By the AeroKool Aviation team · Hialeah, Florida
MTBUR vs MTBF is the comparison between how often a unit is removed without a plan and how often it is confirmed to have failed, and the gap between the two numbers is a finding someone made, not a rounding difference. It is the gap, not either figure alone, that tells you whether removals are tracking real failures.
Both numbers are built from the same aircraft hours and the same removals. They diverge only because of what happens between the removal and the record: a unit comes off, a shop tests it, and a label is assigned. That label decides which number the event lands in.
This guide covers what each number measures, why the comparison is harder to read than it looks, where the gap comes from, and what a credible figure has to carry with it. It is written for the fleet engineer, reliability analyst, or buyer who has to compare figures from two sources. For the wider pattern behind the numbers, see why aircraft component repeat removals persist.
What MTBUR vs MTBF Actually Decides
Two Numbers, Two Questions
As the terms are commonly used in industry, mean time between failures asks how long a population of units runs between confirmed failures, and mean time between unscheduled removals asks how long it runs between removals nobody planned. Neither term is defined in the regulations or advisory circulars cited in this guide, so that sentence describes common usage, not a standard.
The distinction matters because a removal is a decision and a failure is a finding. A removal is made on the aircraft, often with incomplete information. A failure is confirmed later, on a bench, by someone applying a test.
In an MTBUR vs MTBF comparison, the useful reading is the relationship between the two figures, not the level of either one. A fleet can post a respectable MTBF and a poor MTBUR at the same time, and the distance between them says more about the process than about the hardware.
Who Reads Each Number
The reliability engineer reads MTBF to judge the hardware. The maintenance planner reads MTBUR to forecast spares, shop visits, and aircraft downtime. A buyer reads whichever one appears in the quote, usually without being told which one it is.
Each reader has a different purpose, and a number built for one purpose misleads when it is used for another.
FAA Advisory Circular 120-79A lists “Individual component failure rates” among its routine data and “Unscheduled parts replacement or unscheduled maintenance” among its non-routine data, and those two entries sit close to the two numbers. 14 CFR §121.373 requires the operator’s continuing analysis and surveillance system to review the maintenance programs and correct deficiencies, so whichever number the operator reads, the reading belongs inside that system.
Why MTBUR vs MTBF Is Harder to Read Than It Looks
The Definitions Are Local
Each operator defines which removals count as unscheduled, which failures count as confirmed, and which hours count as operating hours. Two operators can run the same fleet and publish figures that differ by definition alone. A figure without its definition cannot be audited.
Consider two operators reporting the same MTBF for the same part number. One counts every troubleshooting swap as a removal, and the other counts only units sent to a shop. Their MTBUR figures can differ widely while the hardware is identical.
The Denominator Moves the Number
The same removals divided by flight hours, flight cycles, calendar time, or time since installation give four different numbers. A unit on short sectors accumulates cycles quickly and hours slowly, and the denominator that suits it is not the one that suits a long-haul fleet.
The population matters as well. A figure can cover installed units only or installed units plus spares, and one part number alone or a whole family of dash numbers.
The Exclusions Change the Count
Some removals never enter the count: units removed to gain access to another task, units removed under a modification or directive, and units swapped during troubleshooting and later cleared. Each exclusion is reasonable on its own. Together they can move the figure more than any change in the hardware.
A defensible exclusion list is written down, short, and applied to every removal the same way. An exclusion applied only when the number looks poor is an adjustment, not a rule.
Confirmed Failure vs. Unconfirmed Removal: Where the Gap Comes From
What the Shop’s Finding Decides
Under the usual convention, a removal that the shop confirms as a failure lands in both numbers, and a removal the shop cannot confirm lands only in the removal count. The more removals the shop cannot confirm, the further MTBUR falls below MTBF. The gap is the unconfirmed removals, written as time.
The finding therefore has two readers. The operator uses it to classify the removal, and the analyst uses it to read the gap. A finding that says only “no fault found” serves neither, because it cannot be classified further and cannot be traced to a test.
That makes the shop’s finding the hinge. Why a no fault found finding is a test result explains why a finding of that kind records what the bench reproduced, not the condition of the unit.
Why a Narrowing Gap Is Not Always Good News
A gap that narrows can mean fewer unconfirmed removals. It can also mean the shop confirms more because it is testing differently, or that the operator assigns more removals to failure because that label is easier to use. Neither explanation shows up in the number.
Ask what changed in the process before crediting the hardware. Where removal reason codes come from shows where the label is born.
The Evidence a Credible Figure Has to Carry
A reliability figure is only as strong as the record behind it. Count, definition, population, and source have to stay linked, or the figure is a label.
The Written Definition
The definition names which removals count as unscheduled, what confirms a failure, and which hours form the denominator. It is written, dated, and attached to the figure whenever the figure travels. When the definition changes, the figure is restated or the break is marked.
Keep the earlier definition with the earlier figure. A trend line that crosses a change in definition shows the change, not the trend.
The Population and the Window
The figure names the part number, the dash numbers included, the aircraft position, the fleet, and the period. A figure for units in one position says little about the same part number in another. How a shop counts component shop visits covers the same question from the shop’s side of the counter.
The window matters because removals cluster. A quarter that included a modification campaign or a fleet-wide troubleshooting push does not describe a normal quarter, and the window should say what happened in it.
The Source of Each Removal Reason
Behind every removal sits a reason, and someone wrote it. 14 CFR §43.9(a)(1) requires a maintenance record entry to describe the work performed or reference data acceptable to the Administrator, and FAA Advisory Circular 120-79A lists “Component teardown reports” among the routine data an operator’s analysis draws on.
A figure is only as good as the findings that classified its removals. Ask whether each classification came from a teardown or from a label.
A classification assigned at the moment of removal and never revisited carries the guess of whoever wrote it. Revisit it when the teardown or test result arrives, and keep both entries so the change is visible.
Common Nonconformities and Failure Points
Reliability figures usually fail through comparison, not through arithmetic. The arithmetic is simple, and the inputs are not.
Numbers Compared Across Different Definitions
A predicted figure, an operator’s observed figure, and a shop’s reported figure may each be right and still not be comparable. They answer different questions about different populations over different windows. Setting them side by side without checking is a box-check, not an analysis.
This is where purchasing decisions go wrong. A quote that cites an observed figure from the shop’s own population, set beside an operator’s figure from a different fleet, looks like a comparison. It is two answers to two questions.
Coding Habits That Move the Rate
When removal reasons are coded loosely, the same fault appears under several descriptions and the rate looks lower than the problem. When they are coded tightly, the rate can look worse in the first year because the pattern becomes visible. The improvement is the visibility, not a decline in reliability.
Averages That Hide the Tail
A mean time between events hides how the events are distributed. A population in which a handful of serial numbers return repeatedly can show the same mean as one in which removals are spread evenly, and the two call for different responses.
Ask for the figure by serial number as well as by population. The serial-number view is where repeat removals become visible, and it is the view an average removes.
What a Reliability Number Does Not Settle
“A Higher MTBUR Is Always Better”
It is better only when the counting rules held still. A higher number can reflect fewer removals, fewer aircraft hours in the period, a wider definition of what counts as scheduled, or a decision to leave suspect units installed longer. Each of those moves the number without moving the hardware.
Read the number as a question, not an answer.
The opposite error exists too. A falling MTBUR can follow a better process, because a team that starts removing suspect units sooner and sending them for evaluation will log more removals before it logs fewer failures.
“A Predicted MTBF Tells You What Your Removals Will Be”
A prediction is a statement about a population under assumed conditions. Your fleet flies its own sectors, in its own environment, with its own installation and its own coding habits. The prediction is the baseline, and the observed rate is the evidence.
When the two diverge, the divergence deserves an investigation, not a verdict on the shop or the unit.
Questions to Ask Before You Accept the Figure
These questions turn a quoted number into evidence. Put them in the request for a reliability report, the vendor review, or the RFQ.
- Is the figure MTBF, MTBUR, or something else, and where is the written definition?
- Which removals are excluded, and who decided?
- What is the denominator: flight hours, cycles, calendar time, or time since installation?
- Which part numbers, dash numbers, and positions are in the population?
- Over what period, and did the definition change within it?
- Who classified each removal as a confirmed failure, and from what evidence?
- How many removals were unconfirmed, and how are they shown?
- Does the figure restate when the definition changes?
Do not read the list as a challenge to the source. A source that answers these questions quickly has usually done the counting properly, and one that cannot has told you something about the figure.
Start with the part number. If the unit is an air cycle machine such as the Embraer E-175 air cycle machine 1000700-7, that number is the first filter on the population a figure refers to. The published capabilities by aircraft and component type show which part numbers sit in the scope AeroKool Aviation publishes, and a part number outside the published list is a question to ask, not an assumption to make.
Where these figures sit inside a program is covered in building a reliability program for ECS hardware. Turnaround time as a separate measure covers the time a unit spends out of service, which neither number captures.
The published programs page lists repair and overhaul, PMA and DER, manufacturing, engineering, and repair management and logistics. The contact form is where the part number and the removal history go.
Frequently Asked Questions
What is the difference between MTBUR and MTBF?
As the terms are commonly used, MTBF measures the time between confirmed failures and MTBUR measures the time between unscheduled removals, whether or not a failure was confirmed. Neither term is defined in the regulations or advisory circulars cited in this guide, so ask the source of a figure for its written definition.
Is MTBUR always lower than MTBF?
Under the usual convention, where MTBUR counts every unscheduled removal and MTBF counts only confirmed failures over the same hours, MTBUR cannot be higher than MTBF. Figures from different sources often use different definitions, populations, and windows, so confirm the basis before assuming the relationship holds.
Why does MTBUR fall when MTBF stays the same?
Because removals are increasing that no shop can confirm as failures. That can come from troubleshooting swaps, from a fault the bench does not reproduce, or from a change in how removal reasons are coded. The hardware may be unchanged while the count of removals is not.
Does a regulation require an operator to report MTBUR or MTBF?
The sources cited in this guide do not require either figure by name. 14 CFR §121.373 requires continuing analysis and surveillance of the maintenance programs, and FAA Advisory Circular 120-79A describes the data such a system uses. Outside Part 121, confirm what your contract and maintenance program require.
Which number should you use to compare two repair stations?
Neither, until the written definitions, populations, and windows match. Ask each source for the definition, the denominator, the exclusions, and how unconfirmed removals are shown. A comparison built on different definitions measures the definitions.
How do no fault found findings affect MTBUR vs MTBF?
Under the usual convention, each unconfirmed removal shortens MTBUR without shortening MTBF. A rising share of no fault found findings therefore widens the gap between the two numbers, and it is a reason to examine the tests and the installation, not only the unit.
A reliability number is a claim about a population, and a claim without its definition cannot be audited. If the figure cannot show which removals it counted, which hours it divided by, and who classified each removal, it is a label with a decimal point.
About AeroKool Aviation. AeroKool Aviation is an FAA-certificated repair station in Hialeah, Florida, specializing in the repair and overhaul of environmental control system components — air cycle machines, cooling turbines, heat exchangers, refrigeration packages, valves and environmental control units — for commercial, regional, business jet, rotorcraft and military operators. The company was founded in 1959 and holds AS9110C and AS9100D registrations, with in-house PMA parts manufacturing and FAA DER repair engineering. Read more in the company profile.
This article is general technical information about how unscheduled-removal and failure-interval figures for aircraft environmental control system components are defined and compared. It is not approved data, it is not a maintenance procedure, and it does not replace the component maintenance manual, structural repair manual, service bulletin, airworthiness directive or other airworthiness data your organization is approved to use. Regulations, standards and manufacturer data are revised, and the requirements that apply to a specific job depend on the part number, the aircraft, the operator’s approved program and the scope of the repair station’s certificate. Verify the applicable requirements against the current manufacturer data, your own quality system, your operator’s continuing analysis and surveillance system, and your FAA principal inspector or EASA competent authority before performing or accepting any maintenance action.
Talk to AeroKool Aviation
AeroKool Aviation is an FAA-certificated repair station in Hialeah, Florida, overhauling air cycle machines, cooling turbines, heat exchangers, valves and environmental control units for commercial, regional, business and military operators since 1959, with an AS9110C quality system and in-house PMA and DER engineering. If you have a component on the bench, a repeat removal you cannot explain, or a part number you cannot source, we will tell you what falls inside our ratings and what does not.
Address
1495 S.E. 10th Avenue Hialeah, FL 33010-5984
eMail
aerokool@aerokool.com | sales@aerokool.com
Phone
305.887.6912 | 800.458.4255

