Allot Tech
Filling machine changeover: compare time, formats and restart evidence
A saved recipe is only one part of a changeover. Use the actual product, bottle, cap and label transition to identify parts, cleaning, settings and release checks. Measure the first accepted new pack separately from stable production so a quick adjustment is not confused with recovered output.
Define the exact A-to-B transition
Name the outgoing and incoming SKU revisions, machine configuration and trained crew. A bottle-size change with the same liquid is different from a formulation change requiring a reviewed cleaning route. Record B-to-A separately if it uses different parts or steps. Work outside the approved product or package envelope needs engineering review rather than a routine recipe change. Follow the machine manual and the site's approved isolation and cleaning procedures; this planning sheet does not authorize access to moving equipment.
Build a station-by-station change matrix
| Station | What may change | Evidence before release |
|---|---|---|
| Bottle handling | Guides, supports, indexing parts, sensors and transfers | Approved format parts and stable bottle transport |
| Product path | Hoses, seals, residual liquid and cleaning route | Material compatibility and the site's cleaning release |
| Filling | Nozzle position, dose basis, feed condition and recipe revision | Individual-head results with the actual product and bottle |
| Closure | Cap supply, guides, tooling and approved application settings | Correct closure fit and the agreed integrity checks |
| Label and code | Artwork, batch data, label position and inspection settings | Complete new pack checked; old materials accounted for |
| Line controls | Recipe revision, station states and interface permissions | Matching configuration across stations, including reject handling |
Prepare the next run before the stop
- Confirm both SKU revisions, drawings and the permitted transition.
- Stage identified, inspected change parts and the approved settings sheet.
- Assign station owners, the restart approver and an escalation contact.
- Prepare product and packaging; separate old labels, caps and work in progress.
- Confirm the cleaning route and release evidence for this product pair.
- Prepare checked measuring equipment and a timed observation sheet.
Release the restart through five checks
- Account for the outgoing material and confirm the required cleaning release.
- Verify installed parts, connections and recipe revisions against the approved setup.
- Confirm guards, tools and station readiness under the authorized restart procedure.
- Check ordered startup samples by filling head, then closure, label, code and pack.
- Release through the designated role; confirm the agreed stable-output window and record rejects and stops.
Measure elapsed time, not added task minutes
Use one clock and one named output point. Record the last accepted outgoing pack, first accepted incoming pack and start of the agreed stable-output window. Define the observation duration and acceptance basis before the trial. Task durations can overlap only where the approved work sequence permits; their sum is not automatically elapsed downtime. Labour minutes and production-loss minutes answer different questions.
Illustrative example: the first accepted new pack arrives 35 minutes after the last good outgoing pack; stable output begins at minute 50. Time to first accepted pack is 35 minutes, the additional stabilization interval is 15 minutes, and the total elapsed interval is 50 minutes. These values are not equipment promises. Do not deduct the same stabilization interval again in a shift calculation.
How much liquid does each short run lose at changeover?
Alongside the timed changeover, keep a mass balance for the outgoing and incoming recipes. Record the bulk product left in the feed tank, pipes and filler, then record how much was discarded, released for reuse or held pending a decision. Weigh the liquid in startup rejects using the same basis. Give each rejected pack one entry, even if filling and final inspection both flag it.
Suppose an orange-to-apple juice change discards 5 kg of orange juice at £2/kg and 2 kg of apple juice in startup rejects at £3/kg. That costs £16 in product. If 800 packs pass final checks, the change adds 2p per accepted pack before packaging and labour costs. Ask each supplier to measure the same recipe sequence, cleaning scope and release point.
Resolve the restart symptom before changing more settings
| Observed symptom | Record first | Next decision |
|---|---|---|
| Wrong or drifting fill after loading a recipe | SKU, head identity, recipe revision, product temperature and feed condition | Hold release; review the approved setup and repeat the defined sample check |
| Bottle jams after a format change | Container dimensions, installed parts and the exact transfer point | Stop through the approved procedure; review format fit and handling evidence |
| Drips or foam after a stop | Stop duration, sample order, nozzle behavior and product state | Agree a product-specific restart trial; avoid an arbitrary discard count |
| Correct fill but wrong finished pack | Cap, artwork, printer data and inspection recipe | Keep affected packs separate and verify the whole pack before release |
Put changeover evidence in the equipment quotation
Ask the supplier to demonstrate a named difficult transition with the intended product and packs. State the crew, parts supplied, cleaning boundary, timing points and acceptance window. Keep all repeated runs, including slower ones. A video showing a fast nozzle adjustment is not a complete line changeover record.
- Specify both directions if their procedures differ.
- List change parts, storage, drawings and training in the quoted scope.
- Retain times, individual results, rejects, deviations and action owners.
- Connect the transition trial to the FAT plan and repeat relevant checks after site installation.
Make one setup reference that joins the recipe to the physical machine
A recipe name is a useful locator, but it is not a complete setup instruction. Two operators can recall the same name while using different guide positions, format parts or capper settings. Give the approved setup pack its own identifier and revision. Link the product-and-pack combination, machine identity, applicable parts and operating references to that pack, so the next crew can recover a known configuration rather than reconstruct it from memory.
Separate the saved recipe from the values actually active during the run. A temporary correction may not have been saved, or the saved file may have been edited since its last accepted trial. Check what the supplied controls can display or export; do not assume every machine has revision history, automatic format recognition or backup functions. Use a controlled manual record for information the machine does not retain.
| Part of the setup pack | What to identify and keep | What this record does not establish |
|---|---|---|
| Product and recipe | Product, bottle and cap codes; recipe identifier, revision and quantity basis | A familiar recipe name does not prove that the intended material or packaging is fitted. |
| Mechanical references | Approved datum or scale references, identified format parts and permitted reference images | A screen capture does not record every manual adjustment or prove correct assembly. |
| Connected stations | The matching closure, label, code and inspection settings with their own references | A filler recipe does not establish that every other station changed with it. |
| Active and stored values | Current permitted process settings, saved version and any documented temporary difference | A file bearing the right name does not show which values were active for the accepted samples. |
Change a setting without losing the approved changeover baseline
Use the reference pack when investigating a repeated changeover that now needs unexplained fine adjustments. For example, a recalled filler recipe may be unchanged while a bottle guide was moved during another job. Compare the physical reference and the active settings first. Do not create a new dose recipe simply to compensate for an unidentified handling difference.
- Preserve the accepted baseline before a trial. Record machine and configuration identity, setup-pack revision and the available settings record. Use the manufacturer's permitted backup method if supplied; partial screenshots are not automatically a restorable backup.
- Log each authorized process adjustment: previous value, new value, unit, reason, person, time and linked sample results. Keep the trial identifiable rather than silently replacing the approved master. The technical owner defines the permitted adjustment range.
- If the planned checks fail, retain the failed settings and identify affected output under the site's controls. Escalate before further changes. Returning to an earlier recipe requires confirmation that the complete earlier configuration is still applicable, including product, parts and required cleaning release; it is not an automatic safe restart.
- After the defined checks are accepted, record the approved revision and who authorized its use. Hand over the active version, any temporary differences and pending actions to the next crew. Check that the record can actually be retrieved and followed during the agreed repeat trial.
For a machine without recipe storage, the same logic can use a versioned setup sheet and the manufacturer's manual adjustment references. For equipment with exports, confirm the supplied restore scope and compatibility with its current configuration before relying on a file. This guide does not prescribe controller programming or permission changes.
Keep the setup reference beside the changeover timing record. A short adjustment time with an undocumented correction is different from a repeatable transition using an approved setup. The useful procurement question is what the supplied machine records, what remains manual, and how the normal trained crew demonstrates the next changeover with the same evidence.
Make a fast changeover claim comparable before you buy
A quoted changeover time only helps if you can repeat its conditions in your own production. Separate what was observed, what needed extra assistance and what remains untested. Keep that distinction beside the timing record, not in a footnote after the order.
Put both supplier trials on the same starting line
Name the exact outgoing and incoming product-and-pack revisions, line output point, crew and cleaning boundary. Ask whether tools, inspected format kits and approved settings were already staged before the clock started. If one demonstration includes a supplier technician or excludes the capper, labeler or release wait, label that difference explicitly. Compare complete scope before treating a shorter number as a better machine.
Separate repeatable evidence from a best-case demonstration
Retain every agreed trial, not only the fastest video. Record the condition that caused each wait: missing parts, material supply, adjustment, inspection or approval. Mark an estimated interval as an estimate until measured. Compare performance with the trained crew and support that will actually be available on your shift; a demonstration with additional people does not establish the same result with fewer people.
Keep startup material and acceptance status traceable
Identify the samples, rejected packs and material held for review during the transition. Record quantity losses separately from elapsed minutes so the same loss is not deducted twice in production planning. If a later check fails, retain the earlier reading and the failure sequence, identify the affected output and document the authorized correction and repeat checks. A new passing sample does not silently cancel the previous failed result.
Four things that can change even when the format looks familiar
- Same nominal bottle volume
- Confirm the actual drawing revision, dimensions, stiffness and handling references. A matching capacity printed on the bottle does not establish the same transport or filling behavior.
- Same nominal neck size
- Identify the closure, liner and neck revisions together. Feeding, application and integrity checks depend on the complete approved combination, not only the nominal diameter.
- Same label artwork
- Record material, adhesive or sleeve specification and relevant inspection settings. Identical graphics do not prove equivalent application, reading or detection on the machine.
- Same recipe name
- Record the controlled recipe revision, installed change parts and manual setup references. A familiar screen label is not evidence that the whole line is in the approved configuration.
Questions to settle before accepting a changeover promise
Does keeping the bottle make a new cap or label a zero-changeover job?
Not automatically. The technical owner should review which feeding, application, inspection and release requirements the new material affects. Some combinations may need different parts or settings; others may only need specified checks. Record the approved scope and actual evidence rather than assuming that an unchanged bottle approves the finished pack.
Can I use the supplier's fastest changeover in my shift plan?
Not on its own. Match the transition, crew, prepared parts, cleaning scope and accepted-output boundary to your production. Retain all agreed observations and explain excluded waits or extra assistance. Use an approved, representative planning basis; the best recorded run is not a guarantee that every shift will repeat it.
What if the first new pack passes but later restart samples fail?
One accepted pack does not complete the agreed stable-output check. Keep the affected output under the site's controls, preserve the failure and timing records, and have the responsible parties approve correction, repeat coverage and release. Do not restart the stopwatch merely to hide an unsuccessful attempt or ship held material as accepted product.
This is a procurement and evidence checklist, not permission to alter operating, isolation or cleaning procedures. The equipment manual and the site's authorized technical and quality processes govern the work. No universal changeover time, discard quantity or sample count is specified here.
Download and complete the changeover record
The editable CSV is a planning and evidence record, not a validated operating procedure. Complete the old and new SKU, revision, evidence, owner and status columns for each applicable task. Mark exclusions explicitly instead of leaving them ambiguous.
Changeover time, not just adjustment time
How long does a filling line changeover actually take?
A defensible quotation states the two formats being changed, the crew size, cleaning scope and when the clock starts and stops. A five-minute filler adjustment does not establish a five-minute changeover for the complete bottling line.
For a timed trial, agree to measure from the final accepted bottle of format A at the specified operating rate to accepted production of format B at its specified rate. Also record how long stable production must continue before the restart is accepted. This avoids treating one good bottle during repeated adjustments as completion.
Separate preparation that can safely happen outside the stopped interval from work that requires stopped and isolated equipment. This distinction follows the SMED changeover framework described by Vorne. Preparing a labelled change-parts trolley is different from reaching inside a running machine. Energy isolation, guarding and the approved cleaning procedure remain mandatory boundaries, not time-saving opportunities.
A 42-minute example for a bottle-format change
Illustrative planning example, not an Allot Tech performance guarantee: the same beverage changes from a 500 ml bottle to a 1 litre bottle. The following activities are sequential; no product-cleaning cycle is included. Actual measurements must come from the proposed machine, formats and trained crew.
| Activity | Minutes | Completion evidence |
|---|---|---|
| Clear the previous format and isolate as required | 5 | Remaining bottles accounted for; equipment released for the approved change procedure. |
| Change and set guides and format parts | 14 | Part numbers, guide settings and fasteners checked against the new format record. |
| Set filling and capping recipes | 6 | Recipe revision, fill quantity and closure settings verified. |
| Set label, code and packing format | 7 | Correct artwork, date code, label position and pack count confirmed. |
| Restart, inspect and demonstrate stable output | 10 | Agreed first-good-product and sustained-rate checks passed. |
Total stopped interval: 5 + 14 + 6 + 7 + 10 = 42 minutes. In an eight-hour shift with 30 minutes of other planned stops and two such changeovers, the remaining production window is 480 – 30 – (2 × 42) = 366 minutes.
If the measured accepted-output rate during that window is 40 bottles/minute, the planning output is 366 × 40 = 14,640 accepted bottles. This is not the same as multiplying 480 minutes by the machine nameplate speed. Any additional unplanned downtime lowers the result. A measured accepted-output rate already includes the quality losses in that measurement; do not subtract them again.
Where two teams work safely in parallel, use the elapsed critical path rather than adding both teams’ labour minutes. Where product change requires cleaning, cooling or sanitation release, include that complete sequence separately. A label-only change and an allergen-related product change must not share one assumed duration.
The format matrix
Which changes must the supplier demonstrate?
| Transition | Items to identify | Evidence before release |
|---|---|---|
| Bottle diameter or height | Guides, handling parts, nozzle height, sensors and conveyor transfer points. | Stable transport, no bottle damage, repeatable settings and correct filling position. |
| Cap or neck finish | Cap feed, chute, chuck and change parts; compatibility cannot be assumed. | Closure fit, application quality and agreed leak or torque checks. |
| Liquid or formulation | Contact parts, residue, seals, cleaning sequence and restart product condition. | Approved cleaning release and real-product filling checks; water trials alone may be insufficient. |
| Label, code or pack count | Artwork version, printer data, inspection recipe, collator and pack settings. | Correct finished pack and controlled removal of previous-format materials. |
What should a repeatable changeover record contain?
Record the A-to-B format pair, product, operator count, start/end times, task times, change-part list, recipe revision, cleaning release, startup rejects and accepted-output observations. Record B-to-A separately if the procedure differs. Preserve slower trials as well as the best run, so the production plan is based on repeatability.
For procurement, connect this timed record to the FAT evidence plan, the filling-head calculation and the comparable filling-machine quotation. Include format tooling, training and cleaning-related parts explicitly in the quoted scope.
Planning examples prepared by Allot Tech. Updated .
