Machine cycle time: locating delays

Two cycles on the same machine can take different times even with identical panel settings. Before changing the program, separate motion from waiting for material, device confirmations and operator decisions. Average cycle time tells you there is a delay, but does not locate it. An investigation needs traces from several cycles and a record of the conditions under which they ran.

Define the measurement boundaries

Choose an unambiguous start and end. Part arrival and confirmed transfer to the next station may be suitable, depending on the machine. Do not compare time from pressing Start with time spent in the automatic sequence. Keep the product, recipe and operating mode alongside each measurement. Label cycles after a format change, during established production and after an alarm separately so that the comparison is meaningful.

Break the cycle into feeding, positioning, processing, checking and unloading, or the stages that actually apply. Identify each entry condition and completion signal. When devices run in parallel, adding their individual durations gives a misleading total. Determine which completion allows the next step to begin. This exposes the bottleneck rather than selecting the slowest-looking drive and assuming it limits output.

Motion time and logical waiting

An active step does not necessarily mean movement. A conveyor may be waiting for space at the next station, while a gripper waits for position feedback. Compare the PLC command, device response and sequence transition. Distinguish physical travel time from delay before starting and waiting after the movement has finished. These are different problems, even when the operator sees the same stationary machine.

For a sensor, check whether the signal is stable and whether the logic expects a level or an edge. Brief dropouts, detection alignment and filtering can affect confirmation. This is not a reason to shorten the filter immediately. Compare the recorded signal with the required device behaviour and involve the person responsible for machine inspection. Keep process conditions and safety functions distinct throughout the investigation.

Handshakes between stations

Where another controller supplies data, establish what ready, request and complete mean, including when each state is set and cleared. One side may wait for a new confirmation while the other retains the previous cycle's status. An active network connection does not, by itself, prove that the process handshake is correct. Look at the actual sequence of exchanged states.

Check communication diagnostics and whether received values are current. When traces come from different devices, establish how their clocks relate. Otherwise, an apparent delay between request and response may be a timestamp difference. Agree interface changes with both sides. Adding a timer in one controller can hide an unclear handshake while leaving restart behaviour unresolved.

Record enough to explain the delay

A useful trend contains the sequence step, relevant confirmations, drive state and reason for waiting. Stop counters distinguish one exceptional interruption from a short delay repeated every cycle. Keep minimum and maximum stage times as well as averages. Averages can conceal an intermittent missing signal that an operator clears manually and never reports as a separate fault.

Collect normal and delayed cycles for the same product. Match the logging interval to the event: a sample every few seconds cannot explain a brief wait between steps. Recording every available tag is rarely helpful either. Select signals to answer a defined question, and include addresses, units and a description. A small trace with clear context is easier to interpret than a large unexplained export.

Change one identified cause

If the evidence points to the sequence, PLC program changes can address a specific condition with a measurable acceptance criterion. Save the current project and agree on a rollback method before downloading changes. Modify one identified element, then measure again under comparable conditions. Changing several timers together makes it difficult to establish which change produced the result.

Verification should include normal production, missing material, interrupted confirmation and restart after stopping. Do not reduce cycle time by removing safeguards, bypassing interlocks or cutting a device's required operating time. Where on-site trials are needed, plan commissioning and upgrades with production and maintenance. Bring the sequence description, representative traces and the circumstances in which the delay occurs.