August 16, 2026 · 15 min read

14 CFR 33.70 / 33.75 / 33.1514 CFR 21.93 / 21.95 / 21.99AD 2026-16-13 / TC CF-2026-35

When the Clock Counted Hours and the Damage Counted Starts: The PW210 Turbine Exhaust Frame AD (2026-16-13) Through a 14 CFR 33.70 and Part 21 Lens

There is a paragraph in FAA AD 2026-16-13 that most readers will skip, and it is the whole post. Paragraph (g), Definition: "For the purpose of this AD, an 'engine start' includes the start of an engine for any reason, even if it is not followed by a flight." A regulator does not write that sentence unless somebody, somewhere upstream, was counting the wrong thing.

This is a story about three separate artifacts that should have existed and did not: a life limit on a major static structural part, a change classification on a material substitution, and an inspection interval anchored to the parameter that actually accumulates the damage. The FAA has now bought all three back in the field, at $125,765 an engine, on a fleet of 48.

1. The public record

On August 14, 2026, the FAA published final rule AD 2026-16-13 (Amendment 39-23443, Docket FAA-2026-7238, 91 FR 52487), effective August 31, 2026. It applies to all Pratt and Whitney Canada Model PW210A, PW210A1, and PW210S turboshaft engines. It supersedes AD 2026-13-09, which had been effective for exactly seven weeks. (Federal Register, 91 FR 52487)

The chain, in order:

| Date | Action | What changed | |---|---|---| | May 29, 2026 | Transport Canada Emergency AD CF-2026-23 | First mandate: repetitive visual inspection of the turbine exhaust frame | | June 29, 2026 | FAA AD 2026-13-09 (91 FR 38991), effective July 14 | Adopts CF-2026-23 by reference; declared an "interim action" | | July 15, 2026 | Transport Canada Emergency AD CF-2026-35 | Supersedes CF-2026-23; splits thresholds by frame material | | August 14, 2026 | FAA AD 2026-16-13 (91 FR 52487), effective Aug 31 | Adopts CF-2026-35; supersedes AD 2026-13-09 |

The unsafe condition in the June AD names the mechanism precisely:

"Operators reported circumferential cracks at the turbine exhaust frame external surface. A manufacturer investigation revealed that the cracks originated from a turbine exhaust frame wall found below the minimum required thickness, with subsequent crack initiation and growth caused by thermal stress from engine starts." (AD 2026-13-09, 91 FR 38991)

Then the August AD adds the second driver:

"A manufacturer's analysis revealed that turbine exhaust frames manufactured from a certain material introduced by P&WC Service Bulletin PW210-72-57123 are more susceptible to developing thermal stress cracks at an earlier stage compared to turbine exhaust frames made of Waspaloy, the material used prior to the issuance of P&WC Service Bulletin PW210-72-57123."

Read those two together. The first says a wall came out of the shop below drawing minimum. The second says the alloy was changed by service bulletin to something less durable. Those are two independent defect sources landing on the same part, and the second one was introduced deliberately, by a document whose entire purpose is to change the configuration of fielded engines.

The consequence, per paragraph (e) of both ADs: turbine exhaust frame failure, "which could lead to loss of axial containment, release of parts, and damage to the helicopter." The TEF is the rear structural case — it carries the aft bearing support of the power turbine and closes the axial load path between the engine core and the airframe mount. Lose it and the power turbine rotor is free to translate along its own axis at operating speed.

Applicability is 48 engines of U.S. registry, across the Sikorsky S-76D (PW210S) and Leonardo AW169 (PW210A) — offshore energy, HEMS, SAR, law enforcement, corporate. (FlightGlobal, June 5, 2026)

Costs, verbatim from the AD: inspection is 0.5 work-hours at $85 = $42.50 per engine, $2,040 fleet-wide. Replacement is 9 work-hours plus $125,000 in parts = $125,765 per engine, and the FAA states it "has no way of determining the number of engines that might need these replacements."

And the sentence that justifies emergency adoption: "At this time, the growth rate of the turbine exhaust frame cracking has not been determined." Compliance for affected engines is therefore 5 hours time-in-service or 20 engine starts, whichever occurs first.

2. The standards lens

This is an engine, so the governing airworthiness code is 14 CFR Part 33, not Part 25 or Part 29 — and Part 33 has its own vocabulary. Four clauses converge on this frame, and each of them names an artifact that is missing.

14 CFR 33.75 — Safety analysis. §33.75 requires an assessment of each failure and its probable consequence, classified as Minor, Major, or Hazardous Engine Effect. §33.75(g)(2) lists Hazardous Engine Effects, and the list explicitly includes non-containment of high-energy debris. "Loss of axial containment, release of parts" is a textbook Hazardous Engine Effect. §33.75(g)(2) then requires the applicant to show that such effects are predicted to occur at a rate not in excess of that defined as Extremely Remote — a probability range of 10⁻⁷ to 10⁻⁸ per engine flight hour.

Note the denominator. Per engine flight hour. The regulation's own probability budget is written in hours, and the damage mechanism here accumulates per start. That mismatch is not the FAA's error; it is a known limitation of the unit, and it is the applicant's job to convert. If your fleet-usage assumption says 1.2 starts per flight hour and your actual HEMS operator runs 3, your Hazardous Engine Effect rate is understated by 2.5x and nothing in the paperwork will tell you.

14 CFR 33.70 — Engine life-limited parts. This is the clause the AD is really about. §33.70 defines engine life-limited parts as "rotor and major static structural parts whose primary failure is likely to result in a hazardous engine effect." A turbine exhaust frame whose failure produces uncontained release is, by the plain text, a major static structural part in that category. §33.70 then requires an approved engineering plan, manufacturing plan, and service management plan, and the resulting life must be published in the Airworthiness Limitations Section of the ICA per §33.4 and Part 33 Appendix A.

If a §33.70 engineering plan existed for this frame, it had to define the damage-accumulation parameter, and for low-cycle thermal fatigue in a hot-section static case that parameter is start cycles, not hours. The AD's paragraph (g) definition — starts count even without a flight — is the FAA writing the missing line of that engineering plan into federal rulemaking, eleven years into the type's service life.

14 CFR 33.15 — Materials. §33.15(b) requires that materials "conform to approved specifications (such as industry or military specifications, or Federal Aviation Administration approved specifications) that ensure their having the strength and other properties assumed in the design data." A wall found below the minimum required thickness is a §33.15(b) and Part 21 Subpart G production-conformity finding, full stop. In supplier-quality terms it is an AS9100D §8.5.1 / §8.5.6 control-of-production and control-of-changes escape, and in AS13004 terms it is a Process FMEA whose wall-thickness characteristic was either not classified as a Key Characteristic or not controlled as one.

14 CFR 21.93, 21.95, 21.97 — Classification of changes. Here is the sharpest question in the whole file. Service Bulletin PW210-72-57123 introduced a different material on a part whose failure mode is Hazardous. Under §21.93, a change is minor only if it has "no appreciable effect on the weight, balance, structural strength, reliability, operational characteristics, or other characteristics affecting the airworthiness of the product." A material substitution on a hot-section structural case that demonstrably reduces thermal-fatigue life has an appreciable effect on reliability and structural strength by definition — after the fact. The engineering question is what the substantiation looked like before the fact, and whether the low-cycle fatigue analysis and the §33.70 life were re-run against the new alloy's LCF curve or inherited from Waspaloy's.

14 CFR 29.1309 and 29.901(c) — the installation side. The S-76D and AW169 are Part 29 / CS-29 transport-category rotorcraft. §29.901(c) requires that engine installation be designed so no single failure or malfunction, or probable combination, will jeopardize safe operation. A twin can lose an engine. It cannot as easily absorb high-energy debris crossing the nacelle into the tail boom, gearbox, or rotor system. That is an aircraft-level Catastrophic failure condition under the AC 29-2C §29.1309 framework, and it inherits from an engine-level Hazardous Engine Effect. The two-tier inheritance is exactly the interface where these things get lost.

MSG-3 / §33.4 continued airworthiness. The final gap: an inspection interval requires a crack-growth curve. The FAA states outright that the growth rate "has not been determined." An interval set without da/dN is not damage tolerance — it is a guess with a short leash. The 5-hour / 20-start threshold is the leash.

3. A worked snippet

3.1 Engine-level failure classification per §33.75

| ID | Failure condition | Engine effect (§33.75) | Aircraft effect (29.1309) | Required rate | Detection credited | |---|---|---|---|---|---| | TEF-FC-01 | Circumferential crack initiation at TEF outer case, external surface | None (latent) | None | n/a | Scheduled visual — not previously scheduled | | TEF-FC-02 | Crack propagates through wall, TEF partial structural loss | Major Engine Effect | Major (IFSD, single-engine continuation) | Remote | Vibration / crew — unverified | | TEF-FC-03 | TEF failure, loss of axial constraint on PT rotor | Hazardous | Catastrophic (debris into tail boom / rotor / cabin) | 10⁻⁷ to 10⁻⁸ per EFH | None | | TEF-FC-04 | As TEF-FC-03, on the SB-57123 material at reduced LCF life | Hazardous | Catastrophic | Same budget, higher realized rate | None |

TEF-FC-04 is the row that did not exist before July 2026. It is the same failure condition as TEF-FC-03 with a different occurrence rate, and the only thing that changed was a material introduced by a service bulletin. A safety analysis is not a document you write once; it is a document that a configuration change invalidates.

3.2 Fault tree — TEF structural failure

TOP: Uncontained release of PT rotor debris (Hazardous Engine Effect,
     inherits to Catastrophic at aircraft level per 29.1309)
      |
     [AND]
      +-- G1: TEF loses axial load-carrying capability
      |     |
      |    [OR]
      |     +-- B1: Through-wall circumferential crack, outer case
      |     |     |
      |     |    [AND]
      |     |     +-- B1a: Local wall thickness below drawing minimum
      |     |     |        (production conformity escape, 14 CFR 33.15(b))
      |     |     +-- B1b: Accumulated LCF damage from start-cycle
      |     |              thermal gradient exceeds material capability
      |     |              |
      |     |             [OR]
      |     |              +-- B1b-i : Waspaloy frame, N_start > N_allow(W)
      |     |              +-- B1b-ii: SB-57123 frame, N_start > N_allow(SB)
      |     |                          where N_allow(SB) < N_allow(W)
      |     +-- B2: Aft bearing support distortion / mount fitting failure
      |
      +-- G2: No detection barrier intervenes before through-crack
            |
           [AND]
            +-- B3: No 33.70 life limit published in the ALS for the TEF
            +-- B4: No scheduled external visual inspection task in the ICA
            +-- B5: No condition monitoring (HUMS/vib) allocated to this
                     failure mode

G2 is the interesting gate. Every barrier under it is an absent artifact, not a failed component. That is the signature of a paperwork gap rather than a hardware gap — and it is why the remedy is a $42.50 inspection rather than a redesign.

3.3 Why the unit matters — illustrative usage profiles

The AD gives one threshold, "5 hours time-in-service or 20 engine starts, whichever occurs first." Under a hours-only assumption those two are interchangeable. Under real utilization they are not. Illustrative, not sourced:

| Operator profile | Typical starts per flight hour | Hours flown when the 20th start occurs | Which limit binds | |---|---|---|---| | Offshore shuttle, long sectors | 0.5 | 40.0 | 5 hr TIS binds first | | Corporate / VIP | 0.8 | 25.0 | 5 hr TIS binds first | | HEMS, short scene flights | 2.0 | 10.0 | 5 hr TIS binds first | | Line maintenance week, ground runs | high, near-zero flight time | ~0 | 20 starts binds first |

The last row is the one paragraph (g) was written for. A maintenance organization doing avionics checkout, post-repair ground runs and hangar warm-ups accumulates the damage parameter while accumulating essentially no time-in-service. If your life limit is expressed in hours, that entire row is invisible to your tracking system. (Tech Times summary of AD 2026-16-13, August 14, 2026)

4. Derived requirements (excerpt)

PW210-TEF-001 — Damage-accumulation parameter The turbine exhaust frame shall be assigned an engine life-limited part life per 14 CFR 33.70, expressed in engine start cycles, with any hours-based limit stated only as a derived secondary limit. An "engine start" shall be defined as any light-off of the engine irrespective of subsequent flight, including ground runs, motoring checks, and maintenance runs. Rationale: the dominant damage mechanism is start-cycle thermal LCF; an hours-based limit does not bound it. Verification: LCF substantiation report plus ALS entry per §33.4 and Part 33 Appendix A. Parent: 14 CFR 33.70(a) engineering plan; AD 2026-16-13 ¶(g).

PW210-TEF-002 — Cycle counting to the airframe The engine control unit shall record a non-resettable cumulative engine-start counter, shall make it available on the maintenance data bus within 2 seconds of a valid start being declared, and shall latch a maintenance message when the counter reaches 90 percent of the applicable ALS start life for the installed frame material. Rationale: a life limit that cannot be read is a life limit that will be exceeded. Verification: HSI-level test plus 500-cycle endurance run with counter audit. Parent: PW210-TEF-001; 14 CFR 33.28 engine control systems.

PW210-TEF-003 — Material-differentiated life Where a service bulletin introduces an alternate material into a §33.70 life-limited part, the part shall carry a distinct part number and a separately substantiated life, and the change shall be classified under 14 CFR 21.93 as a major change unless an LCF substantiation demonstrates capability greater than or equal to the superseded material across the full certified temperature range. Rationale: SB PW210-72-57123 introduced a less durable material onto a part whose primary failure is Hazardous. Verification: change-classification record plus comparative LCF test at rated start transient. Parent: 14 CFR 21.93 / 21.95 / 21.97; 14 CFR 33.15(a).

PW210-TEF-004 — Wall thickness as a Key Characteristic The TEF outer case wall thickness at the crack-origin region shall be designated a Key Characteristic, shall be measured on 100 percent of production units by a method with gage R&R at or below 10 percent of tolerance, and results shall be retained for the life of the part. Rationale: the initiating population was a wall below drawing minimum — a conformity escape, not a design margin issue. Verification: AS13004 PFMEA and Control Plan review; AS9100D §8.5.1 audit; MSA study report. Parent: 14 CFR 33.15(b); 14 CFR Part 21 Subpart G.

PW210-TEF-005 — Interval substantiation No repetitive inspection interval shall be published for the TEF until a crack-growth rate has been characterized such that the interval provides at least two inspection opportunities between detectable crack length and critical crack length at the 95th-percentile growth rate. Rationale: the FAA states the growth rate "has not been determined"; the current interval is a containment measure, not a damage-tolerance interval. Verification: fracture-mechanics report with da/dN data at representative start-transient thermal loading. Parent: 14 CFR 33.70(a); AD 2026-16-13 justification for immediate adoption.

5. What the headline really tells us

The headline version reads like routine bookkeeping: FAA tightens inspection intervals on a helicopter engine part. Forty-eight engines. Forty-two dollars and fifty cents an inspection. Nobody has died. There is not even a public in-flight failure attributed to this mechanism yet — the whole file is preventive, built out of shop findings and a manufacturer's retrospective analysis.

That is exactly why it is worth reading. This is what the system looks like when it works late but not too late, and the artifacts it had to reconstruct in the field are the ones every program is tempted to skip.

The missing artifact is not an inspection. It is a §33.70 life, in the correct unit, on a major static structural part — and behind it a change-classification decision that let a material substitution ride into the fleet on a service bulletin without re-running the low-cycle fatigue substantiation that the original alloy's life depended on. Everything downstream is the cost of rebuilding that one number: the emergency AD, the seven-week supersession, the 5-hour leash, the $125,765 contingency the FAA cannot even scope because it does not know how many frames are affected.

The generalizable question has nothing to do with turboshafts. Take the component you are responsible for and answer two things. First: what unit does its damage actually accumulate in — hours, cycles, thermal transients, charge/discharge, actuations, boots, reconnects — and is that the unit your life limit, your service interval, and your reliability budget are written in? Second: when you last changed its material, supplier, or process, did you re-run the analysis that produced its limit, or did you inherit the number?

If the answers are "hours, because that is what the tracking system stores" and "we inherited it," you are running the PW210 file. You just have not had the shop finding yet.

The regulator had to write down what an engine start is. Somebody upstream should have written it first.

Sources

— Jherrod Thomas, The Lion of Functional Safety™