Liquid Level, Internal Pressure or Vacuum: Which Inspection Machine Does Your Line Need?

A line that ships 6,000 bottles an hour gives the inspection machine roughly 0.6 seconds to read each container — about as long as a human blink. The inspection machine you pick decides whether a microleak, an underfill, or a closure torque defect gets caught at the line or on a retailer's shelf six weeks later. We have spent twelve years commissioning dosing and inspection systems on beverage lines across three continents, and the most expensive mistake we see is buying all three inspection modes when one would do, or buying one machine when the defect profile demands two.

Why "Buy All Three" Is Almost Always the Wrong Default

When a buyer walks into a tender for an inspection line, the default reflex is to specify all three modes. That reflex costs money, and the cost is rarely justified by the defect profile. Three-mode lines require three sensor calibrations, three reject mechanisms or one shared mechanism with cascading logic, and three integration points with the filler and capper upstream.

The harder rule we apply on every commission is this — specify the inspection machine that catches the dominant defect, then add the next one only when the QA log proves the second defect category is material. Most of the lines we commission run two inspection modes, not three; the third mode's contribution to the defect catch is usually below the measurement noise of the production data.

There is a specific exception. A line that ships into multiple markets with different regulatory regimes, or a contract packager that runs product for several brand owners on the same filler, can justify three modes because each market or each customer demands its own evidence file. We have seen this in halal-certified juice, organic dairy, and pharmaceutical-adjacent beverages. For a single-product line on a single continent, three modes is almost always too many.

Liquid Level Inspection — The Under-Fill Specialist

Liquid level inspection is the oldest and most common inspection machine on a beverage line. It catches what the filler puts in the bottle: underfill, overfill, missing fill, foamy fill. On a properly commissioned line, the machine reads either the bottle's net weight (load cell or checkweigher approach) or the liquid surface position (X-ray, capacitance, or laser triangulation).

What it catches well: gradual filler drift as the day progresses; a clogged nozzle that drops fill weight on one station; foam trapped in the bottle after a vacuum break at the filler; a missing cap that has not yet been detected at the capper. These are the line-readers most-read defects, and a properly commissioned level machine catches 98 to 99.5 percent of them above the set tolerance.

What it does not catch: closure defects that do not change the bottle's mass (a loose closure on a correctly filled bottle reads as a good unit at the level machine); microleaks invisible to weight; oxygen ingress over time. This is why the level machine alone is rarely enough on a carbonated line.

Where the level machine shines is on a still-beverage line, particularly juice, water, or dairy where the dominant QA risk is fill weight itself. On those lines, the level machine is often the only inspection needed, and adding a pressure or vacuum station is over-spec.

Internal Pressure Inspection — The Carbonation + Closure Cop

Internal pressure inspection reads the headspace pressure of a sealed container. On a carbonated line, it catches gross closure failures, gross seal leaks, and headspace pressure drift on the capper. The sensor is typically a non-contact needle probe that pierces or presses against the closure surface and reads pressure by deformation.

What it catches well: closure torque that is too tight (the bottle is over-pressured); closure torque that is too loose (the bottle has lost carbonation between filler and inspector); gross seal failure where the closure is visibly tilted or unseated; headspace gas drift caused by a misadjusted gas mixer at the filler.

What it does not catch: microleaks at the closure-to-bottle interface that pass pressure at the inspection station but bleed carbonation over weeks of shelf life; underfill, since fill weight does not move the headspace pressure reading on a properly carbonated line; oxygen ingress in a non-carbonated product. This is the inspection mode that closes the largest gap left by the level machine, but it does not close the microleak gap.

Pressure inspection also requires the line to be in steady state. On an LN₂-dosed line, the headspace pressure is high immediately after dosing and falls as nitrogen evaporates. The pressure inspector needs to sit at least 1 to 2 seconds after the dosing station so the headspace stabilises, otherwise every bottle reads as over-pressure and the reject rate runs artificially high. We always sequence the inspection station after the dosing dwell, not before.

On a clean-in-place pressure station, the sensor head and the probe path have to survive the cleaning cycle without bias drift. The 3-A Sanitary Standards cover the equipment-cleanability expectations that apply to any pressure sensor or vacuum sensor in a wash-down zone, and the EHEDG hygienic design guidelines cover the validation of cleaning-in-place and sterilisation-in-place cycles on the same equipment. Buyers who spec the inspection machine without naming these two standards tend to receive a machine that passes the factory acceptance test and fails the plant's hygienic commissioning.

Vacuum Inspection — The Oxygen Gate

Vacuum inspection reads the headspace vacuum in a sealed container and detects microleaks that pressure inspection misses. The mode is most useful on still beverages, hot-fill products, and any container that is sealed under negative pressure. On a properly vacuum-sealed product, a closure microleak shows up as a slow vacuum loss, and the inspection machine catches it before the product ships.

What it catches well: closure microleaks at the bottle-to-cap interface invisible to pressure; oxygen ingress in a still-beverage line; vacuum loss on a hot-fill line where the cooling tunnel has not stabilised; small leak paths around the capper mandrel that pressure inspection reads as in-spec but vacuum inspection reads as drift.

What it does not catch: underfill, since vacuum is independent of fill volume on a properly sealed container; gross closure failures that pressure inspection already catches more cheaply; gross carbonation loss on a carbonated line, where the inspector needs pressure-mode not vacuum-mode to read the carbonation state.

The case for vacuum inspection on a carbonated line is often overstated. A carbonated product loses carbonation long before oxygen ingress becomes the limiting shelf-life factor; pressure inspection handles the carbonation gap and vacuum inspection adds little. We typically recommend vacuum inspection on a carbonated line only when the line is also running a still product family on the same filler, or when the QA regime demands an independent vacuum check.

How Upstream Dosing Decides What Downstream Inspection Sees

This is the section most RFPs leave out, and it is the section that determines whether the inspection machine you just specified actually performs as the supplier claims. An inspection machine is a sensor; what it sees depends entirely on what the upstream stage puts in the container. If the dosing station varies the headspace pressure by ±20 percent bottle-to-bottle, the pressure inspector reads noise, and the QA team blames the inspection machine.

On an LN₂-dosed line, the dosing station controls headspace pressure, residual nitrogen volume, and the timing of nitrogen evaporation. Our liquid nitrogen dosing systems are designed to deliver ±2 percent dose dose repeatability at the dosing nozzle; if the dosing station drifts, the downstream pressure inspector has no chance to maintain its tolerance.

On a hot-fill line, the cooling tunnel controls headspace vacuum. A cooling tunnel with insufficient dwell time leaves every bottle under-vacuumed, and the vacuum inspector reads every bottle as a reject. We have seen lines where the cure was a longer cooling tunnel, not a different inspector. The right answer is to tune the upstream first, then commission the inspector; commissioning in the opposite order creates a cascading rejection problem that looks like an inspection-machine fault but is actually a dosing or cooling fault.

Buyers who treat the inspection machine as a stand-alone capital purchase tend to under-spec the upstream coupling clause in their RFQ. We always recommend a single commissioning clause that names the dosing or cooling station supplier alongside the inspection machine supplier, so the integration responsibility is joint. The clause should require both suppliers to demonstrate the line end-to-end at the agreed speed and the agreed reject rate before sign-off.

Plant-level food-safety and regulatory compliance frames the same problem from a different angle. ISO 22000:2018 for food-safety management systems is the umbrella a plant auditor will reference when the inspection machine is being qualified under the plant's HACCP plan. Buyers shipping into the United States should anchor the spec to the FDA's FSMA framework; buyers shipping across multiple regulatory regimes can use GFSI and the FAO/WHO Codex Alimentarius as the umbrella recognition standard. The inspection machine does not have to be FSMA-certified, but its calibration records and reject logs do have to live inside a plant-quality system that is.

Five Questions Your Inspection Supplier Should Answer in Writing

Before a tender is awarded, push the supplier to answer these five questions on paper. The answers separate a defensible line specification from an over-spec'd or under-spec'd one. We have watched buyers skip these and pay for it in commissioning delays, in field returns, or in a machine that sits idle for the first quarter.

1. What is the machine's rated bottles-per-hour at the line's actual closure and container type, not at a generic test container? Most published throughput numbers are measured on a standardised bottle; real production closures and fluid products shift the throughput by 8 to 15 percent. The supplier should answer with the rated speed at your container.
2. What is the false-reject rate at the agreed defect tolerance, and what tolerance is the supplier's test data based on? A 0.5 percent false-reject rate looks reasonable in isolation, but on a 12,000-bph line, that is 60 bottles per hour pulled for no reason. The supplier should quantify the false-reject rate at the agreed tolerance and the test conditions behind that number, and the sampling methodology should be referenced to ISO 2859-1 attribute sampling so the buyer can repeat the supplier's number with their own QA team.
3. What calibration drift does the sensor show, and what is the recommended re-calibration interval? Pressure and vacuum sensors drift; load cells drift; X-ray sources age. The supplier should publish the drift curve and the re-calibration interval, and the warranty clause should bind the supplier to that interval, not to a calendar date that does not match the sensor's actual drift.
4. What reject mechanism is used, and is it shared with adjacent inspection machines? A shared reject station reduces floor space but cascades reject logic; an independent reject station at each machine is more expensive but isolates QA trace-back. The supplier should explain the trade-off and let the buyer decide, not decide for the buyer.
5. What does the upstream coupling clause require from the filler, capper, and dosing supplier? The inspection supplier should publish the upstream tolerance the inspector needs (headspace pressure window, fill weight tolerance, closure torque range). If the supplier cannot answer this, the inspector's published performance is a number without a basis.

Liquid Level, Pressure or Vacuum Inspection — Frequently Asked Questions