Labeling Machine Glass Bottles: Setup and Trial Runs

A labeling machine holds a label position on glass only when three things have been written down first: the feature on the bottle the machine senses, the tolerance measured from that feature, and the real diameter range of the bottle at the label band. The trial run then has to prove, on production bottles and production label rolls, that the whole measured spread stays inside the agreed window at production speed and somewhat beyond it. A trial that ends with people agreeing the bottles "look fine" has proved nothing and usually leads to a second one.

What follows is the machine side of the job. Whether the adhesive survives chilling, condensation or a wet pasteuriser is a materials matter handled under label adhesion testing. What the label must say and how big it has to be is a market and artwork matter, shown for one product category in our notes on labels for olive oil bottles. Both should be settled before machine settings are discussed.

Six inputs to settle before you ask for a labeler quotation

A machine supplier configures against whatever it is told. Leave several inputs open and the quotation gets built on guesses nobody has verified, and the gaps surface later as a changeover kit, a different roll format or a stopped line. Settling them on paper takes one meeting.

1. Labeling method. This follows from the bottle and the market, not from taste. It also decides which bottle properties matter at all.
2. Reference feature. The machine has no idea where the shoulder is; it only knows where a sensor reported the start of a bottle. That may be the base interrupting a beam, a mechanical trigger on the conveyor, or a registration mark or moulded feature when the label has to be oriented. Choosing the reference, and confirming the bottle presents it the same way every time, is the step skipped most often.
3. Position tolerance in machine terms. "Centered" is a wish. A top edge at a stated distance below the shoulder, with a plus or minus range and an angular tolerance around the circumference, is something a machine can be set to.
4. Surface over the label area. Described as the actual treatment, not as a look.
5. Line speed and direction of travel.
6. What happens between filler and labeler. People tend to remember this at the first trial. A bottle arriving wet or oily is a different labeling problem from the same bottle arriving dry.

What each labeling method needs from the bottle

Pressure sensitive labels are the usual choice on glass: the label rides on a liner and the applicator barely touches the container. Shrink sleeves and wet glue each solve a problem pressure sensitive cannot, and each shifts the burden onto a different property of the glass. Cut and stack, wrap around and in mould labels lengthen the list, but these three cover most projects.

Pressure sensitive

This is the method most sensitive to position control, because the label lands on a moving surface. Any change in body diameter alters the surface speed where the label touches, and with it how far the label runs before it grabs. An oval body alters that speed twice in every revolution; a taper alters it all the way up the label. Blown on applicators generally cope with a tapered container better than wipe on ones, which explains why a label that behaves on a straight wall can ride up or down on a tapered body.

Shrink sleeve

Here the logic flips. Sleeve placement, and so label position, comes mostly from the applicator, the conveyor speed and the sleeve length, not from the glass surface. A sleeve suits a strongly curved or tapered body because it covers the full circumference, but it needs a sleeve applicator plus a shrink tunnel. Abrupt neck transitions, sharp shoulders, deep embossing and pronounced horizontal ribs are where film webs, folds or tears. The tunnel also puts the glass through a thermal cycle, so a container carrying a lot of internal stress can fail there. That is a container fault, even though the labeling station is where it first shows.

Wet glue

Wet glue copes with dust and moisture that would beat a pressure sensitive adhesive, but it needs a glue system, a label magazine and positive rotation of the bottle. Friction between bottle and pad, and between bottle and drive belt, determines the wrap. A slippery surface, an odd base diameter or an out of round body makes the rolling rate uneven and the label angle with it. The glue film has to be even too, so attention returns to the bottle surface more than to the label edge.

Bottle geometry at the label band

No glass container is perfectly round, and a pressure sensitive label shows it. The nominal diameter tells you little. Measure the largest and smallest diameter within the label band, the taper across the height the label covers, and where any shoulder radius starts if it falls inside the label area. Each one maps to its own defect.

• Ovality. On a wrap around label the laydown speed follows the surface. Where the diameter is bigger the surface runs faster and the label covers a shorter arc, and the reverse where it is smaller. Within one turn the label moves ahead and falls back, which appears as a leading edge that wanders, a seam gap that opens and closes, or a faintly wavy top edge. Slight ovality can be invisible and still out of specification, so test it with a gauge, not by eye.
• Taper. With wipe on or blow on application the contact point shifts while the label transfers. A body narrowing upward gets a slightly rotated label whose top edge lags the bottom. In a sleeve applicator the same body takes a sleeve that is tight at one end or slack at the other before shrinking, and a wrinkle at the shoulder can result.
• Shoulder radius. A label meant to finish just under the shoulder depends on where that radius begins. If the blend differs between moulds, the visible gap changes although the label position is right. A visual acceptance criterion is justified here, since the customer sees the label in relation to the shoulder.

State the labeling requirement against the measured label band, and keep it consistent with the bottle drawing. The conventions for writing and inspecting those figures are covered under dimensional tolerances for glass bottles. A label tolerance tighter than the bottle dimension underneath it is beyond any machine.

Surface finish and how the applicator presses the label down

Frosted, matte, satin, coated and untreated are descriptions of appearance. The labeler needs something else: a known finish that stays the same across the label area, so the adhesive wets it and trapped air gets out.

Two cautions sit outside the table. A deep acid etch or a heavily textured moulded surface should go through an adhesion test; a sample that looked acceptable is not evidence. And organic coatings fired onto the glass do not behave like coatings cured at lower temperatures. The gap between them appears after a wet or chilled cycle, not at the moment the label goes on.

Then match the pressing method to the surface. A brush that suits flint may mark a soft coating. A roller presses evenly on a round body and unevenly on an oval one. An air jet is gentle and handles irregular shapes, but it adds no pressure whatsoever, leaving everything to the adhesive and the surface. Confirm each pairing on a trial.

Converting a position tolerance into machine settings

A tolerance taken from a marketing brief becomes machine parameters in four steps. Working through them in sequence removes most of the disagreement that otherwise breaks out at the trial.

1. Fix the datum. Name the feature the machine senses and the feature the label is measured from. Normally that is the container base for height and a defined point on the circumference for angle. Put both in writing: the same figure taken from the base and taken from the shoulder is two different requirements, and buyer and supplier can read one sentence and imagine different things.
2. Turn the vertical tolerance into a timing window. The applicator's trigger point and delay translate conveyor position into label position. For a tolerance of plus or minus one and a half millimetres, divide that distance by the conveyor speed; the result is typically a very brief interval in milliseconds. A faster line therefore has less room than the same line run slowly, and a tolerance held at one speed may be out of reach at another.
3. Test the angular tolerance against the orienting mechanism. To centre a label on a moulded feature the machine must detect it, and detection accuracy becomes the angular tolerance. A round bottle with no feature lands at an effectively random angle unless a stop, a screw or a servo orienter is fitted, and that is a machine modification.
4. State the label's own tolerance. Length variation within the printing tolerance moves the leading and trailing edges even on a perfect machine. Roll tension, die cutting and splice accuracy add to it and deserve the same attention as the bottle dimension.

Terms to get right in the enquiry

Labeling has its own vocabulary, distinct from filling. A wrong term produces a quotation for a machine you did not ask for.

• Line speed is given in containers per minute or in metres per minute of conveyor travel. They are not interchangeable, and one machine can be described as bottles per minute at a short pitch or metres per minute at a long one. Ask which figure is the rating and which is derived, because accuracy normally tracks conveyor speed, not the count.
• Dwell time is how long the bottle stays against the applicator or inside the shrink tunnel. It is the usual limit on running faster.
• Label pitch is the spacing between leading edges of successive labels. It has to equal the container pitch, or the applicator runs short of labels or applies two.
• Wrap length is the label length around the circumference: shorter than the circumference when a seam gap is wanted, matched to it for a full wrap.
• Registration accuracy is placement relative to a sensed feature. Position accuracy is holding the required distance from the datum. Suppliers quote them separately, and giving both lets them offer the right orientation equipment.
• Tunnel temperature and residence time describe what the container must withstand in sleeve work. They are not a labeler performance figure.

If you give only containers per minute, the supplier works from an assumed pitch, and a wrong assumption yields a machine rated for a speed you never meant. Give the available dwell time and the supplier can tell you plainly whether your speed is feasible with the chosen method.

Planning the trial run

The trial exists to show whether machine, bottle, label and required position can coexist at production speed, and to locate the point where they stop doing so. Presentations and runs on stand-in bottles do not serve that and can be cut from the day.

What to bring

• Production bottles, not gold samples, in a quantity that includes the extremes of measured diameter and ovality at the label band.
• Production labels in the production roll format. Loose sample sheets carry none of the tension or splices a roll does.
• Filled bottles, where fill level or a wet surface influences labeling. An empty bottle labeled on a demonstration table says nothing about one coming off a filler.

What to record

Log everything as a distribution. Note cycle count and speed for every setting so the run can be repeated. Measure label position from the production datum on every bottle so the spread shows. Record room temperature and humidity, since both strongly affect pressure sensitive adhesive, and a cool morning room can flatter a label that later fails in a warm afternoon plant. Finally, push a setting past its limit on purpose and write down how the machine fails; that is how you learn how much margin exists.

Use a speed ladder

Begin under the planned production speed, raise it in steps and capture the position distribution each time. The step where the spread widens is the practical ceiling of that configuration, and its distance from production speed is your real margin. Running only at target speed shows the target was reachable that day with those bottles, nothing more.

Trial run checklist

Each row gives the check, the way to carry it out, the measured result that counts as a pass, how often to repeat it once the line is producing, and who has to sign off.

Watching for position drift in production

On a running line the figure that matters is the mean, and whether it is moving. A centred mean with a spread that fits the window can run indefinitely. A mean creeping toward one edge looks perfect right up to the first reject, and by then the label stock or the bottle lot gets the blame.

Sampling can borrow the language of incoming inspection, using a recognised scheme such as ISO 2859-1 and an agreed AQL for sample size and acceptance. Frequency is a different matter: base it on how quickly the process can shift, not on batch size. A sensible start is a small sample at the beginning of each shift or run, a check following every changeover, roll change and bottle lot change, and closer checks through the first hours of a new bottle and label pairing. From each sample take the largest deviation from nominal, the range from highest to lowest, and any movement of the mean since the last check.

Agree the action triggers beforehand:

• One bottle outside the window. A defect has been made. Find the pallet and check the setting.
• Range wider, mean unchanged. Variation in the bottle or the label is growing, often the earliest hint of a lot change nobody flagged.
• Mean moved by half the tolerance, same direction, three consecutive checks. That is a trend. Correct it before the limit is hit.

Inkjet or vision inspection can measure every bottle where it is installed. It still needs a defined reference and limit, and it must be validated against manual measurement, or it will just repeat its own opinion very consistently.

On a line handling several formats, keep the history separately for each. Drift seen with a single bottle points to that bottle or its setting. Drift shared by all formats points to the machine: a worn belt, a stretched chain, a sensor going off.

Format sheets, locked parameters and wear parts to request

The settings that make the line work are the buyer's property. They should arrive in a form a new operator can follow without phoning the supplier, yet this handover is the piece most often left unfinished, and it governs how fast a changeover or fault recovery goes.

• A format sheet per bottle in the machine's own units: applicator trigger position, delay, conveyor speed range, roller or brush pressure, air pressure for an air jet, label pitch, sensor positions, and tunnel temperature and residence time where they apply.
• A photograph of every station correctly set. One image of a properly adjusted roller beats a paragraph of description.
• A list of which parameters operators may alter and which are locked. A helpful tweak to a locked value is a frequent source of unexplained defects.
• Spare parts and wear items with expected life in operating hours or containers, not months, since a line on three shifts and a line on one age very differently. Belts and pads, the label sensor and applicator head wear parts usually matter most, joined on wet glue systems by the glue roller and doctor blade. Holding these locally turns a breakdown into routine maintenance.
• Changeover data for a multi-format range: which bottles share an applicator, which require a change part, and a target time for each transition, treated as a working target to verify on your line. A range planned around a few bottle families changes over far more easily than one assembled bottle by bottle, where each format may need its own setup procedure.

What to send with a trial request

Three items come first, because they determine which machine configuration is even in play: the labeling method (with applicator type if already chosen), the surface treatment over the label band together with the largest and smallest body diameter there, and line speed in both containers per minute and metres per minute, with available dwell time and direction of travel.

Then add the label position with its datum and angular tolerance, label dimensions and roll format, and the bottle finish and intended fill where they change the surface the label meets. With that set we can indicate which parameters are likely to be critical for your combination, what the trial should measure, which pass criteria suit the width of your bottle distribution, and where altering the bottle or the method would be the cheaper fix compared with altering the machine.

If the container is still undecided, begin with the glass bottle collections. Diameter, roundness and surface set the limits of what a labeler can do well before any setting is touched.

Questions buyers ask about labeling machines for glass bottles

Can one machine apply pressure sensitive, shrink sleeve and wet glue labels?

As a rule, no. Each needs its own applicator: vacuum drum or air jet, sleeve applicator with shrink tunnel, or glue system with label magazine. Some machines share a common base and accept interchangeable heads, which is worth raising if your range mixes methods. Even then, switching method means changing tooling, not a setting.

Why does the label sit higher on one bottle than on the next?

There are three usual suspects. Bottle height or the triggering feature varies, so labeling starts from a different point. Body diameter at the band varies, changing how far the label travels before contact. Or the roll varies in label length or position through tension and splicing. Measure the bottles that failed, not the good ones, and the cause generally becomes clear.

How much position variation is normal?

No universal figure exists. The achievable spread depends on applicator, speed, bottle roundness and the label, so establish it by measurement at the trial. Agree a window with the brand owner and confirm the mean sits near its centre with room either side. If the window is tighter than the bottle dimension allows, change the bottle or the method; demanding a better machine will not help.

Can a labeler compensate for an out of tolerance bottle?

Partly, and planning around it is a mistake. Spring loaded rollers, servo correction and sensor feedback soak up small variations, and a vision system can reject bottles whose label falls outside the window. Consistent placement on a surface whose diameter and position are themselves inconsistent is impossible, though, because the label goes wherever that surface takes it. A consistent container is usually the more reliable and less costly route.

Can check frequency be reduced once the line is stable?

Yes, after the process has demonstrated stability. Keep the checks that follow a changeover, a roll change or a new bottle lot, since a change is exactly when settings are most likely to be wrong.

Is a lifting label a machine fault?

Not necessarily. A label that lifts in a chiller was placed correctly and matched badly: the adhesive, the surface and the actual conditions do not suit each other. That is answered by testing the specific combination, not by adjusting the applicator. Likewise, a machine cannot fix a regulatory limit on label size or content, and a better applicator will not satisfy a regulation.