Lehr Profile: The Cold-End Decision The Hot End Forgets

Here's a belief I'll argue with: most plants think their toughest quality problem lives at the hot end. It doesn't. It lives in a 40-70 minute tunnel most people walk past without a second look.

The lehr does the one job nobody in the plant gets excited about. It takes ware at forming temperature and brings it down through the strain point slowly enough that residual stress doesn't lock into the wall. Get that wrong and the ware looks perfect leaving the lehr. It'll fail in a warehouse in Ohio three days later, and someone will blame the pallet.

Why cold-end checking gets blamed on everything except the lehr

Cold-end checking is delayed stress cracking. It typically shows up 24 to 72 hours after the ware has already left the line, sitting in a warehouse or on a truck. Because it doesn't appear at the inspection station, it gets logged as a logistics problem, a palletiser problem, sometimes a glass-strength problem. It is almost never traced back to the annealing curve that actually caused it.

I audited a GCC-region bottled-water line in 2022 where this exact pattern had been running for months. Returns were climbing. Cold-end team blamed the coating batch. Warehouse blamed handling. Nobody had pulled the zone-3/zone-4 setpoint logs against the job change history, because nobody owned that comparison. When we did, the drift lined up with a container-weight change three weeks earlier that had never triggered a lehr re-qualification.

That's the pattern I see plant after plant: the IS machine job change gets re-qualified rigorously — gob weight, timing curves, mould temperature — and the lehr gets treated as a fixed asset that just runs. It isn't fixed. It's a per-job variable, and it's the single most common blind spot operators self-report when I ask them directly.

The two temperatures that actually matter

Soda-lime container glass has an annealing point around 548-560°C and a strain point near 505°C. Everything about lehr profile design sits between those two numbers. The soak at or near the annealing point relieves stress; the controlled ramp down through the strain point is what locks the relief in — or doesn't.

Shift the soak setpoint by 10-15°C off that band and you can push centre-wall stress readings out of spec with zero cosmetic sign. No baffle marks, no stones, nothing an inspector would flag by eye. Target residual stress at the centre wall normally sits under roughly 100 nm/mm on a polariscope reading, with edge and finish allowances running higher, around 150-200 nm/mm. Rush the ramp through the strain point and the ware exits looking flawless while carrying stress that'll open up a crack somewhere between here and the customer's filling line.

Lehr residence time typically runs 40-70 minutes across six to ten independently zoned burner sections. Heavier or thick-walled containers need more dwell near the annealing point before that ramp starts. Lightweighted NNPB containers need less dwell but tighter tolerance control, because there's less glass mass to buffer an error.

The ware that fails three days later in a warehouse was never a forming problem. It was a decision made in the lehr that nobody wrote down.

Coatings are the tell nobody reads correctly

Hot-end coating goes on just before lehr entry, typically a monolayer tin oxide applied around 550-600°C substrate temperature. Cold-end coating, usually polyethylene-wax based, goes on at lehr exit once ware has cooled below roughly 120-150°C. The entire anneal cycle sits sandwiched between those two windows.

Get the entry or exit temperature wrong and neither coating bonds the way it should. What you get is scuffing and line-abrasion complaints downstream, and the reflex is to blame the coating chemistry. I've sat in those meetings. The coating supplier gets a phone call before anyone pulls the lehr entry temperature log. Nine times out of ten the coating was fine. The thermal window feeding it wasn't.

Who actually owns this on your floor

On a typical container line the lehr operator logs zone setpoints and runs polariscope spot-checks, commonly every 15-30 minutes per section. The cold-end inspection operator runs the automated stress and dimensional checks. A line mechanic covers both ends. In my experience the lehr operator role is the least resourced and least trained position on the shift, despite controlling the defect mode with the longest latency of anything on the line.

That's backwards. The hot-end superintendent should own recipe lock and sign off on any lehr re-qualification tied to a job change, full stop — the operator on shift shouldn't be adjusting zone setpoints without that sign-off, and in practice on under-instrumented lines they sometimes do, because nobody's watching. A 1-2% rise in cold-end reject rate correlating with a zone-3/zone-4 drift of only 5-8°C is common where thermocouples get calibrated annually instead of per changeover. Annual calibration on a component running per-job variance is not a control, it's a hope.

And the 0600 handover is where this usually falls apart. On most lines I've seen, the night-shift's swabbing and polariscope spot-check data doesn't make it into the morning briefing in any usable form — it's on a clipboard somewhere, or in someone's head. That's exactly the tribal-knowledge gap the SKU Library inside our Job Change Tool was built to close: the lehr profile for a given SKU gets locked and versioned, not remembered by whoever happens to be on nights.

What Europe's cullet mix and the GCC's cheap gas both do to your curve

In Europe, FEVE's Close the Glass Loop data puts average EU container glass recycled-content input around 80%, with Belgium and Germany running above 90%. Higher cullet ratios aren't a problem on their own, but they change the glass's specific heat and thermal conductivity slightly batch to batch. A lehr profile validated against one cullet blend can drift out of tolerance on the next delivery without a single operator touching a setpoint. Generic efficiency consultancies rarely connect that dot, because cullet ratio sits in the batching department's scope and lehr drift sits in yours.

In the Gulf, the dynamic runs the other way. GCC states still price industrial natural gas well below export parity, historically in the US$1.25-3.00/MMBtu range against multiples higher on the international LNG spot market. That lets Gulf plants run richer lehr soak profiles without the fuel-cost discipline European operators are forced into by ETS carbon pricing on furnace-and-lehr combined energy draw. It's a real advantage today. It also means when GCC operators do finally face a carbon-cost driver, or when EU CBAM scope extends and starts pulling regional export competitiveness into the conversation, the plants that already run disciplined annealing curves — not just cheap ones — will be the ones that don't have to relearn the process under pressure.

What to check before your next job change

If you take one thing out of this: don't let a container-weight or geometry change go through without a lehr stress re-qualification. It costs you twenty minutes on the floor. It costs a lot more than that in returned pallets three weeks later.

• Pull zone-3 and zone-4 setpoint logs against the last three job changes on that section
• Confirm thermocouple calibration is per-changeover, not annual
• Cross-check polariscope spot-check cadence against actual shift logs, not the SOP
• Compare current cullet ratio against the ratio the profile was last validated on
• Verify hot-end and cold-end coating window temperatures against lehr entry/exit specs

This is the kind of gap a full hot end audit is built to find, because it looks at the lehr as a per-job variable rather than a fixed asset that just happens to sit between forming and inspection. A vendor-neutral container glass consultant isn't selling you a lehr. There's no incentive to tell you it's fine when it isn't.

So next job change: does your lehr get re-qualified, or does it just get trusted?