A vertical load test tells you how much axial force a bottle standing on its base survives, in newtons or kilonewtons, and that figure only becomes a safe number of layers once you have calculated what the bottom bottle in your stack actually has to carry. Work out the load first from carton weight and layer count, add margin for transport and uneven sharing, then compare it with a hold-load result on the real pack, using the weakest samples rather than the average. In most glass shipments the corrugated carton gives way before the glass does, so stack height belongs to the whole pack and not to the bottle alone.
What the test records
The equipment is a compression tester with two parallel platens. The bottle stands upright on the lower one or on a rigid fixture, force is applied through its top along the vertical axis, and a load cell tracks the force as it climbs. Buyers meet the same measurement under several names: vertical load resistance, top load test, compression test on a glass container, or stacking strength test when it is done on the filled pack. Whatever the purchase specification calls it, it is one axial measurement.
A single run yields three different figures, and a report has to say which one it quotes:
- Load at first damage is where the first change can be seen or heard. It normally sits below the peak.
- Peak or maximum load is the highest reading before the bottle collapses or the trace drops off.
- Sustained load is what the bottle holds for a stated period without cracking. When load is applied slowly, this can be well under the peak.
A supplier who offers "compression strength" of several kilonewtons with no indication of which of these it is has not given you a design input. The same goes for a figure with no method or fixture attached.
Why the weakest bottle sets the limit
Glass is brittle and elastic. A plastic bottle bulges and springs back; glass holds the load until it fractures, and the fracture comes without warning. Because there is no yield point to show how close you are, the only sound approach is to test enough samples to see the spread of failure loads and then design well below its lower end.
That means the mean is not your design value. Use a minimum, or the mean less a multiple of the standard deviation. One broken bottle in a load-bearing layer hands its share to the bottles beside it, so a pallet is as strong as its weakest container.
What the figure does not cover
A vertical load result describes one loading direction, on an empty or filled container, under one set of conditions. It is silent on impact resistance, thermal shock, internal pressure and closure behaviour. It will not tell you what happens when a pallet is dropped, when bottles are rinsed across a wide temperature gap, or when the product is carbonated. Treating a good top load number as general proof of strength is a frequent and costly mistake.
Calculating the load before testing anything
A test exists to verify a calculated requirement. Skip the calculation and the result has nothing to be measured against: a pass may mean the bottle is heavier than it needs to be, and a fail leaves nobody able to say how far short the design is.
Begin with the bottom carton. The cartons resting on it number the stack height minus one, so its compressive load is roughly that count times the gross weight of a single carton. This is the load on the carton, not yet on the glass. Board, partitions or honeycomb, dividers and bottles all share it.
How it is shared depends on the pack. In a snug carton with a full partition and a lid resting on the necks, much of the force travels through the board and never reaches the bottles. With no partition and bottles sitting directly on each other, the glass takes almost everything. That second arrangement is the one a vertical load test must imitate.
The four margins to add
- Dynamic loading. On trucks, trains and during container handling, vertical acceleration at the pallet exceeds gravity. A factor of about two on the static figure is a usual working assumption for road plus sea. Check it against your route; it is not a universal constant.
- Design safety factor. Corrugated carton specifications typically carry three to five to cover uncertainty in the estimate. Where you land in that range depends on what the goods are worth and what a collapse would cost.
- Eccentric loading. Cartons drift out of line and pallets overhang. A load meant for twelve bottles can settle on three.
- Duration and humidity. Both weaken the parts of the pack that creep, which is covered further down.
A worked example
These figures are illustrative; substitute your own carton data. Take a carton of twelve bottles with a gross weight of nine kilograms, palletised six layers high, with two pallets stacked in the container.
The bottom carton on the lower pallet has five cartons above it on its own pallet and six more from the pallet on top. Eleven cartons come to ninety-nine kilograms, or about nine hundred and seventy newtons over the carton footprint. If board and partition take half and the glass takes the remainder, each bottle averages around forty newtons. A dynamic factor of two brings that to eighty newtons.
The design target should sit above that figure with further margin, since sharing is never even. For a single pallet stack the requirement ends up in the low hundreds of newtons per bottle, and it rises steeply with taller stacks and double stacking. Seeing that rise before the pallet plan is frozen is the point of doing the sum.
Two corrections for stacked and racked pallets
First, account for the pallet itself. When one pallet sits on another, its own weight and the load on its bottom deck pass down through the cartons underneath. When a pallet rests on a rack across unsupported spans, its lowest carton layer bears that pallet's full load with nothing below to help.
Second, pallet feet, stringers and deck boards press on limited areas instead of spreading force evenly over the carton beneath. Both effects push the true per-bottle requirement above what a simple layer count suggests. They are also good reasons to run the hold-load test on the actual pack geometry and not just on a bare bottle between flat platens.
Stacking scenarios and how to frame the test for each
The layer counts here are typical, not fixed. Replace each one with the real pallet plan for your shipment.
| Scenario | Usual stack | Estimating the bottom-carton load | Test set-up | Main risk | Confirm with the glass plant or packaging supplier |
|---|---|---|---|---|---|
| Domestic road freight on one unstacked pallet | Four to six carton layers, nothing above the pallet | Count the cartons over the bottom layer, four to five of them, and multiply by gross carton weight. | Constant-rate compression on the bare bottle to describe the design; hold-load on the packed carton at the estimated load. | Low. Knocks during loading and unloading matter more than compression in transit. | Use of partitions or dividers, whether the lid rests on the necks, and whether shrink wrapping applies top pressure. |
| Full container load, double-stacked pallets in a 40 ft high cube | Two pallets, five to six carton layers each | Everything above the bottom layer on the lower pallet plus the entire upper pallet, whose load travels down through the lower one. | Hold-load at the calculated figure with the dynamic factor included, on the real pack with its partition and carton. | High. Weeks of sustained load at sea in humid air; board creep leads the failures. | Box compression test value at service humidity, loaded weight of the upper pallet, layer pattern in the container. |
| LCL consolidation with mixed pallets | Varies; a second pallet is often set partly across a neighbour | Distribution is uneven. The worst case is a carton taking part of a pallet on part of its footprint, so work out an average and a point load. | Constant-rate on the bare bottle and a carton compression test at the point-load case. | High and hard to predict. Partial loading, rehandling at the terminal and rebuilt stacks are routine. | Who stacks the pallet, whether slip sheets go in, and how the terminal treats the pallet between vessels. |
| Racked warehouse storage over unsupported beams | Two to four racking levels, each pallet bearing only its own load | The lowest carton over an open span takes every carton above it on that pallet, unaided. | Hold-load at the full single-pallet bottom-carton figure, assuming no help from the deck. | Moderate to high in long storage. Deformation is gradual and often noticed only when the pallet comes out. | Beam spacing, whether the pallet is built for racking, and how long goods will be stored at destination. |
| Bottle-on-bottle stacking, no partition or dividers | Pack-dependent; all force goes through glass | Total weight of the cartons above divided by the bottles in the bottom layer. Deduct nothing for the carton. | Compression through a fixture copying the crown, finish ring or shoulder contact of the bottle underneath. A broad flat platen is the wrong tool. | High. Contact is a small ring, unevenness loads a few bottles, and the weakest samples govern. | Exact contact interface on the pallet, finish dimensions and lip thickness, sample size behind any quoted minimum. |
| Parcel shipment of a single or twin pack | Bulk bags and roll cages; height cannot be predicted | Handling and whatever lands on top dominate, so set a high robustness target instead of a calculated service load. | Constant-rate plus compression on the packed shipper, taken far past nominal service load. | Moderate for compression; handling impact contributes heavily. | Strength of the outer shipper for single-parcel handling, and whether an impact test is also required. |
Bottle dimensions that set the result
Force in an upright bottle travels from the base up the body wall, across the shoulder and into the neck and finish. Two bottles with the same capacity can differ several-fold in measured resistance, and the cause is nearly always shape and where the glass sits, not glass composition.
Body diameter and wall thickness
For a given bottle weight, a wider body stretches the glass around a longer circumference and leaves a thinner sidewall, which buckles sooner. A narrow container with heavy walls does well because its wall is thick in proportion to its diameter. That explains why a small pharmaceutical or spirit bottle often posts a high figure while a large lightweight wide-mouth jar of similar nominal weight posts a far lower one. Compare only bottles with a similar ratio of diameter to wall thickness.
Distribution counts for more than the average. Moulded glass is never even: heavier at the heel and base, lighter mid-body, inconsistent around the shoulder. The thinnest, least supported zone carries the load, so a thin patch at the shoulder will break there however good the rest looks. Request a wall thickness distribution report and read the minimum values around the circumference.
Base and heel
A marked push-up or concave base sends force into a ring at the heel. The heel radius is the feature to watch. A generous, smooth radius spreads stress and resists a vertical crack at the foot of the wall; a sharp corner concentrates it and is where such a crack begins.
Wide-mouth jars often have a nearly flat base, with force running almost straight up the wall. Unusual bases, such as a champagne-style punt or a moulded foot, concentrate force at the contact points, and measured capacity falls well short of what the bottle's bulk implies.
Finish and shoulder
The top of the bottle governs how force enters the body. In a pack without partitions, each bottle rests on the crown or finish of the one below, and force arrives on a small ring instead of the whole shoulder. This is a leading reason a bottle passes under a broad platen and then breaks in a real carton. Finish diameter and edge condition, lip thickness, and any bead or crown all form part of the path, so the test should copy the interface the bottle will meet on the pallet.
Annealing
Annealing relieves the residual stress that forming leaves behind. Where it is done poorly, locked-in stress adds to the stress from the applied load, and two bottles identical in shape and weight can test differently. Annealing quality is its own test subject and is only flagged here.
Method conditions that change the number
ISO 8113 describes a method for determining the vertical load resistance of glass containers. National standards in the same family and customer-specific methods sit beside it. They read alike and give different figures, so two laboratories can both report "vertical load" and disagree. The cause is usually one of four things.
Loading rate
In a constant-rate test the crosshead moves down at a set speed while force is plotted against displacement. The curve is worth more than the final figure: its slope gives stiffness, a step marks a section starting to yield locally, and the fall marks failure. Speeding up the crosshead raises the apparent failure load because the glass has less time to accommodate stress. On a brittle material the effect is large, and it is the most common source of disagreement between labs. A result without the speed stated cannot be used.
Run-to-failure or hold-load
Constant-rate testing destroys the sample and gives a failure load. That suits design comparison but does not resemble service. A hold-load test applies a set force for a set time and then checks for cracks. Glass can fail through slow crack growth under sustained stress, so a bottle that rides out a brief peak may still crack minutes later at a lower load.
A pallet spends weeks under load in a humid warehouse, which makes hold-and-inspect the closer model. The best practice is to run both: constant-rate to characterise the design, hold-load to qualify it.
Sample conditioning
Moisture has no significant effect on glass, but it weakens the pack around it. If the test includes a carton or partition, a carton conditioned in a humid room will give a lower result than a dry one. For bare-bottle tests the relevant conditions are sample temperature and platen state. A cold sample straight from a winter warehouse is a different test object from one left at room temperature for the standard period.
Platens and fixtures
Platens need to be parallel, rigid and clean. A soft pad placed to protect the base spreads force and masks a base defect. A hard particle caught beneath the base creates a local stress point, breaks the bottle early and gets logged as a weak container.
The top contact must be defined as well. A broad flat platen tests shoulder and body. A fixture shaped like the crown or neck ring of the bottle below tests the true pallet condition. Each is valid; they answer separate questions.
Why the carton usually gives way before the glass
Under static stacking, the bottle is almost never the weak link. Glass is stiff, strong in compression and does not creep at warehouse temperatures. Corrugated board fails on all three counts. When a stack collapses, the usual sequence is carton first, then broken glass as load shifts or the pallet topples. Knowing this keeps a buyer from adding glass weight to fix a packaging problem.
Humidity
Board loses much of its compression strength as relative humidity climbs, because the paper and its fibre structure soften. Packaging literature commonly cites a loss in the region of thirty to fifty percent between a dry, controlled room and the damp air inside a sea container or a tropical warehouse. A carton can pass a short test in an air-conditioned plant on packing day and later fail in service under a load it handled in the lab. The answer is not thicker glass. Specify the carton by a box compression test value measured at the humidity the shipment will meet, margin included.
Time under load
Board also creeps. A box held near its short-term compression strength deforms slowly and eventually buckles though the load never changed. Packaging designers therefore treat long-term capacity as a fraction of the short-term test value, with the fraction depending on duration and humidity. Compare your load estimate with sustained capacity, not peak, and prefer a hold-load compression test on the finished pack to a single peak reading.
Overhang and stacking pattern
When a carton layer extends past the pallet deck, the outer boxes lose their support and their compression strength drops sharply. The loss is often expressed as a percentage for a given overhang in millimetres, and it grows fast. Misaligned layers, mixed carton sizes on a pallet, interlocking patterns that leave gaps under some boxes and put others on their corners, and missing or broken deck boards all push load onto a minority of containers.
Without partitions, that concentrated load lands straight on the finish or shoulder of the bottles beneath, a condition a broad-platen test never reproduces.
Specify glass and carton together
A bare-container test gives the upper bound of what the glass can contribute. It cannot show whether carton, partition and pallet pattern deliver load evenly to the bottles. When the two halves are specified separately, the service failure tends to fall on whoever did not own the other half of the calculation, and it gets resolved through a claim instead of through data.
Loads added by the filling line and palletiser
Requirements are often derived from the shipment alone and then exceeded inside the plant. Several stations on a filling and packing line put their own force on the bottle before it reaches a pallet. Take the highest of these along with the shipping load when setting the requirement.
- Accumulation pressure. On a single-file conveyor, a gate or the mass of the queue holds bottles back. The leading bottles take a horizontal push that turns into stress at contact points and bending at the base. On a long accumulation table this can be substantial, and it shows as a heel or shoulder crack, not a clean compression fracture. The input you need is the maximum accumulation pressure the line is set to.
- Capping and sealing. Screw capping heads, press-on cappers and crown crimpers push down through the finish into the shoulder, right where the path is narrowest. A head set high or a stiff closure can use up a significant fraction of compression capacity, and over a shift the repetition makes it a fatigue condition too. A bottle that is marginal for stacking can be tipped past its limit at every cap, so discuss the head setting alongside the stacking requirement.
- Handling equipment. Depalletiser grippers clamp layers from the sides and can throw vertical load onto outer bottles during the lift. Labelers press bottles against a backing pad with rollers and belts. Metal detectors, checkweighers and reject stations shove containers sideways into guides.
- Wrapping platens. Shrink tunnel and stretch wrapper platens bear on the top of the load. A stretch wrapper's top platen can load the top layer beyond the figure calculated for it.
The finished pallet needs the same attention. A slip sheet or interlayer board spreads force and eases point concentration, which generally helps. A top platen that bears only on the highest cartons can crush that layer and change nothing lower down. Loading by clamp truck instead of fork applies clamping force to the sides while the pallet base takes the full mass. Write these handling choices into the packaging specification; do not leave them to the loading crew.
Reading the fracture after a failed test
Crack position and shape show which part of the load path was overloaded, and so what to change.
| What you see | What it indicates | Where to look |
|---|---|---|
| Vertical crack climbing from heel into body wall | The classic compression failure. It begins where force turns from heel ring into wall and travels upward as the wall buckles outward. | Thin sidewall, sharp heel radius, or a base that channels load into a ring smaller than the wall. The fix is in base and wall distribution, seldom in composition. |
| Diagonal or horizontal crack around the shoulder radius | Overload in the upper path, where force enters through a narrow finish contact and must spread. More frequent with a small top platen or neck-to-base fixture. | If it happens under a broad platen, the shoulder radius is probably too tight or the wall thin there. A shoulder crack that appears after unloading suggests residual stress, meaning annealing and not geometry. |
| Chipped finish, split lip, crack across the neck ring or through the thread | Force was concentrated on the finish ring. This is the expected contact in a partition-free pack and is easily mistaken for handling damage. | Finish diameter, lip thickness and smoothness of the bore edge. Add them to the specification wherever the pack transmits load through the finish. |
| Star crack radiating from the bottom, or a crack pattern through the base | Usually a different mechanism: a foreign object, an uneven platen or a handling impact. The wall has not failed in load-carrying. | Sample inspection. Check each bottle on a light box and examine the base before testing, or a flawed sample will give a low result that tells you nothing about the design. |
The costliest misreading is to mistake compression for thermal shock or impact. A thermal shock crack typically starts at a stress concentration such as a surface defect, may curve through the wall, and comes from a temperature difference. An impact crack at the heel generally shows a bruise, a conchoidal fracture surface or a small origin mark, with the area around it often intact. Compression failures have neither bruise nor origin point, and the crack tends to follow the direction of the applied force. When a field failure turns up with no test data, the fracture surface is your primary evidence. Photograph it before anyone throws the sample away.
Related tests that a vertical load result cannot replace
Each mechanism below needs its own line in a container specification, because a pass in one test is not evidence for another.
- Impact. A filled bottle that is dropped, struck against a guide rail or thrown off a pallet fails from impact energy at the heel or body. The relevant method is a controlled drop, and the heights and surfaces that matter for filled packs are set out in our guide to the drop test for glass bottles.
- Temperature difference. Hot filling followed by a cold rinse, or a hot bottle meeting a cold surface, cracks glass through differential expansion regardless of its top load figure. The thermal shock test guide lists the process steps where this becomes a hard requirement.
- Container loading. How many bottles go into a 20 ft or 40 ft container, how cartons and pallets are laid out, and how much space pallets and dunnage take up are worked through in the CBM guide for glass bottle shipments.
- Bulk supply. How bottles are bulk packed, depalletised and fed to a line ahead of filling is a separate topic, covered under glass containers supplied in bulk.
What to send for a stacking assessment
When we review a stacking question, we need three things from the buyer:
- The bottle shape or drawing.
- The packing format: carton weight, partition and pallet pattern.
- The stack height you plan to load and how pallets will be arranged in the container or on the rack.
If the pack passes load through the finish or crown, include the closure type and finish designation. Together these define the load path and what the bottom bottle must carry. From them we can set out the load estimate, the test configuration that matches your interface (bare container, packed carton or both), the hold-load criterion, and the margin to keep.
Frequently asked questions
What does a vertical load test on a glass bottle actually measure?
The axial compressive force an upright bottle withstands, given in newtons or kilonewtons. Each run can report load at first damage, peak load, or the load held over a defined time. To be usable, a report must name which one, along with loading rate, the top fixture and the condition of the samples.
How many layers can safely go on top of my bottles?
There is no general rule; calculate it for your pack. Estimate the bottom-carton load from the cartons above and their gross weight, apply a dynamic factor for transport and a margin for uneven sharing, then set that against a hold-load result on the actual pack, reading the low end of the sample spread. Carton, partition and pallet pattern each take part of the load, so the answer depends on all of them.
Why does my supplier's figure differ from my own test?
The two tests probably differ in crosshead speed, top fixture, sample conditioning, or which of the three figures was reported. Faster loading inflates the apparent failure load. A flat platen spreads force that a real crown or finish contact would concentrate. A mean and a minimum can be far apart. Get the method, fixture, rate and sample size before you compare.
Should the bottle be tested empty or filled?
Empty is standard for the compression behaviour of the container, since it removes a variable. Filled and packed is closer to service for the packaging decision: contents add mass, the closure alters the top contact, and carton and partition share the load. Characterise the bottle empty, qualify the pack filled, and state which one any quoted number refers to.
Does humidity reduce the stacking strength of a shipment?
Yes, mainly through the carton. The thirty to fifty percent loss cited for board in humid conditions, combined with creep under sustained load, is why the carton should be specified at service humidity and the stack designed against sustained capacity. Glass itself is largely unaffected, but once a carton deforms, the load shifts onto fewer bottles.
Can a base flaw be mistaken for a compression crack at the heel?
It can. A base defect or a foreign object beneath the bottle may produce a crack that looks like the classic heel failure caused by a thin sidewall or sharp heel radius. Inspecting samples before the test keeps handling defects from being recorded as design failures.