Reducing Feeder Loss During SMT Placement
Feeder loss is the share of components that leave the reel without ending up on a board. It is a quiet cost: it appears in the stock room as a shortage, in the schedule as an unplanned reel change, and on the finished board occasionally as a missing part that nobody saw. On a high mix line it is one of the largest controllable contributors to material waste.
What the Number Actually Measures
Feeder loss counts every component that is picked up and then discarded, plus every component that is damaged, misaligned or left behind in the tape. Some of it is unavoidable, such as the parts at the start and end of a reel.
The rest is not. Pick errors, dropped parts, tape advance faults and nozzle clogging all show up in the same number, which is why the metric is only useful once it is broken down by cause. A line tracking a single percentage learns little; a line tracking causes finds the process that needs attention.
Mechanical Causes
Feeders are mechanisms, and mechanisms wear. The sprocket that advances the tape and reel develops play, the spring that holds the cover tape loses tension, and the pickup position drifts by a fraction of a millimetre. At 0402 scale a fraction of a millimetre is the difference between a clean pick and a miss.
Vibration compounds the problem. A feeder bank that is not rigidly mounted lets the tape shift between the peel step and the pickup step, so the part is off centre when the nozzle arrives.
Tape and Reel Quality
The carrier tape is a precision part, and cheap tape is not the same as good tape. Pocket depth, pocket position and cover tape peel force all have tolerances, and a reel at the edge of them will pass inspection and still feed badly.
Moisture matters as well. Tape that has absorbed humidity can deform under the heat of a reflow-adjacent environment, changing the pocket geometry. Storing reels in a dry cabinet and observing the floor life of moisture sensitive parts removes an entire category of component loss.
Nozzle Selection and Wear
The nozzle is the point of contact, and its condition sets the pick quality. A worn tip leaks vacuum, so the part is held weakly and is lost during acceleration rather than at pickup. The symptom looks like a feeder fault and is diagnosed as one far too often.
Choosing the right nozzle matters as much as maintaining it. A nozzle that is slightly too small will pick reliably and place poorly, because the part rotates around the contact point. Matching the tip to the component body dimension is a direct improvement in placement accuracy. The same discipline appears in our SMT prototype BOM checklist, where package data is confirmed before the run.
Feeder Setup and Calibration
Every feeder needs its pickup position taught and verified. Doing this once at the start of a run and trusting it for the whole batch assumes that nothing moves, which is rarely true across a long shift.
Scheduled calibration, plus a re-teach after any feeder is remounted, keeps the pickup window centred. Recording the offset for each feeder makes a repeat build faster and removes the trial and error that generates scrap at the start of a run.
Vision and Recognition
Vision systems reject parts that are within specification when the lighting, the threshold or the library image is wrong. Each false rejection becomes feeder loss, and the operator response is usually to widen the tolerance until real defects pass.
Tuning the recognition parameters per package, and reviewing the reject log rather than only the total, separates false rejects from genuine ones. A PCB design optimisation review often reveals that the same part is being rejected on every build, which points to the library rather than the process.
Environmental and Handling Factors
Static discharge damages components in ways that only appear later, as a shift in leakage or a marginal joint that fails after a few thermal cycles. Grounding the feeder bank, the operator and the reel storage area removes most of the risk at negligible cost.
Handling matters too. Reels dropped on a hard floor deform pockets that were previously within tolerance, and the damage is invisible. Storage on a rack, with each reel returned to the same position, prevents the casual handling that creates feeder loss on the next run.
Material Flow and Reel Management
Loss also occurs between the stock room and the machine. Partial reels that are not labelled with the remaining quantity are counted as full, so the shortfall appears mid-run. Splices made incorrectly create a step in the tape that jams the feeder.
Tracking partial reels as first-in-first-out stock, and verifying the count before the run rather than during it, converts a schedule disruption into a routine top-up. That is often the single cheapest reduction in component loss available.
Sustaining the Improvement
Improvement decays when the measurement stops. A weekly review of rejects by cause keeps the calibration schedule honest and shows when a feeder or a nozzle has reached the end of its life rather than merely being out of adjustment.
The target is not a perfect number. It is a number that is understood, broken down by cause and stable enough that a change in it means something. At that point feeder loss stops being an unexplained scrap figure and becomes a process indicator that can be managed.
Statistical Process Control on the Line
Feeder loss responds well to simple statistical control. Recording the loss per feeder per shift turns a machine level figure into a per station signal, and a feeder that drifts outside its normal band is identified before it produces a visible defect.
The value is in the trend rather than the absolute number. A feeder whose loss has doubled over two weeks is telling the maintenance team something, and catching that early costs far less than discovering it through a shortage at the end of a production run.
Training and Operator Feedback
Operators see the causes before the data does. A run of pick failures on one station, a nozzle that needs cleaning more often than usual, or a tape that peels with a different feel are all observations made within minutes, and they only become useful if there is a channel to report them.
A short shift handover log, backed by a response from the process engineer, closes the loop. Without it, the same feeder generates loss for weeks while the reports attribute the shortfall to supplier quality or to the reel count.
FAQ
What is a reasonable feeder loss target? It depends on package size and mix. Fine pitch chip components on a high mix line will always lose more than large parts on a stable run, so the target should be set per product rather than for the whole factory.
Does the placement machine or the feeder cause most loss? On a well maintained machine, the feeder and the nozzle dominate. That is fortunate, because both are inexpensive to maintain and easy to measure.
How can loss be reduced without new equipment? Calibrate feeders on a schedule, replace worn nozzles, store reels dry and review reject logs. Together these usually deliver most of the available improvement.
