A due date falls in three days, the aircraft is on a turnaround far from its base, and the planner offers the tolerance he has been applying for years to maintenance programme tasks. This time the answer is no — and not because the airworthiness manager is being pedantic. That particular line was never written to maintain anything: its interval is a term in the safety demonstration that made it possible to certify the aircraft, and the operator has no control over that term.
Two lineages, one single task card
At the technician’s workstation, nothing distinguishes the two families. Same format, same rules? Same card format, same reference in the production system, same certification signature. Yet two independent chains led there, and they do not obey the same rules.
The first is the one everybody knows. The maintenance working groups run every significant item through a decision logic: is the function evident to the crew, does the failure have an effect on safety, is there an applicable and effective task, at what interval. The result, consolidated by the maintenance review board, becomes the basis of the manufacturer’s recommended programme, then the operator’s maintenance programme. That chain is alive: in-service feedback exists precisely to make it evolve, task by task, under the control of the oversight authority.
The second chain does not go that way. During certification, the system safety assessment evaluates each failure condition and demonstrates that its probability remains below an objective all the more severe as its consequences are serious. When that demonstration only holds on the condition that a hidden failure be discovered within a certain time, the assumption becomes a requirement. The task that follows from it is not a maintenance recommendation: it is an operating limitation attached to the type certificate, of the same rank as a maximum speed or a limiting weight. Nobody negotiates a maximum speed with their planning department; the status is the same, only the form differs.
Some maintenance tasks come from safety calculations during aircraft certification. These tasks can’t be changed by airlines because they’re tied to the aircraft’s certification. Other tasks come from regular maintenance logic and can be adjusted based on reliability data.
On an Airbus A320, a certification-derived task requires checking the standby hydraulic system every 1,000 flight hours. This interval comes from safety calculations showing that waiting longer would increase the risk of undetected failures. An airline can’t extend this interval even if inspections never find issues, because the task exists to prevent hidden failures, not to fix wear.
What the task is looking for: the latent failure, not wear
An ordinary maintenance task fights a physical process — wear, corrosion, fatigue, a setting drifting out. It makes sense even if one knows nothing of the architecture: something is degrading, so one goes and sees how far it has got. The task born of certification fights nothing of the kind. It restores nothing. It does one single thing: it resets ignorance to zero.
The reasoning fits in three sentences. A redundant architecture is only worth something if its branches are genuinely available on the day they are needed. Now a standby branch can fail without anyone noticing: no warning, no perceptible effect in flight, nothing in the technical log. Redundancy then becomes a fiction, and the failure condition no longer needs two independent events to occur — the first is already there, and it is waiting.
The time during which it waits is exactly the time elapsed since the last check. If the hidden failure can occur at any moment within the interval, the aircraft flies, on average, about half of that interval with redundancy that is only apparent. Doubling the interval doubles that average exposure. This is not the caution of a careful drafter, it is an arithmetic property of the arrangement.
From this follows the form of these tasks. They are overwhelmingly functional or operational checks: the function is put into operation and asked to make itself known. One is not looking for an incipient defect, one is looking for a response. And some of these checks are not entrusted to maintenance at all: they appear in a normal procedure, carried out by the crew before the first flight of the day. Testing a warning circuit or bringing a standby means into operation may, in certain cases, be the item credited by the demonstration.
The right reflex. When a normal procedure contains a check that visibly serves no purpose for the day’s flight, do not file it away as routine: find out which demonstration it belongs to before a documentation revision simplifies it away.
The interval is a variable of the demonstration, not an operational parameter
For a failure condition that requires the combination of a hidden failure and a second event, the calculated probability brings out the product of the failure rates and the exposure time. The check interval is there, inside the product. At an unchanged failure rate, lengthening the interval increases the contribution of that term proportionally; shortening it reduces the contribution. The relationship is that of a product, and it is no more negotiable than the result of a multiplication.

What gets lost in operations is that certification did not approve a task: it approved a pair, a task and its interval. Keeping the line in the programme while changing the second term amounts to preserving the façade of a requirement from which what made it valid has been removed. And the operator, in all good faith, cannot know what margin remains: the published value comes out of a model in which other contributions have already consumed their share — common modes, maintenance errors, flight duration assumptions. The apparent margin of one line is not the margin of the system.
Why it is not shortened indefinitely either
The symmetry deserves to be stated, because it explains why the published value cannot be tightened unilaterally either. Every intervention has its safety cost: opening a panel, disconnecting a connector, removal and refitting, restoring the configuration, a check that leaves a system in an unusual position. Multiplying the checks on a standby function means multiplying the opportunities to leave it inoperative. The interval retained is the bottom of a valley, not an arbitrary ceiling.
The candidate task, and the trade-off that makes it disappear
The analysis does not directly produce requirements: it first produces a list of candidate tasks. Before freezing them, they are compared with what the recommended programme already provides for. If a maintenance task detects the same hidden failure, with the same effectiveness, at an interval no greater, creating a separate requirement would be a costly duplicate: the candidate is declared covered and disappears from the final list. The saving is real. So is the trap.
For the covering task now has two reasons to exist, and only one is visible from the maintenance programme. It appears there as an ordinary task, with its history, its results, its periodicity — and nothing indicates that a whole section of the safety demonstration rests on it.
The trap. A reliability review proposes to lengthen that task because it never produces a finding. That is exactly the expected result: it is watching for a rare event, and its zero discovery rate is proof that the system is working, not the argument for an extension. If nobody knows that it covers a candidate task, the coverage disappears without any signal being triggered, and the demonstration silently becomes false.
| Criterion | Maintenance programme task | Certification-derived task |
|---|---|---|
| Origin | Decision logic, working groups, maintenance review board | System safety assessment, under type certification |
| What it targets | A degradation process, or a function whose loss would be evident | A latent failure that would make redundancy fictitious |
| Governing document | Operator’s maintenance programme, based on the recommended programme | Airworthiness limitations section of the instructions for continued airworthiness |
| Revision of the interval | Possible on a reliability case, within the approval of the programme | Beyond the operator’s reach: request to the design holder, decision of the certification authority |
| Due date tolerance | The one the approved programme explicitly provides for | Only the one the limitations document itself provides for, sometimes none |
| Overrun found | Deviation handled by the management system, restoration of compliance and cause analysis | Non-compliance with a limitation: the aircraft no longer meets the conditions of its certificate |
| What an audit comes looking for | The technical justification and the follow-up of results | Traceability of accomplishment at the exact interval, without interpretation |
What “asking for a deferral” really means
Three doors exist, and only one is open to the operator without delay. The first is the tolerance that the limitations document itself provides for. Some manufacturers mark explicitly, task by task, what admits a slip and what admits none; where the marking exists, it is read, not interpreted. In its absence, the absence of tolerance is the rule, not a drafter’s oversight.
The second door is the revision of the limitation itself. It goes through the design approval holder, who alone has the model and can redo the analysis with the new value; the certification authority approves; the result is a new issue of the document, binding on the whole fleet concerned. The lead time is typically counted in months, never in hours on a turnaround. The third door, the individual exemption granted to one operator, exists in certain frameworks, but it is narrow, formal, and never opens the day before the flight.
One confusion deserves to be cut short: the minimum equipment list is not a way out. It deals with an identified inoperative item, with the conditions and limitations that the demonstration provided for that precise case. A check that has not been done is not an inoperative item: it is an unknown state, and nothing in the list covers the unknown. The two situations resemble each other — in both cases one is considering flying with something missing — but one is a constructed tolerance, the other a hole in the construction.
The right reflex. Before even drafting a request, trace the task back to its source document and check three things: which section it comes from, what tolerance that document provides for it, and whether the indirect approval procedure your organisation may hold does or does not cover these tasks — exclusion is the usual situation. Then date the last actual accomplishment, not the one in the plan.
The neighbours in the same section
The limitations section does not house only the tasks born of the safety assessment. It brings together families of very different origins, united by one point in common: their content belongs to the approved design, not to the operator’s programme. Confusing them leads to applying the logic of one to the other.
| Family | What sets the due date | What a deviation produces |
|---|---|---|
| Life-limited parts | A fatigue justification, supported by test | Mandatory removal at the due date, with no tolerance in principle |
| Damage tolerance inspections | A crack propagation rate and a probability of detection | A defect can reach its critical size between two visits |
| Tasks born of the safety assessment | The admissible exposure time to a hidden failure | Redundancy becomes fictitious with no signal to announce it |
| Critical design configuration control limitations | No due date at all: a configuration to be preserved at every intervention | A repair that looks unremarkable can re-establish an ignition path |
| Wiring and systems ageing | A zonal analysis and fleet in-service experience | A diffuse degradation, with no identifiable triggering event |
The fourth line is the most misunderstood, because it has no due date at all. A critical design configuration control limitation is not scheduled: it is complied with every time the zone concerned is touched, and it survives every revision of the programme. It is passed on through the accomplishment documentation and through personnel training, not through a schedule. It is also the only one that can be destroyed by a repair correctly carried out but poorly informed.
What to go and check in your own documentation
Open your maintenance programme and look at how the origin of the tasks is carried there. Many organisations have flattened the limitations section into the same repository as the rest, for the sake of consistency: the line survives, the status disappears. If the source does not appear on the task sheet, the first blanket escalation will carry it away without anyone having decided any such thing. Then ask the design holder for the status of candidate tasks covered by programme tasks: that document exists, it is rarely in the operator’s hands, and it is what protects against unintentional extensions.
The real grey area, however, is neither there nor in the hangars: it is in the cabin and on the flight deck. Checks credited to the crew appear in no maintenance schedule and feed into no reliability indicator. A procedure revision motivated by preparation time can remove one link of the demonstration without a single continuing airworthiness process noticing. This is the most fragile point of the whole arrangement, and it is handled by a single rule: any deletion of a check from a normal procedure goes through a question to the design holder. Exchanges between operators on these trade-offs make useful reading on the professional forum.
For anyone preparing a type rating or a technical interview, there is a shift of view to make: the limitations section of a manual does not contain only performance figures and weights. Some of its lines are check intervals, and knowing where they come from changes the way one answers the question “why this value?”. The preparation courses that cover continuing airworthiness make it a subject in its own right, and rightly so.
The above describes a mechanism and has no value as an enforceable reference: the applicable text, the approved limitations of the aircraft and the operator’s documentation prevail in all circumstances over this article.
Questions fréquentes
Can the programme's schedule tolerance be applied to a certification maintenance task?
No, unless the limitations document itself provides a tolerance for that task. The general tolerance of a maintenance programme applies to tasks arising from maintenance logic, not to those that constitute an operating limitation of the type certificate. In the absence of an explicit mention, it must be taken that no slip is permitted.
How can these tasks be recognised in an already consolidated maintenance programme?
You have to go back to the source: the task sheet must carry the originating document, and not only an internal reference. Manufacturers generally mark these tasks in the airworthiness limitations section, with an indication of what does or does not admit a slip. If your repository has merged all origins into a single numbering scheme, ask the design holder for the cross-reference table.
Does a history with no findings at all justify lengthening the interval?
No, and this is the most frequent misreading. These tasks watch for a rare failure whose discovery would already be an occurrence; a zero finding rate is the normal result, not a sign of over-maintenance. The justification for an extension goes through the safety model, and therefore through the design holder, not through the operator’s accomplishment statistics.
What should be done if the due date is exceeded while the aircraft is on a turnaround?
The aircraft no longer meets the conditions of its certificate and cannot return to commercial operation on that basis. The minimum equipment list does not cover this case, since it is not a matter of an identified inoperative item but of an unknown state. The only ways out are accomplishing the task on site or a formal decision by the authority, which is not obtained within the day.
Who can change the interval of an airworthiness limitation?
The design approval holder makes the request after redoing the safety assessment with the new value, and the certification authority approves it. The operator and its continuing airworthiness management organisation have no means of doing so, including through an indirect approval procedure for the programme. The result, once approved, takes the form of a new issue of the limitations document, binding on all operators concerned by that document.
Is the crew affected by these tasks?
Yes, more often than is believed. Some checks of hidden functions are carried out by the crew within normal procedures, and the safety demonstration takes them into account in the same way as a workshop task. Any proposal to lighten a normal procedure should therefore be verified with the design holder before being applied.

Leave a Reply
You must be logged in to post a comment.