CUB Mould Design: The Standard Australian Plants Needed

It's April 2015 and I've got three plants on a call, all insisting they run cavity-up-bottom (CUB) mould changes the same way. They don't. One plant's cavity-up reading was another plant's cavity-down, and nobody had noticed until a mould set travelled between sites and cost us most of a shift chasing a fit problem that was never really a fit problem.

That's the thing about mould design standards. Everyone assumes they're shared until a set moves, or a new hire from a different plant reads the gauge sheet and asks why the orientation mark is on the wrong side. CUB sounds like a small thing. It isn't, and it's a good lens for a bigger problem: most Australian plants I've audited have excellent people and no shared language for the mould itself.

Why 'cavity up' meant three different things on three lines

The root cause wasn't the mould shop. It was documentation, or the lack of it. Two of the three plants had CUB orientation written into laminated cards that lived in the section operator's back pocket. The third had it in a supervisor's head. When that supervisor was on leave, the 0600 handover missed the orientation note entirely, because it had never been written down as a locked spec, just passed on verbally shift to shift.

The mould shop and gauge technician caught it eventually, but not before three sections ran out of tolerance for the better part of a changeover window. This is a shift-handover problem wearing a mould-design costume. I see it constantly (and yes, I know your fitter will tell you the team all knows the convention by now, ask him to write it down anyway).

What a CUB standard actually has to lock down

A real CUB standard is not a drawing filed away in the mould shop. It's a locked spec that ties orientation to the finish tolerance, the gauge points, and the thermal behaviour of the mould halves, because those three things drift together. Finish (thread/lug) diameter on standard beverage ware needs to hold to roughly ±0.10-0.15mm, with overall height inside ±1.0-1.5mm. Miss that band and the hot end can pass ware that a capping line rejects two hours later.

Checkmarks at the heel and side seam are the defect that exposes a bad CUB read fastest. Generic troubleshooting charts point straight at plunger or blank timing, but on the plants I've walked, the actual cause is more often a mould temperature differential exceeding 15-20°C across the parison mould halves. That differential shows up when the CUB orientation is inconsistent between sections, because swab cycles and cooling air get set up assuming a mirror-image mould that isn't actually a mirror image on that line.

Swab cycle is worth a mention here too. Plants that swab on a fixed time interval are managing the wrong variable. Gob count drives mould surface temperature faster than a clock does, so the better practice is one swab application per three to five gobs on high-speed sections, adjusted by mould material. Bronze plunger and neck-ring inserts dissipate heat faster than cast iron, which matters more on lightweighted, high-speed SKUs running 90-140 containers a minute than it does on heavy traditional ware.

A CUB standard that only lives in the mould shop isn't a standard. It's a filing system nobody on the floor can read.

The mould-shop bottleneck nobody budgets for

Australian plants aren't alone in underinvesting in mould engineering headcount while SKU count climbs. I saw the same pattern on a 2022 audit of a GCC container glass operation, where private-label and craft-beverage SKUs were multiplying faster than the mould shop could keep pace with. Saudi Arabia's Vetro Arabia, the Saint-Gobain/Obeikan joint venture running a furnace at Yanbu with capacity above 200,000 tonnes a year, sits in a region where job-change frequency has risen sharply without a matched increase in in-house mould-engineering capability, which stretches changeover downtime well past European benchmarks.

Egypt's export-facing plants face a related squeeze. Furnace campaigns there typically run eight to ten years, in line with global norms, but mould-shop capability is the recognised bottleneck limiting how fast a line can turn over SKUs. It's the same story I've watched play out across Australian multi-plant footprints: the furnace and forming end get the capital attention, and the mould shop runs on whoever's available that week.

But that's exactly where a CUB standard earns its keep. If the orientation, the gauge points and the thermal setup are locked and versioned, a mould-shop technician who's never touched that specific set can pick up the gauge sheet and get it right the first time. Without that lock, every changeover becomes a small reinvention.

Standardising CUB across a fleet without freezing the mould shop

The fix isn't more paperwork. It's fewer, better documents that are actually locked. On the Australian rollout, we ended up standardising three things across every plant in the fleet:

• Cavity orientation marked identically on every mould set, with no plant-specific variant
• Gauge check points recorded against the same reference datum, not a local convention
• Swab cycle intervals tied to gob count per section, not a fixed clock interval

This is precisely what the Job Change Tool's SKU Library is built to hold. Every recipe, mould set and CUB orientation versioned and locked, so the changeover isn't running on whichever fitter's notebook made it to that shift. Mapped against the mould-change stage of the 9-stage Job Change Lifecycle, with a named mould shop owner and a hot-end superintendent who signs off recipe lock, best-practice plants get like-for-like finish changes under 20-30 minutes per section. Plants without off-line pre-fit and gauge-checking are still running 45-60 minutes or more, and most of that gap is orientation confusion, not mechanical difficulty.

Any consultant selling you a generic job-change benchmark without asking whether your fleet runs bronze or cast iron inserts, or whether your mould orientation is actually the same across plants, hasn't run this problem before. That's the case for working with a vendor-neutral container glass consultant rather than an OEM-affiliated shop that frames every defect around its own equipment upgrade path. Lean Glass founder Zaid Hassoneh ran into this exact orientation-drift problem across multiple sites before it had a name, well before the Job Change Tool existed to catch it early.

Twenty-three minutes. That was the changeover time once the CUB standard was locked and the gauge sheet matched across every plant in the fleet. Not a mould problem. A documentation problem wearing a mould-shop costume.

If your plants are still running mould orientation off whoever remembers it best, that's worth a proper look before it costs you a shift the way it cost us in 2015. Our forming audit is where we usually find this first, well before it shows up on a KPI board.