How to restart a machine safely after a long shutdown ART-427

Why downtime changes the restart risk

During a stoppage, equipment may look unchanged from outside, yet confidence in its present condition is lost. Room conditions, moisture and dust, a changed workshop layout, work by other contractors, disconnected utilities, or completed and unfinished service can all matter. Even a short note saying “the machine was switched off” does not answer the important questions: who stopped it and why, whether faults were active, what was disconnected, and whether the normal configuration was preserved.

Returning a machine is not the same as commissioning or troubleshooting it. An operator can collect facts, check permitted visible signs and follow the manufacturer’s defined startup procedure. They must not compensate for an unknown condition by changing settings, clearing alarms, moving safeguards or carrying out a “test” with guards removed. Hazardous-energy control is a formal procedure; the local lockout and return-to-service process takes precedence over workshop habit.

Separate three questions before switching on

First separate administrative readiness, physical condition and process readiness. This prevents the presence of electrical power from being mistaken for permission to cut.

The table does not set technical tolerances. Its purpose is to turn the vague phrase “it stood for a long time” into a list of checkable facts.

A practical sequence for a controlled restart

Begin with a handover record. State the machine identifier, time, reason for the stoppage, person responsible for restart and the source instruction. If a fault existed before the shutdown, do not mark it solved merely because the screen now shows no message. Ask service or the responsible engineer to confirm that the finding was closed.

Next make an external walk-around within the limits of the local procedure. Check the integrity of guards and accessible panels, the absence of foreign objects in the work zone, obvious leakage, and damage to accessible external cables or hoses. Do not open electrical cabinets, the laser-source enclosure, optical-path covers or service doors unless your role and instruction expressly permit it. Do not touch parts to assess their temperature.

Then confirm the work environment: no nearby activity is creating dust, moisture or obstructions or changing air flow; emergency passages remain accessible; and the gas and extraction arrangement has not changed. No visible problem does not prove that utilities are sound. They are checked only in the order defined by the manufacturer and authorised specialists.

Perform normal power-on only under the current instructions for the model. Do not add universal waits, pressures, temperatures or commands: they depend on configuration. During startup, record messages, unusual sounds, smells, indicators and the moment at which they appear. Do not clear alarms “to check”, change service settings, defeat interlocks or hold a sensor by hand. If the system does not complete its normal algorithm, stop at that stage.

Only after readiness is confirmed may a controlled test be performed under the internal procedure and documentation. Its purpose is not to produce an acceptable part but to confirm that the controlled restart completed without findings. Record the outcome, test material, program and observations. If there is a deviation, do not tune parameters by trial and error; pass the facts to service.

Information to prepare for service

A useful request begins with context, not a diagnosis. Provide the model and serial number when available safely; the date of last normal operation; reason and duration of downtime; all active messages; photographs of accessible external condition; work that was performed; and changes to power, room conditions or utilities. If logs of backups and service work exist, do not edit them retrospectively.

It is especially important to state whether downtime was preceded by an impact, axis fault, leak, burnt smell, unstable power or a safeguard problem. That turns a routine planned restart into a case requiring a different assessment. Do not conceal a “temporary” solution used before the stoppage: it may be central to safe diagnosis.

Common mistakes

• Starting with an urgent production order. The first restart must not be a way to catch up with the plan.
• Assuming a clear screen means the machine is sound.
• Bypassing guards or sensors to find a cause faster.
• Supplying gas, opening circuits or an electrical cabinet without the assigned authorisation.
• Changing cutting settings when it is not known whether the cause is technological at all.
• Failing to record who performed the restart and what was checked.

Treating the restart as a controlled event

Plan a return to service from the moment the stoppage begins. If downtime is expected, the responsible person records the job state, active messages, service actions completed, date of the next review and contacts of those who must confirm readiness. This is not bureaucracy: after several weeks, a shift rarely remembers why a cover was removed, who disconnected an auxiliary module or which symptom came before the shutdown.

For restart, define a short approval route: the operator collects facts and performs permitted actions, the supervisor confirms the work area is ready, and the service specialist decides on technical deviations. A person hurrying to close a production plan must not decide alone about guards, alarms or an unknown service state. Roles may be combined only where internal rules and qualifications expressly allow it.

Separate “it powered on” from “it was accepted back into operation”. The first status describes the system’s response to a normal command. The second means all required checks have been completed, no blocking finding remains and the control result was accepted under local procedure. Both events belong in the log; otherwise after a later failure it will be unclear whether the machine ever returned to a confirmed condition.

If restart is postponed, do not leave equipment in an intermediate state without a label. State exactly what is awaited: service, documentation review, access authorisation or correction of workshop conditions. Do not ask the next shift to “just try it”. Handing over an unfinished situation without facts often produces repeated restarts and loss of the original symptoms.

Passing a restart between shifts without losing context

Shift handover should end not with “the machine has not been started yet”, but with a short record that lets another person continue from the present safe status. It is enough to state when the previous stage ended, who performed it, which documents were reviewed, what was seen during the external inspection, whether findings remain open and who takes the next decision. If any of these facts is absent, that is not a minor gap; it is a reason to pause the next stage until it is clarified.

Use simple statuses understood by every role: “downtime recorded”, “awaiting document review”, “awaiting service decision”, “admitted to normal procedure”, “return to plan confirmed”. They do not replace a technical conclusion, but prevent a power-on fact from being confused with permission for production. Each status needs an owner for the next action and a time of last update. Do not use the status “done” while even one required record or specialist response is still outstanding.

At handover, separate confirmed facts from statements such as “it is probably fine”. Facts include shutdown date, available record of cause, visible signs, exact message text and service-ticket number. Assumptions include an explanation of why an event happened or an assessment of an assembly without inspection. An assumption may be written only as a question for service, not as a reason to change operating mode or close the event.

Checking whether a downtime record is complete

Before deciding on normal startup, the responsible person can review the downtime record in four groups. First is identification: which exact machine, when it was stopped and who made the record. Second is context: whether the stoppage was planned, whether deviations, service or changed workshop conditions preceded it. Third is status: what has been checked within permitted actions and which findings remain open. Fourth is decision: who must confirm transition to the normal procedure and where the outcome will be recorded.

This review is not a technical inspection and does not require access to enclosed equipment. Its purpose is to find missing facts before someone starts rushing because of the production plan. For example, “the equipment stood over the weekend” is not enough when nobody can say whether an unfinished service request existed before it stopped. Likewise, an old checklist does not prove that it concerns the current downtime. When in doubt, add a fact or escalate the question; do not infer an answer from a similar case.

The decision: production or escalation

The decision to return to production must be distinct from the decision to end downtime. Production may resume only when the non-technical control points required by the procedure are complete and the responsible role confirms that no blocker remains open. If a document is missing, it is unclear who closed a service finding, or an unexpected fact appears during the normal procedure, the correct outcome is escalation. That does not mean the machine is necessarily faulty; it means there is insufficient evidence for a safe decision.

An escalation message does not need a diagnosis. Give the machine identifier, short downtime chronology, facts already recorded, exact messages, photographs of accessible external condition if needed, and a specific question: may the next normal step begin, or is an inspection required? Service or the responsible engineer then receives material for a decision, while the shift does not turn uncertainty into an experiment.

After a decision, record its basis as well: reference to the normal procedure, name or role of the person confirming the transition, and the control-point result. This is not formal approval for its own sake; it prevents the next stoppage from starting with a search for why the machine was returned to plan. If confirmation was verbal, add a short note in the approved system instead of leaving it only in a shift’s memory.

Keep the decision record available with the downtime log and service ticket. Then supervisor, operator and service see one sequence of facts rather than three different verbal versions. If the responsible person changes, the new owner records acceptance of the open item separately.

Planned downtime versus other scenarios

Planned stoppage usually has a known record and a controlled state. Emergency loss of power, a safety trip, collision, water ingress or repair without a closed report are different scenarios. They must not be put through an ordinary restart checklist, even when the downtime was brief. One unexplained fact matters more than the calendar duration of downtime.

Do not try to define universally how many hours or days make a shutdown “long”. For one configuration room conditions may be decisive; for another, the type of work performed or utility status. The threshold, additional inspections and responsible roles are set by manufacturer documentation and the site safety system.

When immediate escalation is required

Do not continue a restart if there is smoke, a burnt smell, sparking, leakage, visible damage, abnormal noise or vibration, a failed guard, a safety alarm or unknown status after repair. Follow the site emergency plan, restrict access and notify the responsible manager and service. Do not resume operation until a competent person decides.

What cannot be determined without data

Without the model, serial documentation, shutdown history and inspection, it is impossible to say which checks, intervals, test program or parameters are permitted. A general article cannot confirm the condition of a laser source, optics, drives, cooling, gas system or electrical assemblies. That is the boundary where the manufacturer, authorised service or the engineer responsible for safety is needed.

Checklist before deciding to return to the plan

• [ ] The reason for and duration of downtime are recorded.
• [ ] Open service findings have been checked.
• [ ] The external inspection found no deviations.
• [ ] Workshop conditions and access meet local rules.
• [ ] Normal startup was performed only under the model documentation.
• [ ] All messages and observations were recorded.
• [ ] The controlled test is permitted by the responsible procedure.
• [ ] A service request was created for any uncertainty.

Need a service check?