OEE & Downtime in Electronics & SMT Manufacturing
Key takeaways
- Measure OEE the same way every shift and compare against your own trend first.
- The big losses: feeder and nozzle micro-stops, changeovers, and placement/solder defects.
- Placement machines stop briefly and often, so the hidden factory is large.
- Defects are costly and hard to rework, so quality losses carry real weight.
An SMT line looks calm and runs fast, which is exactly why its losses hide so well. The placement machines stop for a fraction of a second, thousands of times a shift, on pickup errors and feeder faults, and high-mix production means constant changeovers. The OEE number rarely tells the whole story without good stop data.
What's a good OEE in electronics?
We found no published OEE range for electronics and SMT assembly alone with a stated method. The usual reference is 85 percent world-class (traced to Nakajima, Introduction to TPM, 1988), and published plant studies found average OEE of about 55 to 65 percent (sources: Ljungberg, 1998; Hedman et al., 2016); the more useful target is your own line's trend. The separators are micro-stop rate on the placement machines and changeover speed in high-mix production. Calculate your OEE.
The biggest losses in electronics
| Loss | Why it's big in electronics | OEE factor |
|---|---|---|
| Feeder & nozzle micro-stops | Pickup errors, feeder faults and nozzle issues stop placement briefly and often | Performance |
| Changeovers & feeder setup | High-mix production means frequent product changes and feeder reloads | Availability |
| Component starvation | A missing reel or splice stops the line | Availability |
| Placement & solder defects | Tombstoning, bridging, misplacement and reflow issues create scrap and rework | Quality |
| Reduced speed | Running below rated placement rate to protect yield | Performance |
The brief, frequent stops on SMT lines are the biggest hidden loss. See what yours cost.
One SMT line shift, worked through
Illustrative numbers for an SMT line (printer, two placement machines, reflow, AOI), so you can check the math against your own shift. Planned production time is 450 minutes. The shift log records two product changeovers of 25 minutes and a 10-minute wait for a missing reel: 60 minutes of downtime. The line beat is rated at 30 seconds a board (2 a minute), but runs at 1.8 a minute because the placement rate was turned down for a fine-pitch part. Pickup errors and feeder faults stop it about 200 times for 12 seconds each, 40 minutes nobody logs.
| Step | Calculation | Result |
|---|---|---|
| Availability | run time 390 min (450 - 60) / 450 | 86.7% |
| Boards made | 350 running min (390 - 40 of micro-stops) x 1.8 a minute | 630 |
| Performance | 630 / 2 a minute / 390 | 80.8% |
| Quality | 606 good of 630 (24 AOI fails scrapped) | 96.2% |
| OEE | 86.7% x 80.8% x 96.2% | 67.3% |
The log explains 60 lost minutes. The line lost 147.0: 60 of changeovers and the missing reel, 40 of micro-stops, 35.0 of running below the rated beat and 12 of scrap. The micro-stops are the biggest loss the log never sees, which is typical of placement machines.
Reason codes for an SMT line
A short list everyone uses beats a long one nobody does. A starting set:
- Changeover: product change, feeder setup, stencil change, first article inspection
- Materials: missing reel, splice, wrong component, moisture sensitive part waiting for bake
- Placement machine: pickup error, nozzle, feeder fault, vision reject
- Printer and reflow: stencil wipe, low paste, reflow profile alarm
- Starved or blocked: no bare boards, AOI or downstream backed up
- Quality: AOI fail, bridging, tombstoning, rework
Log them in the free OEE tracker and the downtime Pareto ranks them for you; the downtime log template does the same on paper.
What a camera shows on an SMT line
A stop code tells you the line stopped, not what happened at the feeder bank or the conveyor in the seconds before. Where cameras are fitted, Fabrico's camera-based OEE software syncs camera video with machine data, so you can replay the moment the line stopped. Cameras also catch short stops and a line that is switched on but not producing, the losses nobody writes down: in the shift above, that is the 40 minutes the log never saw.
What downtime costs in electronics
High-value boards, tight delivery schedules and defects that are expensive or impossible to rework mean both lost throughput and scrap hit hard. On a serial SMT line, a stop on one machine starves the rest, so the lost-output number climbs quickly. Estimate your downtime cost.
How leading electronics plants improve
- Measure placement micro-stops automatically and attribute them by machine and cause.
- Cut changeover time with offline feeder setup and standard work.
- Catch defects at the source with vision and SPC, not just at final test.
The partner we recommend, Fabrico, reads stops from the line and puts the video of a stop next to its machine data where cameras are fitted, with AI-suggested likely causes, then lets a technician open the work order from the stop. It is EU-built, with data hosted in the AWS EU region and holds ISO 27001 / 20000-1 / 9001 (supports audit-readiness). Fabrico is a partner we recommend; the tools here are free regardless.
Is OEE measured per placement machine or per line?
Both are useful. Per-machine OEE finds the bottleneck placement head; line OEE shows how starvation and blocking between machines drag the whole line down.
What is the fastest win on an SMT line?
Usually reducing feeder and nozzle micro-stops, because they are frequent, attributable and large in aggregate once you actually capture them.
Does this cover through-hole and final assembly?
The same principles apply: micro-stops, changeovers and defect control dominate, though the specific failure modes differ from SMT.
Common equipment to troubleshoot: AOI machines · Crimping machines · Coil winding machines · Cable extrusion lines · Ultrasonic cleaners · Pick-and-place · Reflow ovens · Wave solder · Vision inspection · Conveyors · full directory