Separate machine demand from feeder capability
The receiving process has a cycle requirement. The feeder must satisfy that demand after normal rejection, recirculation and transfer losses. If the machine pauses or accelerates, a buffer may be needed to decouple the two systems.
| Normal demand | Typical sustained consumption during production. |
|---|---|
| Peak demand | Highest short-term demand during the defined operating sequence. |
| Feeder accepted output | Correct parts delivered at the final agreed hand-off. |
| Buffer capacity | Number of accepted parts available between feeder and process. |
Measure accepted output, not raw movement
Components circulating in a bowl or passing an intermediate sensor are not necessarily usable output. Count only parts that meet the agreed orientation, spacing, quality and transfer condition.
A practical rate and buffer calculation
Start with the machine cycle and number of parts consumed per cycle. Add a reasonable margin based on process variability and recovery requirements, then size the buffer to cover short feeder interruptions without creating damaging queue pressure.
- Machine cycles per minute × parts per cycle = basic demand
- Define normal and maximum production modes separately
- State whether rejects and inspection losses are included
- Agree where the output count is measured
- Define buffer recovery time after a stop or refill
Account for real component and batch variation
Rate can change with surface finish, oil, static, dimensional variation, mixed cavities, bulk depth and temperature. Trials should include representative batches rather than only hand-selected samples.
Choose a meaningful test duration
A short demonstration can miss gradual queue build-up, warming, refill effects and intermittent tangling. The test duration and sample quantity should be long enough to expose normal production behaviour.
Define permitted operator interventions
Acceptance should distinguish normal replenishment from manual untangling, jam clearance, setting changes or hand placement. Recording interventions prevents an apparently high average rate from hiding repeated disruption.
What the acceptance report should record
Record the component batch, controller settings, load condition, test start and finish, accepted count, rejects, interventions, stoppages and final output condition.
| Test basis | Component revision, batch and incoming condition. |
|---|---|
| Settings | Recipes, drive values, sensor positions and tooling configuration. |
| Results | Accepted count, elapsed time, calculated rate and buffer recovery. |
| Exceptions | Jams, doubles, missed picks, manual interventions and component damage. |
Frequently asked questions
How much margin should a feeder have above machine demand?
There is no universal percentage. Margin depends on cycle variability, buffer size, recovery needs, component behaviour and the consequence of starving the process.
Should hopper refill time be included in the rate test?
If refilling occurs during normal production, the test should reproduce it or clearly define the autonomy and refill assumptions.
Can peak feeder rate be used in a specification?
It can be recorded, but it should not replace a sustained accepted rate measured at the final hand-off.
Where should output be counted?
At the point defined as accepted delivery—normally after orientation, inspection and singulation, immediately before the receiving process.