Transformer-Rectifiers

The overview made the point that the A330 is an "AC aeroplane": the 28 V DC network is not generated, it is made from AC by the transformer-rectifiers (TRs). This article opens the DC half of the chapter by taking the four TRs in depth. The single fact that organises everything below is this: the four units are physically identical and interchangeable, yet they are rated differently, protected differently, and fail differently — because each sits in a different job. They are also the only "autonomous" power components on the aeroplane — each TR drives its own isolation contactor by internal logic, outside the ECMUs' control — and that one architectural choice runs through every protection rule in this article.

What is left to the neighbouring articles: the DC bus topology and tie-contactor logic (DC Network Transfer), the batteries and the BCL (Batteries and the BCL), the TR/battery abnormal procedures (TR and Battery Faults), and the ground-service 2PU contactor (Ground Service and Maintenance Bus).

1. Four identical units, four different jobs

FCOM names the family and, in its last sentence, states the autonomy principle:

"Two main Transformer Rectifiers TR1 and TR2 and one essential TR supply the aircraft's electrical system with DC current. A fourth TR is dedicated to APU start or APU battery charging. Each TR controls its contactor by internal logic."

Per FCOM DSC-24-10-20-20. The weight is in the last sentence: the TR contactors are the only power contactors not managed by the ECMUs, which manage every other AC and DC main contactor. Each TR judges its own faults and opens its own contactor.

The topology — who feeds each TR, and which buses each TR feeds — looks like this (per AMM 24-30-00 and 24-32-00 / 24-34-00):

  AC BUS 1          AC BUS 1          AC BUS 2          AC BUS 2
   (1XP)           via 3XC *          via 2PU           via 7PU
     │                 │                 │                 │
     ▼                 ▼                 ▼                 ▼
┌─────────┐      ┌──────────┐      ┌─────────┐      ┌─────────┐
│  TR 1   │      │  ESS TR  │      │  TR 2   │      │ APU TR  │
│ (1PU1)  │      │  (1PE)   │      │ (1PU2)  │      │ (1PU3)  │
│ forced  │      │ natural  │      │ forced  │      │ natural │
│  vent   │      │ convect  │      │  vent   │      │ convect │
└────┬────┘      └────┬─────┘      └────┬────┘      └────┬────┘
  5PU1│           3PE │            5PU2 │            7PU │
     ▼                 ▼                 ▼               ├──► 5PB ──► APU BAT charge
 DC BUS 1         DC ESS BUS         DC BUS 2           ├──► 5KA ──► APU starter
  (1PP)            (4PP)              (2PP)             └──► 309PP / 909PP
     │ 1PC1            │ 1PH              │ 1PN
     ▼                 ▼                 ▼
 DC BAT BUS       DC SHED ESS       DC SERVICE
  (3PP)            (8PP)             BUS (6PP)

 * 3XC selects the ESS TR source: AC BUS 1 normally; AC BUS 2 if
   AC BUS 1 is lost; the emergency generator (CSM/G) in emergency.
   4PC (recovery path): DC BAT BUS ──► DC ESS, used only when the
   ESS TR loss is NOT due to overcurrent and TR1 & TR2 are available.

Three reading notes fall out of the diagram:

• TR1 and the ESS TR both hang off the left network (AC BUS 1, the ESS TR via the 3XC source-selection contactor — see AC ESS Feed and Transfer), so the DC BAT BUS and the whole DC ESS family inherit the "left-heavy" bias of the AC side.
• The TR isolation contactors 5PU1, 5PU2 and 3PE are the self-controlled ones. The DC tie contactors (1PC1/1PC2) belong to the ECMUs, and the distribution contactors (1PN, 1PH, etc.) are unpacked in DC Network Transfer.
• The APU TR feeds nothing in the normal DC network. It idles on 309PP/909PP and only does work for APU battery charging (via 5PB) or APU starting (via 5KA).

The ratings: identical hardware, different numbers

Per AMM 24-32-00 §6.A and AMM 24-34-00 §6.A. The resolution of the apparent paradox is the cooling environment: TR1 and TR2 are installed where the aircraft ventilation system extracts heat, so the same hardware is rated to 200 A continuous; the ESS TR and APU TR rely on natural convection, so the same hardware is rated to only 100 A. A rating is a property of hardware plus installation, not of the box alone. Maintenance may swap any unit for any other; the rating then follows the slot, not the part.

[!warning]- "Four identical, interchangeable TRs — so they must all carry the same current."

Half right. The hardware is genuinely identical and interchangeable, but the continuous rating is not the same: TR1/TR2 are force-ventilated and rated 200 A, while the ESS TR and APU TR are convection-cooled and rated 100 A. The rating is a hardware-plus-installation property and follows the slot, not the box.

2. The autonomy principle and the cross-reporting wiring

The TR contactor chain in the normal configuration is fully specified in the 24-30 overview: TR1 feeds 1PP through 5PU1, and 1PP feeds 3PP through 1PC1; TR2 feeds 2PP through 5PU2 and 6PP through 1PN; the ESS TR feeds 4PP through 3PE and 8PP through 1PH. Of these, 5PU1, 5PU2 and 3PE — the TR isolation contactors — are driven by the TR's own internal logic, which is the physical embodiment of the FCOM line "Each TR controls its contactor by internal logic." The DC tie contactors 1PC1/1PC2 belong to both ECMUs; the rest of the distribution contactor logic is in DC Network Transfer.

Each TR carries a monitoring card with a dual supply (an internal 28 V plus an external sub-bus — the APU TR being the exception, internal only). The card has three output paths: to the SDAC (voltage / current / fault, for the ELEC DC page), to the CMS (a type-3 system: one discrete line reports FAULT, one receives the reset command), and to the ECMU. The ECMU path is wired crossed:

"TR2 supplies these data to ECMU1 and TR1 to ECMU2."

Per AMM 24-32-00 §5.A(4). Why crossed? (Integrative reasoning, not verbatim.) When a TR fails, the contactor the ECMU has to act on lives on the opposite side — losing TR1 means the ECMU has to close TR2's tie contactor 1PC2 — so feeding the opposite-side ECMU the local TR's death notice directly removes one hop from the reconfiguration command. It is the same crossover philosophy as "ECMU1 manages EPC B, ECMU2 manages EPC A" (see ECMU and Contactor Management).

3. The five protections

Each TR protects itself with five mechanisms. The thresholds are identical across units; only the inhibits and the reclosing behaviour differ by job.

Per AMM 24-32-00 §6.B. Three of these reward a closer look.

The minimum-current inhibit (protection 2). Its purpose is to catch "powered but producing nothing = the unit is internally dead". But when the input supply is lost, zero output is a physical certainty, not a fault — so the protection is inhibited in that case:

"The minimum current protection is inhibited for ESS TR in case of loss of input power supply."

Per AMM 24-34-00 §6.B(2)(b). The same inhibit applies to TR1 and TR2. The APU TR goes one step further: it normally idles at zero output (it only does work during charging or starting), so for it the protection is permanently inhibited.

The winding over-temperature reclose (protection 5). Above a 350 °C hot spot the contactor opens; once the temperature falls below 230 °C the reclose behaviour splits by configuration:

"recloses if a reset is performed in normal or essential configurations, recloses automatically in APU configuration."

Per AMM 24-32-00 §6.B(2)(e). So in the normal and essential configurations a reset is required before the contactor recloses; in the APU configuration it recloses automatically on cooldown. A NOTE adds that if the temperature sensor itself fails, the APU mode latches open and reports FAULT, while the MAIN / ESSENTIAL modes take no action. The APU TR's "reclose-by-itself, don't-report" privilege has a physical reason — APU starting is a large current surge and the windings will get hot — and it is the same philosophy expressed on the display side, where FCOM keeps the APU TR's parameters green during the start even when they look abnormal (see §7).

The reset path. Once a contactor latches open, reset is by CMS maintenance test or by a pushbutton on the unit:

"Reset can be done either with the CMS or directly with the RESET pushbutton switch behind the front grill of the TR itself."

Per AMM 24-32-00 §6.B(3)(a). The pushbutton is behind the front grill of the TR in the avionics compartment — a pilot cannot reach it — so an in-flight latch-open stays latched until landing. Even on the ground, using that button first requires the network to be powered:

"the aircraft electrical network must be energized (if the ground power unit is not available, energize the essential network only with the ELEC BAT pushbutton switches)."

Per AMM 24-32-00 §6.B(3)(a) NOTE (general principle also at AMM 24-00-00 §3.C).

[!warning]- "If a TR faults in flight, the crew can reset it from the flight deck and carry on."

No. Reset is only by the CMS maintenance test or by the RESET pushbutton behind the TR's front grill — and that button is in the avionics compartment, out of reach in flight. An in-flight TR latch-open is latched until landing.

4. The ESS TR overcurrent and the 4PC lock-out — the red line

The ESS TR's overcurrent provision is the most consequential rule in this article, because it locks out the very contactor that would otherwise rescue the DC ESS buses:

"In case of overcurrent detection (threshold approximately 900 A + or - 50 A, 1 s), the ESS TR contactor 3PE is locked open until a resetting is performed. Moreover the 28VDC BAT BUS/ESS BUS SUPPLY contactor (4PC) is locked open and the DC ESS busses are lost (transfer from DC BAT BUS not allowed). NOTE: The contactor 4PC serves to restore the DC ESS busses from the DC BAT BUS in case of ESS TR loss, if TR1 and TR2 are available (when ESS TR loss results from any fault other than overcurrent)."

Per AMM 24-34-00 §6.B(2)(c). This is the DC-side full version of the "overcurrent is not reconfigured" red line from the overview. An overcurrent says the short may sit on the DC ESS bus side — and in that case closing the DC BAT BUS onto it (4PC closed) would only keep feeding the short, and would drag the DC BAT BUS down with it. So the system seals the rescue route shut as well:

 ESS TR overcurrent (~900 ±50 A, 1 s)
        │
        ├──► 3PE locked open  ──►  ESS TR isolated
        │
        └──► (via relay 2PE)  ──►  4PC locked open
                                        │
                                        ▼
                       DC BAT BUS ──X──► DC ESS   (recovery barred)
                                        │
                                        ▼
                            DC ESS + DC SHED ESS LOST

FCOM gives the pilot-side note in plain terms:

"If a TR is lost due to overcurrent detection, reconfiguration does not occur and the related DC BUS is lost."

Per FCOM DSC-24-10-30-30. It is the same family of philosophy as the AC-side rule "generator overcurrent → lock the BTC" (see GCU and AC Generation Control): when a bus short is suspected, the network would rather drop the bus than send reinforcements into it.

One mechanism detail (the executing hardware is unpacked in DC Network Transfer): 4PC is not driven directly by the ESS TR but through a relay —

"the contactor 4PC is locked open via the relay 2PE."

Per AMM 24-35-00 §4.B(1). A terminology warning: the same 4PC is written "28VDC BUS 3 / ESS BUS SPLY contactor" in 24-30 and "28VDC BAT BUS / ESS BUS SUPPLY contactor" in 24-34 — BUS 3 (3PP) is the DC BAT BUS, so the two names denote one contactor. The symmetric object on the main-TR side is the ECMU locking 1PC1/1PC2 open to bar 1PP↔2PP transfer on a main-TR overcurrent.

[!warning]- "An ESS TR fault drops the DC ESS buses, but the DC BAT BUS can always recover them through 4PC."

Only conditionally. The 4PC recovery path works only for a non-overcurrent ESS TR loss (with TR1 and TR2 available). On an overcurrent loss, 4PC is locked open with 3PE, the DC ESS family is dropped, and transfer from the DC BAT BUS is not allowed — because the overcurrent points at a short on the DC ESS side, and feeding it would only keep the short alive.

5. Reconfiguration paths and the 7-second line

For the non-overcurrent losses, the TRs cover for each other symmetrically:

"(a) TR1 If there is loss of the TR1, the TR2 automatically restores supply to the DC BUS 1 and the DC BAT BUS. (b) TR2 TR2 loss leads to the symmetrical recovery of DC BUS 2 from the TR1. (c) In the event of ESS TR loss and if the TR 2 is available, the DC ESS BUS and DC SHED ESS BUS are automatically recovered by the TR1 through the DC BAT BUS and the DC BUS 1. If the TR 2 is not available, the DC ESS BUS and the DC SHED ESS BUS are no longer supplied and the bus 2PP is supplied by the TR 1. (d) TR1 and TR2 In the event of TR1 and TR2 loss, the DC ESS BUS and DC SHED ESS BUS remain supplied by the ESS TR."

Per AMM 24-00-00. The full spectrum, with the contactor hardware named:

Two counter-intuitive points the table makes precise. First, the ESS TR cover has a precondition — before TR1 can take on the ESS TR's job it must first have TR2 holding DC BUS 2, so with only one main TR left, the DC ESS family is dropped (and 2PP then draws from TR1). Second, with both main TRs lost but the ESS TR alive, the DC ESS family survives — the ESS TR is an independent third unit outside the main-TR fault chain.

The 7-second line is stated for the double-main-TR loss:

"In the event of TR1 and TR2 loss, DC BUS 1 and DC BUS 2 are immediately lost and DC BAT BUS after 7 seconds."

Per AMM 24-30-00 §3.B. What is that 7 seconds? (Integrative reasoning; the full mechanism is in Batteries and the BCL.) After both main TRs are lost, the batteries — through the BCL — bridge the DC BAT BUS for 7 seconds (FCOM's BAT pushbutton note: "In flight DC generation lost (limited to 7 s)"), giving the system a window to reconfigure or for the crew to act, then dropping 3PP to preserve battery charge.

[!warning]- "If TR1 and TR2 both fail, the DC ESS buses go with them."

No. On a double main-TR loss, DC BUS 1/2 are lost immediately and the DC BAT BUS is lost after 7 s, but the DC ESS family is still fed by the ESS TR — an independent third unit that is not on the main-TR fault chain. This "both mains lost, DC ESS alive" state is the hardware background to the DC BUS 1+2 failure (see DC Bus Faults).

6. The APU TR — the part-timer

The APU TR is the only unit of the four that holds a part-time job. It idles, watching 309PP/909PP (the APU bus segment), and does work in only two cases:

• Charging — when the APU BCL detects the APU battery below 26.5 V, the APU TR charges it through 7PU + 5PB.
• Starting — it supplies the APU starter through 7PU + 5KA, alone or coupled with the APU battery.

Its endurance is capped on the electrical side:

"The APU TR can supply three APU starts each hour with an interval of one minute between two starts, one after the other."

Per AMM 24-32-00 §6.A(3) NOTE. Read the wording carefully: this "three per hour, one-minute interval" is the AMM electrical-side TR thermal-capacity limit (the 3500 A / 1 s short-time rating exists to take the start surge); it is not the ATA-49 starter limit, which separately requires a wait after a failed start. The two limits coexist; the more restrictive one applies.

ATA-49 anchors the APU TR's starting role at the source, so it is not an inference:

"The APU battery supplies sufficient power to start the APU. If powered through the external power or through the main engine generators, the APU Transformer Rectifier Unit (TRU) supplies the power to start the APU."

Per AMM 49-00-00. So on the ground, with external power or the main generators available the APU TR does the starting; otherwise the APU battery does it.

One orphan property reinforces the part-time status: the APU TR's monitoring card is internally supplied only, and the ECMU does not reconfigure it — because, more fundamentally, the APU TR does not even report to the ECMU:

"The APU TR does not give a fault signal to the ECMUs. The TR1 gives a fault signal to the ECMU2 and the TR2 to the ECMU1."

Per AMM 24-32-00 §6.B(3)(a) NOTE (the FCOM side states the same as "The ECMU provides automatic reconfiguration (except for APU TR).", per FCOM DSC-24-10-30-30). If the APU TR is lost, it is simply lost — APU starting then falls back to the APU battery (the ATA-49 view).

[!warning]- "ELEC DC page shows APU TR at 25 V / 100 A on the ground — low volts, current pinned, it must be faulty."

No. While the APU is starting, high current with low voltage is the normal face of a surge load, and the APU TR's parameters are deliberately held green during the start even when they look abnormal. Read it as "APU starting", not "TR fault".

7. ECAM ELEC DC page and pilot scenarios

On the ELEC DC page (per FCOM DSC-24-20), each TR shows three fields:

• TR label — white (amber on fault or abnormal parameter).
• Voltage — green; amber if < 25 V or > 31 V.
• Current — green; amber on TR fault or < 2 A (this is the display face of protection 2).

The APU TR has its privilege noted above: during an APU start its parameters stay green. The ESS TR row carries an extra set of supply-source markers (AC 1 / AC 2 / EMER GEN — which source is feeding it, the visible result of the 3XC switching from AC ESS Feed and Transfer and Emergency Generator).

Five scenarios that tie the mechanisms to the flight deck:

1. Cruise, TR 1 FAULT. Check the reconfiguration on the ELEC DC page — DC BUS 1 and the DC BAT BUS should already be carried by TR2 (the 1PC chain). The procedure is in TR and Battery Faults; remember the TR cannot be reset in flight (the panel button is in the avionics compartment).
2. ESS TR FAULT with the whole DC ESS family amber. First question: is it an overcurrent loss (4PC locked out, DC ESS not recoverable) or an ordinary loss (4PC should restore it)? The ECAM procedure branches; knowing the mechanism keeps you ahead of it.
3. Double TR FAULT. Count the 7 seconds — the DC BAT BUS is about to drop while the batteries do the final bridging; the DC ESS family is still alive (the ESS TR is independent). This is the hardware background to the DC BUS 1+2 failure (see DC Bus Faults).
4. Repeated APU start attempts. The electrical-side limit is three starts per hour with a one-minute interval (APU TR thermal capacity), to be combined with the ATA-49 starter limit — the more restrictive applies.
5. ELEC DC page shows APU TR at 25 V / 100 A on the ground. The APU is starting — high current at low voltage is the normal surge picture, not a fault.

8. Dispatch view

Even the dispatch treatment follows the "identical hardware, different job" theme — the four TRs are not signed out the same way (per the operator MEL):

Per the operator MEL (TR dispatch items). The reading: the three network TRs (TR1, TR2, ESS TR) are all no-go — the main DC network and the survival floor admit no missing unit — while only the APU TR can be dispatched with conditions, because it is the part-time unit. The dispatch cost (per the operator MEL maintenance procedure): with the APU TR inoperative the ground APU can only be started on the APU battery (which then needs the APU battery at or above its minimum, otherwise the APU is treated as inoperative), and an in-flight APU battery start is not allowed. This is the dispatch-side extension of "identical hardware, different job": the job decides not only the rating, the protection set and the failure mode, but even whether a failure permits dispatch. The full picture across all electrical sub-systems is in MEL Dispatch.

A teaching image to carry the chapter: four quadruplet brothers, the job decides the treatment. Same genes (interchangeable). The two eldest (TR1/TR2) sit in the air-conditioned office (forced ventilation) and do double the work (200 A). The third (ESS TR) guards the vault (the ESS family) in an office with no air-conditioning (natural convection), so works a 100 A shift — but his troubles are the most serious, because a suspected break-in (overcurrent) gets even the spare key confiscated (4PC). The fourth (APU TR) is the on-call firefighter — sits idle most of the time (zero output, minimum-current protection permanently off), all burst power when the call comes (the 3500 A start surge, hot windings not counted as injury), but answers only three calls an hour.

Self-test

[!note]- Q1. Name each TR's source and customers.

TR1 ← AC BUS 1, feeds DC BUS 1 via 5PU1 (and DC BAT BUS via 1PC1). TR2 ← AC BUS 2 via 2PU, feeds DC BUS 2 via 5PU2 and DC SERVICE BUS via 1PN. ESS TR ← 3XC (normally AC BUS 1; AC BUS 2 if AC BUS 1 is lost; CSM/G in emergency), feeds DC ESS BUS via 3PE and DC SHED ESS BUS via 1PH. APU TR ← AC BUS 2 via 7PU, idles on 309PP/909PP and works only to charge the APU battery (below 26.5 V) via 5PB or to feed the APU starter via 5KA.

[!note]- Q2. How can the four TRs be "identical and interchangeable" yet rated differently?

The hardware is genuinely identical and interchangeable; the rating difference comes from the installation. TR1/TR2 are force-ventilated by the aircraft system and rated 200 A continuous (220 A with 55 °C + 24 g/s cooling air); the ESS TR and APU TR are convection-cooled and rated 100 A. The short-time overload limits differ too — the ESS TR holds 300 A for only 1 min and peaks at 1200 A / 3 s. A rating is a hardware-plus-installation property.

[!note]- Q3. List the five protections and the key inhibits.

(1) Overheat 175 °C / 5 s → locked open. (2) Minimum current < 1.5 A / 5 s → locked open (**inhibited on loss of input for TR1/2 and ESS TR; permanently inhibited for the APU TR**, which idles). (3) Overcurrent > 900 A / 1 s → locked open (TR1/2 only; the ESS TR has the separate 4PC provision). (4) Diode open/short / 5 s → locked open. (5) Winding hot spot > 350 °C opens, recloses below 230 °C (normal/ESS need a reset; the APU recloses automatically and does not report — the start-surge exemption; on a sensor failure the APU latches open with FAULT, MAIN/ESS take no action). Reset is by CMS or the RESET button behind the TR's front grill — in the avionics compartment, network must be energised, unreachable in flight.

[!note]- Q4. What makes an ESS TR overcurrent loss special?

Beyond 3PE locking open, the 4PC contactor (the DC BAT BUS → DC ESS recovery path) is locked open with it (via relay 2PE) — so the DC ESS family is dropped outright and transfer from the DC BAT BUS is not allowed. The reason: an overcurrent points at a short on the DC ESS bus side, so sending reinforcements would only feed the short and drag the DC BAT BUS down. 4PC acts as a recovery path only for a non-overcurrent loss with TR1 and TR2 available. Same family as the AC-side "overcurrent locks the BTC".

[!note]- Q5. Give the reconfiguration paths and the 7-second line.

One main TR lost → the opposite TR covers via 1PC1/1PC2 (ECMU dual control). ESS TR lost (non-overcurrent) → TR1 recovers DC ESS via DC BUS 1 → DC BAT BUS → 4PC, provided TR2 is still holding DC BUS 2. ESS TR + one main TR lost → DC ESS family dropped. TR1 + TR2 both lost → DC BUS 1/2 lost at once, the batteries bridge the DC BAT BUS for 7 s before it drops, and the DC ESS family stays alive on the ESS TR. The APU TR has no one to reconfigure it (ECMU exception list + internally supplied card).

Key takeaways

References

Per AMM 24-30-00 (normal contactor chain, double-main-TR loss / DC BAT BUS 7 s), AMM 24-32-00 (TR1/TR2/APU TR supply chains, interchangeability and the ratings table, the five protections and their inhibits, monitoring-card outputs and cross-reporting, the reset pushbutton, the three-starts-per-hour limit), AMM 24-34-00 (ESS TR sources via 3XC, the 100 A / 1200 A figures, the overcurrent 4PC lock-out), AMM 24-35-00 (4PC locked open via relay 2PE), AMM 24-00-00 (TR symmetric recovery, reset-button general principle / network energised), AMM 49-00-00 (APU TR vs APU battery start coupling); FCOM DSC-24-10-20-20 (TR family, internal-logic contactor), DSC-24-10-30-30/-40 (overcurrent-not-reconfigured note, ECMU reconfiguration except the APU TR), DSC-24-20 (ELEC DC page indications, APU-start green hold); and the operator MEL (TR dispatch items). The "quadruplet" analogy and the cross-reporting rationale are integrative synthesis of the cited sources, introducing no facts beyond them.

Independent study material, not an Airbus publication. Refer to current operator FCOM, FCTM, and QRH for operational use.