Cycle time is rarely a specification written by one person. In most industrial programmes it emerges from a review meeting: the customer asks what the line or the machine can complete in a shift, the engineering team answers with a number, and that number becomes the target everyone is measured against. When the figure is wrong, the cost shows up later - in overtime, in expedited parts, in a cleaning bay that cannot clear its queue, or in a prototype that arrives after the customer's pilot window has closed.
Assessing cycle-time needs before development starts is therefore an engineering discipline, not a sales exercise. The work is to define the unit of output, measure the real process around it, and decide how much capacity margin to build into the design. This article sets out the sequence we use when a buyer brings us a throughput target for cleaning equipment, process equipment or a custom automation cell.
Start from demand, not from a machine
The first question is not "how fast can this run". It is how much output the site actually needs, in the shift pattern it already has. Write the demand down in the unit the site cares about: vehicles per day, cassettes per shift, panels per hour, washes per week. Then separate the three patterns that pull the design in different directions.
- Average load sets the size of the asset. A duty that runs at a steady trickle does not need the same machine as one that must clear a backlog.
- Peak load sets the cycle time. A depot that receives most vehicles in two morning hours has to clear that window, whatever the daily average suggests.
- Duty cycle and recovery sets the machine class. Continuous operation needs different tank capacity, thermal management and consumables handling than intermittent use with rest periods.
A useful test is to ask the operations owner what happens on the busiest day of a normal month, and how long they are willing to wait before the queue becomes a customer complaint. That answer, more than the monthly total, drives the required cycle time.
Convert demand into a required takt time
Once the peak window and the volume inside it are known, the arithmetic is simple. Divide the available working time in that window by the number of units that must be completed. The result is the takt time - the maximum time per unit the process is allowed to take.
Two adjustments are usually needed before this number is trusted. First, remove planned non-productive time: changeover, tank refilling, filter or consumable replacement, scheduled breaks, and cleaning of the equipment itself. Second, decide whether one unit of equipment can be loaded while another unit is being worked. Where the operator can prepare the next item during a process, the internal cycle can be longer than the takt time; where it cannot, the cycle must beat the takt time on its own.
The number that survives these adjustments is the specification the design has to satisfy. Everything after this point is a comparison between that requirement and what the process can realistically deliver.
Measure the current process before setting the target
If the site already performs the task by another method, measure it. Do the measurement at the site, on the actual items, with the actual operators, across a full shift rather than during a demonstration. Record four separate durations, because they are governed by different causes and improved by different design decisions.
| Segment | What it contains | Typical design lever |
|---|---|---|
| Setup and handling | Moving the item in, positioning, connecting, removing, drying time before the item leaves the area | Fixturing, access, workflow layout, parallel staging |
| Active process | The time the process itself needs to reach the required result | Process parameters, machine power, tooling, coverage design |
| Quality check and rework | Inspection, and any pass repeated because the first one failed | Defined pass criteria, in-process indicators, repeat-rate control |
| Queue and waiting | Idle time between items caused by staffing, upstream supply or space limits | Scheduling, buffer space, second station, shift structure |
The rework component matters more than most teams expect. If one item in five fails the cleanliness or finish standard and has to be repeated, the effective cycle time rises well above the machine's nominal figure. Any target that ignores repeat work will be missed on the shop floor.
Test candidate equipment on your own workload
Manufacturer specifications describe a machine under favourable conditions. They rarely describe your contamination, your part geometry, or your operator's technique. Before a cycle-time figure is accepted into a specification, run the candidate equipment on a representative sample of your own work items and count the time to a completed, accepted result - not the time to finish a machine cycle.
Define the acceptance standard first. "Clean enough" is not a test condition; a written criterion is. It might be that visible grease is removed from the sampled surfaces, that the finish is uniform with no streaking at the edges, or that a swab or wipe test at defined points passes. Only with that criterion fixed can the trial produce a trustworthy cycle time.
Run the trial through the real duty cycle rather than a few isolated examples. A machine that holds its performance for two cycles and then loses steam pressure, temperature or vacuum as consumables deplete will look acceptable in a short demonstration and fail over a shift. Note the point at which refilling, draining or part replacement interrupts work, and count that interruption as part of the cycle.
For process steps where the requirement is removing heavy oil or grease from machinery, an on-site trial also settles questions that no datasheet answers: whether the method reaches the geometry of the assembly, whether the surroundings can tolerate the process, and whether the site can handle the residue that comes off. We describe this approach for industrial steam cleaning equipment used in factory maintenance, where the workload varies from a single machine to a whole line.
Decide what to automate, and what to leave to people
The fastest cycle is not always the right answer. Full automation usually shortens the active process, but it adds fixturing, changeover and programming time that only pays back at volume. For families of parts that change often, or for low volumes with high variety, a semi-automatic station operated by trained staff frequently delivers a better effective throughput than a fully automated cell.
A practical way to split the decision is to identify which steps are variable and which are fixed. Fixed steps with well-defined geometry - a pass, a bond, a press, a rinse - reward automation. Steps that depend on judgement, inspection or item-specific handling usually do not, and automating them tends to move the bottleneck rather than remove it.
Where a cell does need to be built to a throughput target, the machine is only part of the answer. Alignment accuracy, feeding method and in-process verification each constrain how fast the cell can run without producing rejects. Where that balance has to be designed rather than selected, our semiconductor process equipment work is a useful reference: bonding and laminating cells for display and electronic assemblies are specified against cycle time and uptime together, because a fast cell that produces defects has no usable throughput at all.
Choose capacity margin on purpose
Once a measured cycle time meets the takt requirement, the remaining decision is how much spare capacity to buy. There is no universal answer, but the trade-off is straightforward and worth making explicit in the specification.
- Too little margin means one disruption - a consumable failure, a staff absence, one difficult item - turns into a missed schedule and overtime.
- Too much margin raises the purchase price and the operating cost of a machine that sits idle most of the week.
- Modular margin is often cheaper than oversized margin. A second station, a spare tank, or a portable unit deployed at peak can cover the same risk at lower capital cost.
Where demand is growing but the future volume is uncertain, ask the supplier whether the design can be extended. A machine that accepts an added station, a larger tank or an upgraded heating module later is usually the better commercial decision than a larger unit bought today to cover a forecast that may not arrive.
Build product development time into the schedule
Cycle-time assessment changes the design brief, and changing the brief takes engineering time. If the required takt is shorter than what existing equipment delivers, the difference is a development task: a new process parameter window, a reworked toolpath, a different heating or vacuum arrangement, or a redesigned handling path. That work needs a schedule of its own, including the prototype builds and the verification runs that prove the achieved cycle time is real.
Verification is where many programmes lose time, because test fixtures and acceptance criteria are left to the end. Fixing them early lets a build be judged against a number rather than an opinion. Where a team does not have that capability in-house, structured hardware testing and prototyping support shortens the loop between a design decision and a measured result.
A short checklist for the assessment meeting
Before the target is signed into a specification, confirm that these items are documented and agreed by both the operations owner and the engineering owner.
- Peak window and volume within it, in the site's own operating unit
- Required takt time, after planned non-productive time is removed
- Measured baseline of the current method, split into handling, process, inspection and waiting
- Written acceptance standard for a completed item, and the observed repeat rate
- Trial results on the site's own work items, across a full shift and full duty cycle
- Consumables, refill and maintenance intervals counted inside the cycle
- Capacity margin chosen deliberately, with an extension route stated in the specification
A cycle-time target is a design input, and it should be treated with the same rigour as any other. Define the demand, calculate the takt, measure the real process, test on your own workload, and choose the margin knowingly. Do that, and the number in the specification is one you can plan a schedule and a budget around rather than one you discover is wrong at commissioning.
