Flexible feeding replaces much of the fixed orientation track with controlled part redistribution, camera location and robotic picking.
Understand flexible feeder platforms, machine vision, robot picking, lighting, cycle balance, changeover and application limits.
The core cell architecture
A typical flexible feeding cell combines five coordinated elements:
- Bulk supply meters components onto the presentation surface.
- Flexible platform vibrates or moves parts to separate, flip and redistribute them.
- Lighting and camera create a controlled image of the available parts.
- Vision software identifies acceptable pick poses and sends coordinates.
- Robot and gripper pick, orient and place the component into the process.
The PLC or robot controller manages replenishment, image triggers, pick results, redistribution and fault recovery.
The operating cycle
Parts are placed on the plate at a controlled density. After settling, the camera captures the field. The vision system scores visible components and returns one or more pick poses. The robot collects a valid part and places it into a nest, assembly, conveyor pocket or inspection fixture. When too few pickable parts remain, the platform redistributes the field and the cycle repeats.
What makes a part pickable
The vision system needs a recognisable outline or feature, adequate contrast and enough exposed surface for the gripper. Overlap, translucency, reflection and flexible geometry can reduce pickable yield. Lighting trials are therefore part of the feeder trial, not a late controls task.
- Backlight is useful for strong silhouettes.
- Diffuse top light can reduce glare on mouldings.
- Polarisation may help with reflective surfaces.
- Structured or multi-angle lighting can reveal height or feature detail.
Cycle balance and sustained output
Do not specify the system from the robot's nominal speed. Use a complete cycle model:
- Average number of picks before redistribution
- Platform motion and settling time
- Image exposure and processing time
- Robot travel, grip confirmation and placement
- Regrip or inspection stages
- Bulk replenishment and recovery after failed picks
Two robots or multi-pick grippers may increase output, but only if the plate consistently presents enough valid components.
Changeover and recipe control
A new format may require a vision recipe, lighting change, platform motion recipe, gripper fingers, bulk metering setting and placement coordinates. Good cells make these dependencies explicit and confirm that the correct mechanical tooling is fitted before production starts.
When fixed tooling is still better
For one stable component at high volume, a dedicated bowl or centrifugal feeder may offer a simpler cycle, smaller robot requirement and easier maintenance. Flexible feeding should be selected because it solves variety or geometry—not because it is visually impressive.
Acceptance test for a flexible feeding cell
Test the complete product range, worst-case colours and finishes, normal bulk density, pick success, sustained placement rate, failed-pick recovery, recipe changeover and final placement tolerance. The test should distinguish vision location errors from grip and robot placement errors.
Frequently asked questions
What components suit flexible feeding?
Parts with a camera-visible pick face, manageable overlap behaviour and a robotically accessible grip are good candidates, especially where several formats are required.
Does the camera inspect quality as well as locate parts?
It can check selected visible features, but location and quality inspection are separate requirements. Critical defects may need a second controlled view.
How is throughput calculated?
Model feeder redistribution, settling, image acquisition, pickable-part ratio, robot motion, placement and recovery. Average sustained output is more useful than robot pick time alone.
Can a flexible feeder replace every bowl feeder?
No. Dedicated bowl tooling may be simpler, faster and lower cost for one stable, high-volume component.