion observed | Extended condition becomes valid | One transition term is satisfied |
| Product detection changes to false | Product-present permissive is lost | Another required term is not satisfied |
| PLC continues executing Step 30 | Transition evaluates false | The state remains active by design |
| HMI displays waiting state | No overall transition completion | Operator sees a stuck machine |
| Process timeout later becomes active | Expected duration exceeded | Secondary symptom, not necessarily initiating cause |
This sequence makes the diagnostic problem specific. Instead of asking why the CPU froze, it asks why the product-detection condition was lost at that stage.
The source of that loss requires its own investigation of sensor, wiring, product position and process conditions.
14. A second case: a pulse arrives before the sequence is ready
Consider a different illustrative control system where a transfer-complete event is represented by a brief pulse from another station.
Station A reaches a waiting state after Station B has already issued its completion pulse. The step transition at Station A requires a fresh rising edge while the step is active.
As a result, Station A remains in the waiting state even though Station B has physically completed its operation.
This points toward a sequence coordination and event-memory design issue, not necessarily a bad sensor.
The corrective engineering objective would be to establish an agreed, documented method for transferring and acknowledging completion states across the stations. Specific implementation and verification depend on the machine's validated control architecture.
A transient signal and a persistent acknowledgment have different semantics; treating them as equivalent can lead to intermittent sequencing faults.
15. Distinguish field feedback from internal logic
When a sequence cannot advance, it is useful to classify the missing condition into one of three groups:
| Classification | Example | Important distinction |
|---|---|---|
| Physical feedback | Sensor reports position or product presence | Actual process and sensor interpretation may differ |
| Internal software state | A latched alarm, mode or recipe condition | Logic may retain information after a physical event clears |
| External system state | Another station or remote device grants permission | Communication or upstream sequencing may determine availability |
A PLC step may be blocked by any of these.
This classification prevents an investigation from focusing only on local hardware when the relevant condition may be internal to the program or supplied by a different controller.
16. A step-transition evidence matrix
For maintenance reporting, a useful matrix separates the requested step, the expected condition and the observed condition.
| Evidence field | Illustrative entry |
|---|---|
| Active step | Transfer confirmation |
| Expected next step | Return actuator |
| Transition requirement | Actuator extended AND product present AND receiver ready |
| Actuator position | Reported true |
| Product presence | Intermittently false |
| Receiver ready | Reported true |
| Timeout status | Becomes active after prolonged wait |
| First meaningful deviation | Product-presence condition drops |
| Remaining uncertainty | Whether sensor, product geometry or another factor caused that drop |
This is a documentation and analysis framework, not a prescription for changing live PLC states.
The most important entry is the first deviation supported by actual records. The eventual timeout may only tell the team that the process did not finish on schedule.
17. How fault logging can make sequences maintainable
A well-designed machine can record which step was active, why it entered the step and which prerequisite prevented the next transition.
The value of logging is not simply collecting more data. It is preserving enough context to reconstruct a failure after the machine has returned to normal operation.
Useful information can include:
- Current and previous sequence state.
- Transition condition or named unmet prerequisite.
- Relevant input or device state changes.
- Recorded timeout or fault classification.
- Operating mode and product or recipe version.
- Time ordering and the limitations of the logging interval.
Different controllers, HMIs and devices may use unsynchronized clocks. Millisecond-level conclusions should not be made from timestamps unless synchronization and recording accuracy are known.
18. Why changing a timer is rarely a complete diagnosis
A sequence that repeatedly times out may appear to be fixed after its timeout is increased.
That change does not automatically address a missing input, mechanical delay, communication condition or incorrectly defined transition.
A longer timer can alter the observed symptom while leaving the original mechanism unchanged.
Before a timeout is treated as unsuitable, the engineering investigation needs an explanation for the required process duration, acceptable variation and the effect of any proposed change on functional and safety requirements.
Timeout changes are design changes, not a shortcut for proving root cause.
19. Common PLC sequence troubleshooting mistakes
- "The machine is waiting, so the PLC has frozen." An unsatisfied transition may be intentional.
- "The cylinder is visibly extended, so the transition must be true." More than one condition may be required.
- "The HMI says waiting for a sensor, so the sensor is defective." Labels may summarize complex logic.
- "The timer did not finish, so it is broken." Its enable state may have changed.
- "An input pulse happened, so the PLC must have seen it." Scan, filtering and task timing influence event observation.
- "Acknowledging the alarm resets the whole sequence." HMI acknowledgement and internal state reset are distinct.
- "Increase the timeout and the problem is repaired." The original condition may still exist.
- "Force the missing bit to see if the sequence moves." Such intervention can defeat protections or cause unexpected machine motion and is not a valid general troubleshooting method.
- "The last visible alarm is the root cause." Event chronology may show an earlier deviation.
20. A structured review for authorized engineering teams
A useful analytical approach is to reconstruct the intended operation using existing documentation and approved records.
- Define the exact symptom. Identify the state that did not advance and the production conditions.
- Confirm execution context. Distinguish controller status from application sequence status.
- Identify the documented transition. Establish every condition required to leave the current state.
- Classify the missing condition. Separate process feedback, internal state and external permissions.
- Preserve event timing. Identify the earliest recorded deviation rather than the last alarm.
- Consider task and signal behavior. Review the meaning of pulses, latches, timers and relevant scan timing.
- Compare competing explanations. Determine whether the evidence supports sensor, mechanical, communication or logic-design causes.
- Document accepted findings. Any change must be managed and validated under the site's approved engineering process.
The purpose is to identify why the program remained in a state, not to make it advance regardless of machine conditions.
Technical conclusion
A PLC sequence that does not advance is often a control-system diagnosis, not a controller-hardware failure.
The key is to distinguish state, transition, permission, timing and feedback. A condition may never become true, may disappear too quickly, may remain latched incorrectly or may be delivered to a different execution context than expected.
Understanding the scan cycle helps explain how and when a program observes I/O. Understanding the sequence model helps explain what the program is waiting for.
When these two views are combined with reliable event history, a seemingly random frozen step can become a specific, testable engineering problem.
For related fault patterns, see Why Does a Machine Stop Without an Alarm?, Intermittent Conveyor Stops After Changeover and Why Does a Pneumatic Solenoid Valve Fail to Actuate?.
Authoritative references
- Siemens — SIMATIC S7-1200: Execution of the User Program: cyclic execution, process-image updates, interrupts and I/O timing.
- Rockwell Automation — Logix 5000 Controllers: Sequential Function Charts: sequence steps, transitions and time-based sequence behavior.
- Rockwell Automation — Logix 5000 Controllers: General Instructions: timer enable, accumulated time, preset and done-state concepts.
Cover photograph: industrial PLC and control-panel modules, by Raymond Sime, used under the Unsplash License. Cropped, resized to 960 × 540 pixels and converted to WebP. It illustrates real industrial automation equipment, not the reconstructed machine incidents.



