Accumulation Table Conveyor: Buffer Points and Capacity

An accumulation table should hold the smallest stock of bottles that rides through the stop it was bought for: line rate in bottles per minute, times the minutes of that recurring stop, times a coverage factor of roughly one point two to one point four. On a glass line the buffer ahead of the filler is normally treated as mandatory, the one ahead of the labeller is the next most common, and the table after the labeller is sized in packing-machine cycles instead of minutes. Which accumulation mode you choose depends on what the bottle surface and shape can tolerate, not on throughput.

The scope here is narrow: the waiting, the spacing and the capacity between two machines. Valve behaviour, nozzle timing and closing belong to our guide to the glass bottle filling line, and the diameter and height bands a buffer has to accept across a delivery are taken from the tolerance standards for glass bottles, not worked out again.

What a buffer does and what it does not do

Machines on a bottling line never share one effective rate, and none of them holds a steady rate either. A filler only reaches nominal speed while it is actually filling. A labeller tends to be slower and pauses for reel changes and registration faults. Rinsers release bottles in surges, and case packers work in cycles. These are ordinary events, yet on a line with no buffer each of them halts everything.

The stop is the cheap part. The expensive parts are the restart, the filled bottles that someone has to pass or reject at that moment, and the time it takes to get fill height back under control.

A buffer therefore decouples: for a short while the machines on each side keep their own time base. It does not store. A table planned as a warehouse costs more, claims floor that a filling hall rarely has, and keeps open or filled containers standing longer, which raises hygiene and temperature questions that have nothing to do with capacity. Nor does it add capacity. Put a thousand bottles in front of a machine running at half its nominal rate and the machine is no faster; the line simply reaches the pace of its bottleneck a little later, and that delay is the whole benefit.

Six points to settle before comparing quotations

A layout drawing gives a table width, a lane arrangement and a motor. It says nothing about minutes of production held, how hard waiting bottles press on one another, whether a second format fits, or how the section behaves when a stop outlasts its sizing. That is why this is the easiest item on a line to buy badly. Settle these six points first.

• The stop being absorbed. Capper micro-stops and a wash changeover call for different sizes and frequently different designs.
• Minutes of full-rate output to hold. This is a commercial judgement about labour, scrap and idle machines, not an equipment parameter.
• The accumulation mode. Single-file back-pressure, low-pressure spacing and multi-lane tables can hold an identical count while exposing the glass to very different amounts of scuffing.
• The position in the sequence. Buffers ahead of the filler, the labeller and the case packer each solve a separate problem.
• Format change. Count the settings that move, time the job, and ask whether a written record can reproduce it.
• Acceptance evidence. Claimed capacity becomes demonstrated capacity only in a trial on your own bottles.

Rate and stop data to take from your own production record

Sizing is arithmetic, so the inputs decide the quality of the result. Collect four figures per machine, and take them from production records, never from a nameplate or brochure.

1. Nominal rate in containers per minute.
2. Effective rate: nominal rate multiplied by the availability achieved on this product with these bottles, this crew and this shift pattern. The gap to nominal is usually wide, and this figure determines whether a table gradually fills during a shift.
3. Stop distribution: which interruptions come back, how long the longest recurring one runs, and how often.
4. Upstream discharge pattern: an even stream, or a surge and then a gap.

Two cautions belong in the enquiry. First, steady-state sizing fails on a bursty line, because the section must swallow the combined effect of many small events across a shift before it gets a chance to drain. Second, size against the slowest machine downstream, which is not always the neighbour. When the labeller's effective rate is below the case packer's, the labeller sets whether the table fills over a shift.

Ask every bidder for the calculation sheet, not just the layout. Someone who has really sized the section can name the machine whose stop is covered, its duration in minutes, the rate used, the coverage factor and the fill level expected in normal running. A proposal that offers only table area, lane count and motor brand cannot be verified, and this is the weakness we see most often in buffer offers.

Calculating capacity from minutes of output

From stop time to a bottle count

Capacity in bottles is the line rate per minute, multiplied by the stoppage minutes to be covered, multiplied by a coverage factor a little above one. The factor is doing real work. A table sized to match the stop exactly is brim-full when the stop ends, so internal pressure peaks and stability is poorest just as the downstream machine restarts, and no room remains for the next event.

One point two to one point four is a common band. Put the agreed value in writing, and lean towards the top of the band when stops come in bursts or the restart ramp is slow.

Purely to show the method: at one hundred and twenty bottles per minute, a three-minute stop requires three hundred and sixty bottles before allowance and about four hundred and thirty to five hundred with the factor applied. At sixty bottles per minute the same stop needs half as many. Real figures must come from the line's own log and be confirmed on site before anything is built.

From a bottle count to floor area

This step is where offers diverge. A bottle takes up about its body diameter squared plus a working clearance, and that clearance varies with single-file, low-pressure or mass-flow operation. The whole top is never usable: edge bottles and those near the discharge combiner pack less tightly than the middle of the bed, and lane guides and sensors occupy space of their own. Required count divided by achievable bottles per square metre gives the area.

If the area will not fit, there are honest choices: cover fewer minutes, add lanes within the same footprint, or accept a shorter covered stop. Keeping the quoted minutes while quietly shrinking the table is not one of them, because the gap shows up later as a line that stops regardless.

Which direction the capacity covers

A buffer works two ways. Kept full, it feeds a stopped downstream side from stock. Kept empty, it lets the downstream machine carry on through an upstream stop until the section runs dry. A single table can do both, though the capacity needed each way may differ. When an offer says it covers three minutes, ask in which direction.

Three accumulation modes compared

All three modes can hold the same count. They differ in scuffing, fallen bottles and changeover labour, not in throughput, and one cannot be swapped for another later.

Single-file back-pressure

Force passes along the row from rear to front, so the bottle at the stop bears the load of all those behind. Pressure grows with lane length, not with bottle count. Contact with neighbours and rails lasts as long as the hold does, which marks the contact band and can dull a coating or a frosted finish. Bottles tip a few positions back from the stop because the push arrives at the base while mass and lean act higher up. The mode is cheap and needs next to no control. Avoid it for decorated, lightweight or tall narrow bottles, and for long lanes.

Low-pressure dynamic

Because bottles never touch, decoration rub largely disappears and pressure-driven falls do not happen. You pay in control complexity: sensors, a variable-speed drive, and a control philosophy that decides when the surface runs and when it holds. The weak points shift to the discharge, where spaced bottles have to become a stream at the rate the next machine wants, and to the sensors, which must detect the extremes of a delivery and not just the typical bottle.

Multi-lane and mass-flow

Far more bottles fit per square metre than in a single lane. The price is two awkward points, the diverter that splits one lane into several and the combiner that brings them back. Both have to be designed around the bottle. An angled guide that behaves on a round body can lift a corner of a square or oval one, and the merge is where fallen bottles are found most often. A mass-flow bed also needs managing: stop the discharge long enough and it fills, and a full bed pushes in every direction instead of down one lane.

Matching mode to bottle

Short, wide, clear and unlabelled is the most forgiving format; it comes through single-file accumulation with no visible damage. Decorated, frosted, sleeve-wrapped or tall narrow bottles should not be accumulated under pressure at all, so specify low-pressure or multi-lane with spacing. On a multi-format line, specify the mode that every format survives, since a buffer is a fixed installation.

Where buffer points sit in the line sequence

Buffers go at joints where rates differ and a stop costs most. Process decides that order, not spare floor. A typical glass hall runs: depalletising and unscrambling, washing or rinsing, draining or drying, buffer, filling and closing, buffer, labelling, buffer, case packing and palletising.

Before the filler

This table sits on the clean side and keeps the costliest machine running through short upstream stops. Its position matters more than its size. The washing stage releases wet bottles in surges, so the table normally follows the drain or dry stage. That avoids holding wet glass and keeps the whole table inside the clean zone. How bottles are prepared before they arrive is described in our page on cleaning methods for glass bottles.

Before the labeller

The labeller is usually the slowest machine, and its reel changes, registration faults and web breaks are routine. A buffer at its infeed absorbs them so the filler keeps going. The condition a bottle must be in for a label to sit properly is a separate matter, covered under glass bottle labeling machines.

After the labeller

This is a different device. Its job is to hand the case packer or palletiser a complete group, so it is sized on the packing cycle and not on a stop pattern.

Two placement errors and the cost of each extra point

The first recurring error is a table that straddles a hygienic boundary because it happened to fit there. The second is a buffer ahead of a stage that needs a settling period after any stop, for example a labeller whose web must be re-registered. Either section will run, and the layout is still wrong.

One buffer per machine is not a default. Every added point brings a table, controls, sensors, guides, a changeover procedure and another line in the cleaning schedule. For each candidate, ask which specific stop it absorbs, and only then what holding those minutes costs.

Evidence to request at each buffer point

Use the table below as an audit worksheet. It deliberately contains no table dimensions, bottle counts or stop durations, because those follow from your rate, your stop log and your bottle; a generic figure would only be copied into a quotation and mis-size the section.

Acceptance tests and the document pack

What matters in a buffer is behaviour, not dimensions. A table may be square, level, correctly painted and correctly motorised, and still hold half of what was quoted. Five tests settle it.

1. Full-accumulation stop. Run into a deliberately stopped downstream machine, let the table build to its production level and hold for the specified stop. It recreates the line's worst condition and tells you more than any measurement. A supplier who never sized the section will push hardest to cut this one short.
2. Run-down. Stop the upstream machine and time how long the section keeps the downstream one fed. Record this upstream coverage on its own; do not average it with the first result.
3. Format extremes. Use the smallest and the largest bottle in a delivery and confirm that guides, combiner and all sensors still see and steer them. A section tuned to the average grips the big one and loses the small one.
4. Changeover timing. Once by the supplier's procedure, then again with the crew who will do it in production.
5. Surface. A full shift at full accumulation, then inspect the first and last bottle of the row at the contact band, on any decoration and at the base heel. A section that passes everything else but marks glass is unfit for a decorated format. Accept a physical sample as proof, not a verbal assurance.

The same pack should contain a layout drawing with each buffer point and its neighbouring machines, and a calculation sheet giving rate, minutes covered, coverage factor and direction. The control philosophy has to say what happens when the section is full, when it is empty and when a sensor fails; the last case is the one most often undefined and most often encountered. Ask for the changeover procedure and a settings record per format at handover, plus a spare parts list for sensors, drive and guide wear strips, which wear first.

Investigating falls and jams in a fixed order

Trouble on a buffer gets blamed on the buffer, often wrongly. A set order of questions reaches the cause sooner and stops people adjusting the table everywhere until it fails in a new way.

1. Same place or random? A repeating spot means something local and mechanical: a moved guide, a worn strip, a shifted sensor bracket, a combiner part. Random locations suggest a systemic cause such as mode, line rate or the bottle.
2. Only near full? Then pressure is responsible and the fix lies in mode and control, not guides. The mode is wrong for the bottle, the fill level exceeds the design, or the infeed keeps running when the discharge stops.
3. Has the bottle changed? A new sleeve, coating, weight or finish diameter alters both geometry and base friction. A cosmetic change can make an accepted section fail.
4. Has the surface changed? Condensation, wash residue, a different lubricant or a shift in hall humidity all affect base friction, most of all where bottles are pushed in a row.
5. Is the mode suitable at all? Tall narrow or decorated bottles under single-file pressure fall and mark whatever the settings.
6. Is it really the combiner? Geometry is tightest where lanes close into one, and a bottle change appears there first.
7. Are the controls doing what was specified? A sensor set to the average bottle misses a small one, and the drive keeps running into an overfilling section.
8. Is it the buffer at all? This check is skipped most. An out-of-round bottle, a foreign object or a mis-timed upstream machine is not a buffer fault, and tuning the table to mask it creates a second, harder problem.

Keep a short note of the conditions present each time. Two or three events are usually enough to identify the cause.

Format range and changeover settings

A buffer generally accepts a narrower format range than the remainder of the line, because it must keep a bottle steady for minutes where a conveyor only moves it for seconds. Fixed lane guides suit one body diameter band. Beyond it, the largest bottle is gripped or the smallest wanders across the lane and meets the combiner at an angle. The limiting format is normally the tallest, narrowest bottle, which has the least base contact and the highest centre of gravity.

At a format change, expect to move lane guide positions, the count of lanes opened to the downstream machine, sensor positions and heights, the drive speed profile, and the entry and exit angles of any diverter or combiner. Write all of them down per format when the section is first set up, so the crew restores a known setting and does not derive a new one under production pressure.

Where formats rotate regularly, time the changeover with a stopwatch and set it against the changes per week. Twenty minutes to reset means something quite different on a line that changes twice a shift.

Small batches and short runs

A buffer justifies its space when the duration of recurring stops, multiplied by what a line stoppage costs, is large beside the section itself, its floor space and its added changeover time. On a line running one short campaign a week at a modest rate, with a crew that reacts quickly, the sum often goes against a buffer. Accepting the stops and intervening by hand is then better than a table that stands empty most of its life.

Batch size does not alter the arithmetic. It alters how many events the section absorbs in a year. Where a small format shares a line with larger ones, a buffer may still be right for another reason: protecting the shared downstream machine. Size it on the largest batch's stop pattern and the smallest format's tolerance, and write the changeover procedure from day one.

What to send when asking for a buffer to be sized

Three groups of data only the buyer holds, and they are best sent together.

• Line speed. Nominal and effective rate per machine in containers per minute, and the rate required at the buffer's position.
• Stop times per stage. Recurring interruptions, the longest recurring one, its frequency, and how long the downstream machine needs to regain full rate, because that ramp is part of what must be covered.
• Bottle format. Body diameter with its tolerance band, height, empty weight, body shape (round, oval, square or fluted), and any coating, sleeve or direct print. Include the format range expected over the line's life. A section sized for one bottle and later given a lighter one is the commonest source of unexplained falls.

Two additions shorten the reply: the buffer points the plant plans to install, which fixes how many sections are involved and their direction, and the floor genuinely free at each one. Capacity that will not fit the floor is not capacity. On the bottle side, we can supply the drawing and tolerance band for the glass we are coordinating, so the section is sized on the delivered range.

With that, a reply can state capacity in bottles and minutes per point with its direction, a suitable mode, a lane arrangement for the available floor, changeover settings per format, the tests to witness and the documents to expect. A bare machine list tends to produce a table drawing, with the true capacity discovered at commissioning.

Frequently asked questions

Why do bottles fall on the accumulation table and not on a straight conveyor run?

Accumulation is the one place where bottles stand in sustained contact. On a straight run each travels with clearance in a moving row. On a buffer they stop, base force keeps building while mass and lean act above it, and a bottle some positions back loses lateral stability. Do not try to strengthen the bottle. Prevent the pressure: slow the infeed when the discharge stops, split the lane into independently driven sections, or switch to spaced accumulation.

Should the table be sized for the worst stop in the plant log?

No. Size it for the recurring stops in your own record, with the coverage factor applied, and nothing beyond that. Check the result against a bursty shift, since a run of small events can fill a table that a single stop would not.

Why is the buffer before the filler treated as mandatory?

The filler is usually the most expensive machine on the line and the one that tolerates stopping and restarting least. Keeping it supplied through short upstream stops protects fill-height control and avoids judging part-processed bottles at every interruption.

Does a second bottle format mean changing the table?

The table usually stays and its settings move: lane guides, open lanes, sensor positions, speed profile, and possibly diverter or combiner angles. The check to make first is whether both formats fall inside the body diameter band the fixed guides allow, with the tallest, narrowest bottle as the test case.

What does a supplier need to know about the bottle and the floor?

Body diameter and tolerance band, height and empty weight, body shape, and any coating, sleeve or direct print. Available floor matters as much as the line data, because the required area often decides between a single-lane and a multi-lane layout.