One ripped belt, one copper quarter
BHP's operational review for the quarter ended 30 September 2026, released on 20 October 2026, records that production at its Copper South Australia segment fell 16% year on year to 60.7 thousand tonnes, "primarily due to the underground conveyor belt failure at Carrapateena" - described in the review's quarterly summary as a belt rip. A replacement belt was installed and was operational within seven weeks, followed by a progressive ramp-up through September, and the conveyor is now operating at full capacity. The failure had first been disclosed with BHP's FY26 results in August. Chief executive Brandon Craig's statement in the review lists the event in the group's own order of causes: "Copper production was lower in the first quarter, reflecting planned grade declines and scheduled maintenance, alongside temporary disruptions from severe weather in Chile, stops for safety and the conveyor failure at Carrapateena." Group copper production fell 19% to 400.7 thousand tonnes in the same quarter.
Restoration time is decided before the failure
The editorial judgement belongs up front: the procurement content of this disclosure is not the rip, it is the seven weeks. How long a single-point underground trunk conveyor stays down after a belt failure is not improvised in the week it happens; it is the sum of decisions taken earlier and elsewhere - whether a replacement belt is held in stock and in what lengths, whether rip detection stops the drive before a tear propagates, whether splicing crews, presses and cure capacity are on site or on call, and how a belt reel is moved through underground workings to the failure point. BHP does not disclose which of these set its seven-week interval, and this article does not guess. But by publishing the interval next to a production number, the review converts an internal spares debate that every mine has - how much belt inventory is enough - into a benchmark with a measured cost: one quarter, one segment, 16%. Buyers of trunk conveyors should treat the seven weeks as a question to put to their own operations, not as a standard to copy.
What the disclosure quantifies, and what it withholds
The quantified layer is unusually firm for a failure story, because it comes from a market disclosure rather than a vendor case study. Quantified: Copper South Australia at 60.7 thousand tonnes, down 16% on the same quarter a year earlier and 33% below the June quarter; group copper at 400.7 thousand tonnes, down 19%; an average realised copper price of US$7.01 a pound, up 53% year on year, which sets the value context of the lost tonnes without this article inventing a revenue figure BHP did not publish. Withheld: the root cause beyond the word "rip", the failed belt's width, strength class and length, the belt's supplier, whether the replacement came from site stock or new manufacture, the repair cost, and the exact split of the quarter's shortfall between the failure and the planned maintenance BHP also cites - the review says "primarily" the failure, and gives no tonnage for either part. Anyone quoting a cost for this event beyond the production percentages is adding numbers BHP has not released.
Guidance held, and the arithmetic behind it
BHP left full-year guidance unchanged after the quarter: 290 to 320 thousand tonnes for Copper South Australia and 1.65 to 1.80 million tonnes for group copper in FY27, with Copper SA unit cost guidance also held at US$0.30 to US$0.80 a pound. Holding guidance after a 16% segment decline is an arithmetic claim - that the remaining three quarters can absorb the loss - and it rests on two statements in the same review: the conveyor is back at full capacity after the September ramp-up, and the segment had buffers elsewhere, with Prominent Hill selling more concentrate than it produced by drawing down inventory built in the second half of FY26. Neither buffer replaces a trunk conveyor; both buy time around one. For buyers, the transferable point is the structure of the disclosure itself: a dated failure, a dated restoration, a quantified segment impact and an explicit guidance decision. That is the accounting standard to demand internally after any critical-belt failure, because it is the format in which boards and insurers will eventually ask for the event anyway.
A single event, not a loss statistic
This article stays in its own lane against two neighbours on this site. The FM loss-data article works from twenty years of insurer statistics across more than 300 mine sites - a probability layer, useful for weighting specifications across a fleet. The safety-devices article treats the protection chain - pull cords, alignment switches, belt-damage detectors - as one specification, the layer that tries to stop an event like this one. Carrapateena is neither: it is a single named event with a named owner, a dated restoration and a published production consequence, and its subject is the layer after protection fails - restoration. Rip detection belongs to the protection layer and appears here only as one of the pre-decisions that set restoration time; nothing in BHP's disclosure says whether detection was fitted or how the rip started, and this article does not retrofit a cause onto an undisclosed failure.
Five checks for trunk-conveyor owners
First, put the replacement question in writing for every single-point trunk conveyor: is a full replacement belt, or at minimum the longest single section, held in stock, where, and under whose release authority. Second, audit rip detection as restoration equipment, not only as protection: coverage, trip logic and the length of belt a tear can consume before the drive stops all feed directly into how much belt the replacement has to cover. Third, contract the splice before it is needed - crew, vulcanising press, cure time and underground access rehearsed as one exercise, because seven weeks is counted in calendar time, not in working days after mobilisation. Fourth, walk the replacement logistics: reel transport through the decline or shaft, handling at the failure point and the return path for the damaged belt, timed against the interval BHP has now published. Finally, adopt disclosure-grade downtime accounting for critical conveyors - failure date, restoration date, production consequence - recorded in the format BHP's review uses, so the next spares argument is settled with your own numbers, not with a competitor's quarter.
Manufacturer announcements and our analysis are distinguished. Product claims require verification for the specific application.
Sources
BHP - Operational review for the quarter ended 30 September 2026 (20 October 2026, full text) ↗BHP's official quarterly disclosure as lodged: Copper South Australia production of 60.7 kt, down 16% year on year, 'primarily due to the underground conveyor belt failure at Carrapateena, disclosed with the FY26 results'; a replacement belt installed and operational within seven weeks with ramp-up through September and the conveyor now at full capacity; the summary's wording 'the Carrapateena belt rip'; group copper 400.7 kt, down 19%; FY27 guidance unchanged (Copper SA 290-320 kt); CEO Brandon Craig quoted verbatim. Belt specification, supplier, root cause and cost are not disclosed.
Alliance News via Morningstar - BHP and Anglo American operational updates (20 October 2026) ↗Wire coverage of the same disclosure: total copper production down 19% to 400,700 tonnes against 493,600 tonnes a year earlier, the conveyor failure at Carrapateena listed among the temporary disruptions alongside Chilean weather and safety stops, and FY27 guidance reaffirmed. Used as independent corroboration of the event layer and the market reading of the review; all conveyor-specific facts rest on source 1.
Kalkine Media - BHP Q1 FY27 results: copper drops 19% (October 2026) ↗Second independent carry of the operational review: Copper South Australia output down 16% to 60.7 kt primarily because of the underground conveyor belt failure at Carrapateena, the conveyor repaired and back to full capacity within seven weeks, and group FY27 copper guidance of 1.65 to 1.80 million tonnes maintained. Corroboration layer only; no conveyor parameters beyond BHP's own wording appear in any of the three sources.
