A glass bottle drop test releases the complete shipping unit, with closures torqued, partitions fitted and labels on, from a height read off the weight band table of the chosen standard onto a rigid surface, striking a corner, then edges, then faces. ISTA, ASTM and ISO each answer a different question, so the transport mode picks the family, not habit. The result only means something if fracture, crack, chip and cosmetic damage were defined in writing before the first carton fell.
What the test has to represent
The test article is the shipment, not the bottle. A bare bottle dropped on a floor tells you about the glass. A carton with its closures, liners, labels and partitions tells you about cushioning and load sharing inside the pack, and that is the thing a buyer needs proved before tens of thousands of units leave port.
Six conditions have to be met for the result to be comparable with anything else.
- The pack is tested in its shipping format. If the same bottle ships bare on layer pads to a filling line and also in individual retail cartons, those are two test articles with two results.
- Closures are applied at production torque. Too loose and the cap backs off on impact; too tight and the neck finish is pre-loaded, which changes the failure mode. Record the torque range in the report. Without it, a leak after a drop cannot be attributed to glass, liner or capping.
- Samples are conditioned. Glass ignores humidity, but fibreboard, paper partitions and liner plastics do not. Samples are held in a controlled atmosphere before testing, with a cold or frozen step added when the route crosses extreme temperatures. Cartons pulled warm off the line and dropped at once are a different pack from the one that will sit in a humid port.
- The impact surface is rigid and flat. The methods call for a heavy steel or concrete face so that the energy goes into the pack. A compliant floor or a warehouse mat absorbs part of the blow and gives a pass that a sorting hub will not repeat. Name the surface in the report.
- Samples come from a normal production run. Cartons packed by regular operators, on a regular shift, with the board grade the order will use. A pack hand-built by its designer flatters the result. Where cartons will be palletised and moved by lift truck before the journey, test cartons that have already been handled; a scuffed carton with a slightly crushed corner does not behave like a fresh one.
- Pass and fail are written first. Buyer and packer agree the criteria as events: a leak, a fracture through the wall, a crack reaching neck or base, a chip in the seal area, a detached closure. Whatever is not listed is not a failure. That is intentional, since a rule with zero tolerance for marks fails every pack and drives cost into cushioning nobody needs.
ISTA, ASTM and ISO answer different questions
These families are not interchangeable, and a report that cites one of them should be checked against the question you actually asked.
| Family | How it is organised | Where it fits |
|---|---|---|
| ISTA | Packaged-product performance tests grouped by rigour: the one-series for non-simulation integrity tests, the two-series for partial simulation, the three-series for general simulation, with dedicated protocols for parcel delivery and less-than-truckload distribution | Usually the most direct route for a pack moving through a carrier network |
| ASTM | Separate test methods, including free-fall drop of loaded containers, compression, vibration and impact, tied together by a distribution cycle practice for shipping containers and systems | Useful when the buyer wants to build a programme element by element |
| ISO | Vertical impact and stacking methods for complete filled transport packages | Often named outright by a customer or a marketplace |
Plenty of sound reports cite more than one family. What matters is that each cited method governs the hazard the shipment will meet.
Let the route choose
Parcel freight passes through automated sortation, where a carton is tipped, merged, kicked and pushed off belts. Parcel protocols therefore use many drops from moderate heights onto varied faces. A palletised order going from plant to port, across an ocean and on to a distribution centre meets fewer single drops and a great deal more compression and vibration, so those elements join the handling element. Air cargo is handled gently but sees pressure and temperature swings, which matter to a liner or a film seal. Groupage is the hardest to model: cartons are handled again at every consolidation and deconsolidation point, so set the handling element at the severe end of the band.
Height, contact points and orientation
With the family fixed, three variables are read from its tables rather than guessed.
- Height is keyed to the weight of the individual shipping unit, not the pallet. Heavier cartons drop from lower heights, because impact energy at any given height climbs with mass.
- Number and location of drops. The classic sequence starts on a corner, moves to the three edges that meet at that corner, then covers the faces. The corner comes first because it puts all the energy into one point, the worst single event a carton can meet.
- Orientation. A top-down drop stands for a carton landing flat, a side drop for a roll, an edge drop for a strike against a step or belt edge.
A plan that drops the same face ten times has tested one face ten times. It has not simulated distribution.
Test plans by transport mode
The table gives a starting plan for each mode. Treat the heights and sample counts as an opening position for a conversation with the test house and the packer: once packed weight and dimensions are known, the standard's own table governs, and a carrier's handling profile may justify something tighter. Quantities assume one pack format. A second format needs its own samples, and the totals add.
| Mode | Test type | Height and impact points | Samples | Breakage criteria | Packing change it tends to force |
|---|---|---|---|---|---|
| Full container by sea | General simulation with handling, compression and vibration elements, or the performance practice for shipping containers | Weight band height for the packed unit; corner, the three edges from it, then the six faces, with a fresh sequence for each test package | Three to five production cartons per format | No wall fracture, no closure leak, no crack reaching neck or base; chips allowed only away from the seal area | Corner protection, interlocking partitions, layer pads between tiers, a pallet pattern with no corner overhang |
| Groupage (less than container load) | Partial simulation at the severe end of the handling band, plus compression to reflect double handling and mixed stacking | At or close to the top of the weight band; each corner dropped at least once | Five to eight cartons per format | No leak, fracture or closure separation; edge chips tolerated except on the sealing surface or shoulder | Higher board grade, thicker partitions, corner boards, a stronger inner form |
| Air freight | General simulation with handling plus an environmental element for pressure and temperature change | Moderate height for the band; a face sequence and at least two edge drops | Three to five cartons per format | No leak, closure movement or liner displacement; cosmetic marks inside the carton tolerated | Inner liner or sealed bag for sensitive contents, moulded tray or foam insert, torque check before dispatch |
| Express parcel and courier | Parcel delivery simulation: many moderate drops on varied faces, mirroring automated sortation | Parcel weight band height; the complete face, edge and corner set | Five cartons per format, with spares for the drop count | The strictest cosmetic rule of any mode, as the consumer sees the retail carton: no scuffed print panel, no loose closure | One unit per carton, moulded pulp or foam end caps, shrink band or tamper strip, a closer carton fit |
| Palletised road distribution | Non-simulation or partial simulation handling, combined with vibration and stacking | Lower heights from the band, repeated; a reduced drop sequence and a long vibration run | Three cartons per pallet position, drawn from top, middle and bottom tiers | Linked to the retail complaint rate; hairline cracks and surface chips are counted | No pallet overhang, better stretch film and edge protectors, slip sheets, a lower stack |
The groupage row is the most demanding for a simple reason: the people moving those cartons did not pack them. The sample count matters for a related reason. One carton that survives says almost nothing about a production run; several cartons show whether a failure is systematic or random.
Breakage criteria in four classes
Sorting damage into classes is what turns a verdict into a design tool.
| Class | What it is | How it is treated |
|---|---|---|
| Fracture | A break through the wall that releases the contents or leaves the bottle structurally open | Never acceptable. It ends that test package; one fractured bottle can spoil a pallet and cut whoever unloads it |
| Crack | A discontinuity that has not opened, typically running from the impact point toward base or shoulder | Normally a failure even with no leak. A cracked bottle may pass filling and give way later on a shelf or in a hand, when it is hard to trace |
| Chipped edge | Material lost at an edge or corner with no crack growing from it | Judged by location |
| Cosmetic only | Scuffs, label scuffing, light body abrasion, marks that stop short of the structure | Recorded and counted; a failure only where the printed carton or the bottle surface is part of the product |
Chips are graded by where they sit
Most disputes are about chips. On the base rim of a bottle bound for an automatic line, a chip alters how the bottle stands on the conveyor and under the filler head; it can lean, fall or jam the line. On the shoulder of an opaque bottle that goes into its own carton and is never handled by the consumer, the same chip is minor. So the written criteria should name the sealing surface, the finish, the base and the shoulder one by one. They should also state the dividing rule: a chip with a crack radiating from it is a crack.
Count cosmetic damage anyway
Cosmetic marks are the early warning. A gradual rise in scuffing from one shipment to the next often shows that a board grade has been quietly lowered or a partition tolerance has drifted, well before anything fractures. A report listing fractures alone discards that signal.
Set a rate, not just a verdict
One chipped bottle in five after a simulated parcel cycle points to an unacceptable rate in real distribution. One chip in fifty may be fine for a bulk industrial pack. The commercial figure is breakage per thousand units delivered. Ask the packer what rate the design is meant to achieve and set the test threshold to that rate, not to zero.
What the failure pattern says about the pack
A drop test earns its keep by locating the weakest link. Where the damage lands tells you where to spend.
| Where failures cluster | Likely cause | Where the fix sits |
|---|---|---|
| Carton corners | The bottles, not the pack, are carrying corner loads | Corner boards or a partition grid redesigned to take the load there; a thicker wall all round is not needed |
| Bottle shoulders | Bottles are moving and hitting each other or the board | A tighter cell, a taller partition, or a neck support that holds the top of the bottle as well as the body |
| Bottle bases | Drops on the bottom edge, or the pallet passing impact upward | Pallet and handling, not the carton |
| Leaks with no breakage | Closure, liner or torque range | The closure specification |
| Crushing across the carton | Compression and stacking, not impact | Add a stacking test; tightening the handling element will not help |
The leak row deserves emphasis. A common reaction to leaking bottles is more cushioning. That softens the blow but does nothing for a liner that fails to recover after compression, or for a torque value nobody controlled at the capper.
Work in rounds: test the current pack, see where the damage lands, change one variable, test again. Altering board grade, partition thickness and carton size together leaves no way of knowing which change did the work. The packing format is a designed variable, and it follows from the procurement route, so anyone still choosing between stock, modified and private-mould shapes should settle that first with our notes on sourcing glass containers in volume. If the closure interface is still open, start from the bottle and jar range.
Weighing breakage against over-packing
Breakage is usually costed too narrowly. The broken bottle is the visible part. Behind it sit the labour to unpack and re-sort the pallet, the lost contents, disposal of soiled packaging, a whole carton written off when a leak spreads, floor space tied up while a claim runs, and a late delivery. For a filled, labelled bottle the packaging is often the smallest part of the loss and the finished product the largest, which is why a rate that looks tolerable per empty bottle reads very differently per filled unit.
Over-packing has its own bill, and it is easier to measure:
- a heavier board grade adds carton cost to every unit;
- a thicker partition cuts the units in each carton;
- a bigger carton raises the cubic measurement, which raises volume-charged freight and lowers units per container;
- a taller pallet may breach a container or trailer height limit, and a heavier one eats payload;
- extra cushioning slows packing and unpacking, and a pack that is awkward to open gets damaged by the person opening it.
The aim is the least expensive pack that holds the agreed breakage rate on the agreed lane. To find it, work out the cost per thousand units delivered for the current pack and for the proposed change, freight effect included, and set that against the cost of the breakage the change should remove. In most cases a modest improvement in corner protection pays back at once, while a wholesale move to heavier board does not.
Where a drop test ends and a transit test begins
A drop test shows what a drop does to the pack. A transit test shows what an entire journey does, and the two differ even when the drop element is identical. The gap is everything that happens around the drop: stacking compression in a container or trailer, vibration on the road leg, low-frequency movement on rail, and handling at every transfer. Glass is brittle and does not build up fatigue as metal does. Cartons, partitions and pallets do, and a carton that has spent three weeks compressed under a high stack will not respond to a drop as a fresh one would.
For a full container by sea, a practical programme therefore joins handling, vibration and compression or stacking, with an environmental element when the route crosses extreme climates. Add a rail leg and the vibration profile changes with it.
Read supplier reports with this in mind. A document titled "drop test" may contain the handling element and nothing else. That is legitimate for what it covers, but it does not say the pack passed a full distribution simulation. Where an insurer, retailer or marketplace asks for a full simulation, check which elements were actually run.
Scale is the other limit. One carton on a test rig cannot tell you how it behaves in the forty first layer of a loaded container after weeks of vibration. Buyers shipping full containers should take the drop result and the stowage plan as a single decision. Units per container, stowage factors and safe stack height are worked through in the CBM and container loading guide; the count it produces holds only if the pack survives in that configuration.
When the test has to be repeated
A result describes a pack, and the pack stays the same only while its inputs do. Write the retest triggers into the supply agreement:
- bottle specification: weight, wall thickness or the mould itself;
- closure or liner material, which alters how the pack fails at the neck;
- carton supplier or board grade, which alters the cushioning;
- partition thickness, cell count or tiers per pallet, which alter the load path;
- transport mode, carrier or route, which alter handling and vibration;
- pallet type or pattern, which alters how impact travels from the floor into the carton.
Any single item justifies a repeat. A re-run on one carton is usually enough to confirm the change has not shifted the result.
Two routine checks sit beside the triggered re-runs. Monitor closure torque on the filling and capping line, because uncontrolled torque can cancel a good carton design without touching the packaging specification. And inspect incoming board and partitions periodically against the tested specification, which catches a slow downgrade that would otherwise surface months later in the damage rate. Neither check is a drop test. Both protect its result.
Writing the report so an insurer can use it
Transport claims tend to fail on paperwork, not on glass. Either the file does not show the goods were packed to a defensible standard, or it does not show the damage happened in transit and not before loading. The test report closes the first gap; the loading record closes the second.
The report must let someone recognise the carton in the claim. It should contain:
- the test article: bottle reference, closure reference, carton dimensions and grade, partition type and count, units per carton, pallet pattern;
- the method: standard, series and level, drop height and the basis for it, drop order and contact points, conditioning, impact surface, sample quantity;
- the pass and fail criteria as agreed beforehand;
- results article by article, cosmetic findings included;
- the signature and date of the party that ran the test.
A one-page statement that the pack "passed" is of little use to an insurer. It cannot be tied to the shipment and it cannot be questioned.
Around the report, the supporting set normally holds photographs of the pack before loading, a loading record with pallet pattern and tier count, photographs of the sealed container or trailer showing the seal number, and an arrival record with the seal photographed intact before opening. Add the closure torque setting used at filling; transit leaks are routinely argued as packing failures, and a recorded torque is the simplest reply.
Check the claim window and survey requirement in your own policy before loading. Many insurers want notice and a survey at delivery, and a claim documented after the goods have been reworked or destroyed is far harder to support. The food contact declaration for the bottle is a separate document set and has no role in a damage claim, though both belong in one file in case the goods are questioned on both counts together.
What we need to draft a test plan
Three facts fix the plan: the bottle type and size, the packing format as it stands today, and the transport mode. Between them they determine the test family, the height band the carton falls into, the sample count and the packing changes the result is likely to support. If units per carton and the pallet pattern are already settled, send those too, so the plan is drawn against the real pack. Any breakage figure we discuss is a range to be confirmed against a specific pack, lane and named carrier.
Frequently asked questions about glass bottle drop testing
What does a glass bottle drop test measure?
The performance of the whole pack, not the strength of the glass by itself. A complete shipping unit is released from a defined height onto a rigid surface at set contact points in a set order, then graded against criteria fixed in advance. It is the usual way to show that a packing design suits a given distribution route.
How many cartons should be tested?
Three per pack format generally establishes a baseline. Five is more comfortable when failures are expected to be intermittent, and groupage shipments justify five to eight because they see the most handling. Two pack formats are tested separately, since they fail differently.
Can one drop height be used for every carton?
No. A single height across all carton weights does not follow the standard, and neither does a height chosen because it was what the plant floor allowed. Take the height from the weight band table for the individual shipping unit, and print both the height and its source table in the report.
Does a chipped bottle fail the test?
That depends on location. On the sealing surface, the finish, or the base rim of a bottle that will run on an automatic filling line, a chip is normally a failure. On the shoulder of an opaque, individually cartoned bottle it is often accepted. If a crack radiates from the chip it is graded as a crack and fails, leak or no leak.
Does a pass mean the shipment will arrive intact?
No. A pass says the tested pack met the agreed criteria under the tested conditions. It does not extend to another pallet pattern, a higher stack, a longer route, a rougher carrier, a new board supplier, or closures applied outside their torque range. Its value depends on shipped goods matching tested goods.
Will a drop test report support an insurance claim?
Yes, provided it identifies the test article precisely, states the standard and the basis for the height, records the pre-agreed criteria, and is signed and dated. File it with the loading and arrival records. Alone, it cannot prove the damage occurred in transit; without it, the claim is likely to be contested as a packing defect.