A bonding line quotation usually arrives with a throughput claim on the front page and almost nothing else that a manufacturing engineer can check. Cycle time is the number that decides whether the line clears your order book, whether the cleanroom space you build is the right size, and whether you buy one line or two. It is also the number most often mis-specified, because it is quoted under ideal conditions and then compared with a factory that has changeovers, jam recovery and an operator on break. Below is a working method for setting and testing cycle-time requirements for automated bonding assembly line equipment before you commit to a build.
Set the requirement from demand, not from ambition
Start with sellable demand over the planning horizon, broken into product families, not a single blended part number. Bonding lines are rarely bought for one SKU: a display or membrane-electrode operation typically runs a family of sizes and constructions through the same equipment. For each family, establish monthly volume, the mix ratio, and the seasonal swing. Then convert to required output per hour at the takt your shifts allow.
Do the conversion on the shift pattern you actually intend to run. A line bought to clear a peak that appears two months a year is a line that sits idle for ten. Conversely, a line sized to an average will fail in the peak months and the fix — overtime, a second line, an outsource — costs more than the headroom would have.
Decompose cycle time into its real stages
A single "seconds per piece" figure is not a specification. Break it into stages and give each one a budget:
- Load and unload — including any manual feed step and any indexing wait.
- Vision or measurement — camera photography, alignment search, and any recalibration that has to recur.
- Alignment and bonding — the axes settling, the pre-bond step, vacuum pull, and the bond itself.
- Process hold — dwell at temperature or under vacuum, which is a physical constant and cannot be negotiated away.
- Transfer and handoff — between stations, including buffer behaviour when a downstream station is occupied.
- Recovery — jam clearing, a rejected unit, a re-alignment. Often omitted from quotations entirely.
The stage that sets the line's output is the longest one in a cycle, and the stages that eat the margin are the ones the quotation leaves out. Ask the supplier to state each stage separately, so a claim of high throughput can be located in a real sequence rather than accepted as a headline.
Demand the assumptions that come with the number
Cycle time is meaningless without its conditions. Get written answers to: how many bond operations does one finished unit require; is the stated time a single-station time or the line's output rate; with how many operators; at what yield; on what product size; with what changeover state; and does the figure include or exclude load/unload. On PowerDe's TGV alignment packaging machine solution, the published page gives a concrete example of why this matters: the machine is described as producing roughly 8–10 pieces per hour of glass vacuum bonding, with the note that one product requires four bond operations. A reader who treats the rate as four times faster than it is has mis-sized the line. The same page publishes ±0.03 mm attachment accuracy, manual feeding with camera positioning, an X/Y/θ alignment platform and a capacity figure, all stated together. That is the format a requirement should take: rate stated alongside the process steps that produce it.
Where a supplier's page does not carry the numbers — and several equipment pages list a product without a full parameter block — do not infer them from a sibling model. Ask for the values in writing for the configuration you are buying.
Convert cycle time into equipment count and floor space
Once you hold a staged cycle time with stated conditions, the design follows. Required units equals required hourly output divided by achievable hourly output per line, adjusted for availability. Then check the footprint against your room: the TGV solution page gives external dimensions of about L2260 × W1700 × H2300 mm including the vacuum pump and FFU height, and mentions a safety light curtain. Compare that against your bay height, aisle width and evacuation routes, and against the space a second line would need if the cycle time falls short.
Also plan the upstream and downstream. A bonding step that runs faster than its feeder or its cure step simply moves the constraint, and the buffer you add to absorb that variance is real floor space and real capital. Size the whole sequence, not the interesting station.
Test the claim: a staged acceptance
Write the acceptance in three gates and put them in the contract.
- Sample test. Run your own parts, at your own tolerances, in the supplier's works. Record rate, accuracy and defects. Insist that the trial uses your materials, not the supplier's demonstration stock.
- Factory acceptance with your materials. Run a sustained block — hours, not minutes — and measure throughput, the stage times against the quoted stage times, and the reject rate.
- Site acceptance. Measure the same figures on your floor, with your operators and your changeovers, and include at least one full changeover in the measurement window.
Record availability at each gate: uptime against scheduled time. Some published pages in the semiconductor process equipment range cite an uptime figure and a cycle time for a specific test condition, which is the right way for a manufacturer to state performance — a number tied to a stated test. Insist on the same convention in your specification: every rate you accept carries the condition it was measured under.
Changeover, mix and future products
Ask explicitly how a change between product constructions is made and how long it takes. On machines that offer selectable program recipes, a changeover may be a recipe recall rather than a mechanical rebuild — a large difference over a year of production. State your changeover frequency in the specification, and require the quoted figure to include it. Then ask the harder question: what happens when next year's product arrives, larger or thinner or with a different film stack. Ask what is reprogrammable, what is a mechanical change, and what the cost and lead time of each option is.
Where your process is genuinely non-standard, a built-to-requirement approach beats adapting a catalogue machine. Shenzhen PowerDe Automation Precision Equipment develops customised process equipment against customer product requirements — the TGV solution above was written around a specific glass panel and PI film format — and supports non-standard automation equipment and integrated engineering team services for the wider line build. Send your product drawings, target hourly output, shift pattern and changeover frequency to the contact page, and ask for a cycle-time breakdown by stage against your parts rather than a single throughput figure.
