Gob Weight CV Control: The Metric That Decides Forming Yield

0.4%. That's the gob weight CV number I want to see on a well-run triple-gob line, section to section, across a shift. Most plants I walk into are sitting somewhere between 0.6% and 1.1% and don't know it, because nobody is plotting the trend past the shift-end average someone scribbles on a whiteboard.

Ask a hot-end superintendent what gob weight CV was last Tuesday and you'll get a shrug. Ask what it is right now, section 3, and you might get an answer. That gap between tracking it and acting on it in real time is where forming yield quietly disappears.

In 2017 I was called into a two-furnace plant outside Cairo running an old Emhart AIS line with 1990s-era cam timing on half its sections. Gas curtailment had already hit that winter, and the plant manager compensated the only way he knew how: pushing the furnace hotter to keep pull rate up. Gob weight CV climbed from 0.5% to 1.3% in six weeks. Six weeks. Nobody connected the dots until check detector rejects had roughly doubled.

That's the thing about gob weight drift. It never announces itself. It shows up as a slow bleed in cold-end rejects that everyone blames on something else, usually the mould or the mix.

What 0.4% actually buys you

Gob weight coefficient of variation is the standard hot-end control statistic tracked per section on IS machines, and it deserves more attention than most plants give it. Well-run operations hold it below roughly 0.5-1.0%. Once you cross 1.5%, reject rates start climbing in a way that's hard to miss on the cold end.

The part that catches plants out is how fast the correlation shows up. A rise of just 0.3-0.5 percentage points in gob weight CV commonly produces a measurable uptick in check detector rejects and cold-end inspection rejects within the same shift, often before anyone on the floor has flagged a visible defect trend. By the time someone walks the line looking for stones or cords, the number already told you something was wrong two hours earlier.

Forming yield, good ware out of total gobs formed, is the number this all rolls up into. Pull a line from a CV of around 1.2% down to 0.6% and you're typically looking at a 1-3 percentage point improvement in forming yield on a high-speed triple-gob line. On a plant running six-figure tonnage a year, that's not a rounding error. That's EBITDA.

Gob weight CV isn't a quality metric. It's a leading indicator that quality hasn't caught up with yet.

The feeder bowl is lying to you more than the plunger

Everyone's first instinct is mechanical: plunger mechanism wear, orifice ring erosion. They're right to check it. Those two are the most common root causes of gob weight drift on narrow-neck press-and-blow lines, and orifice rings should come off on a defined tonnage or cycle-count schedule rather than waiting for a visible defect to force the issue.

And the cause that gets missed more often sits upstream of the plunger entirely: feeder bowl temperature stratification. Most plants aim for a spread of plus or minus 2-3°C across the bowl and never actually check whether they're holding it. When one side of the bowl runs cooler than the other, you get section-to-section gob weight variation that looks mechanical but isn't, and no amount of plunger rebuilding fixes it.

The defect modes tell you which one you're chasing. Underweight gob shows up as unfilled containers, incomplete parison fill. Excess glass shows up as a thick heel and settle-wave marks. Off-centre gob gives you double-gob events or baffle-mark asymmetry, and that one usually points straight back at the centring tool reading (and yes, I know your fitter says it's fine, check it anyway).

• Unfilled containers and incomplete parison fill point to underweight gob
• Thick heel and settle-wave marks point to overweight gob
• Baffle-mark asymmetry and double-gob events point to off-centre gob

Read the gob formation sequence carefully and you'll see why: the feeder is where shape and weight get set together. Fix one without checking the other and you've just moved the defect three sections down the line.

Ownership beats another sensor

Here's where generic consultancies get it wrong. An OEM-affiliated consultancy tends to frame CV drift as a hardware problem and price you a new plunger mechanism or feeder control retrofit before checking whether the real cause is thermal or procedural. A generic Lean or Six Sigma boutique hands you a standard SPC control chart that ignores viscosity-temperature sensitivity and refractory wear curves, so the control limits don't reflect where the furnace actually sits in its campaign life. That's the case for bringing in a vendor-neutral container glass consultant instead of an equipment vendor's audit team: nobody in the room has a plunger mechanism to sell you.

The gob weight CV trend should be owned jointly by the IS machine operator, who makes the real-time adjustment, and the forehearth or feeder technician, who owns upstream temperature and cullet ratio. Split those two without a shared dashboard and every correction cycle gets slower. I've watched it happen more times than I can count.

The 0600 handover is usually where it falls apart. On most lines I audit, the night-shift swabbing data doesn't make it into the morning briefing seven times out of ten. Not a furnace problem. A handover problem. That's exactly the gap the 9-stage Job Change Lifecycle inside the Job Change Tool is built to close, because job changes and mould swaps are disproportionate contributors to short-duration CV spikes. Plants that track a separate settle-in reject rate for the first 15-30 minutes after a changeover, instead of blending it into the steady-state number, catch this before it becomes a trend.

Look, the data says one thing and the floor says another, more often than any vendor slide deck admits. That's exactly why the correction has to be a role fix before it's a technology fix.

Why this bites harder when you can't just run hotter

In gas-secure markets, plenty of plants paper over forming variability by running the furnace a few degrees hotter and letting pull rate cover the mess. That option is closing fast in the Gulf. Gas curtailment episodes in Egypt in recent winters, combined with GCC fiscal reform pricing gas closer to its export opportunity cost rather than a flat subsidised rate, have made pull-rate stability and gob weight consistency a bigger yield lever there than in gas-secure European plants. You cannot out-heat a forming problem when the gas bill is watching.

Saudi Arabia's Vision 2030 industrial diversification push has brought new domestic glass packaging capacity online to cut import reliance for beverage and pharma bottles, and that new capacity is being commissioned and audited to a higher forming-control bar from day one, not retrofitted onto it later.

Europe and the US are feeling a different version of the same squeeze. EU ETS Phase IV is phasing down free carbon allowances for container glass manufacturing year on year, and the Packaging and Packaging Waste Regulation is tightening dimensional and weight tolerance requirements through 2026-2030, which means a CV target that used to be good enough won't clear the new spec. In the US, O-I and Ardagh have both been idling and consolidating furnace capacity, which raises the yield stakes on every tonne the remaining furnaces melt.

Wherever the plant sits, the finance-facing argument is the same one most consultants miss: a sustained CV improvement isn't only a quality win, it's a carbon-cost and compliance-cost lever too, because every rejected gob is glass you already paid the melting energy bill for.

Lean Glass's founder, Zaid Hassoneh, started on the floor at O-I Brisbane in 2005, made plant manager by 2019, and later ran the $220M USD Arglass Yamamura greenfield build in the US from 2019 to 2022. He's chased this exact drift down more forehearths than he can count, which is why the Job Change Tool treats gob weight CV as a cross-functional problem first and an equipment problem second.

None of this needs new capital equipment to start. It needs someone to plot gob weight CV by section, by shift, and actually look at it before the cold end tells you the story secondhand. That's the starting point of a proper forming audit, and it's usually where the cheapest EBITDA in the plant is hiding.