SMT Nozzle Pick Failure: Troubleshooting Guide 2026
01:30 on a Wednesday. Two high-speed placement machines in a Dongguan consumer electronics plant raise Pick-up Error within minutes of each other. Nobody changed the program or the material. The pick failure rate on both machines has climbed from 0.3% to 4.7% in two hours.
The equipment engineer pulls one nozzle, holds it up to the inspection lamp and looks straight down the bore. The 0.4 mm nozzle is no longer 0.4 mm. The bore has opened into a small trumpet – a few microns of metal removed by millions of picks.
A pick failure looks minor: the machine misses the pad, repositions and picks again. On a machine running 40,000 placements an hour, a nozzle failing 3% of the time means tens of thousands of wasted picks an hour, each one pulling the vacuum down, re-triggering the feeder and loading the correction logic onto cycle time. Yield and capacity fall together.
Most shops respond the same way: replace the nozzle. That is often correct. But a nozzle is only the last part of a vacuum chain, and four other links fail with exactly the same symptom. Swap enough nozzles without separating the causes and you simply buy a bigger pile of nozzles for a problem that lives somewhere else.
This guide follows the order an engineer actually works in on the floor: a five-minute triage, a replacement standard for each part in the chain, and a maintenance schedule that stops the scenario above from happening.
1. Separate the Three Root Causes Before You Touch a Spare Part
The first mistake on the floor is going straight to the nozzle. The correct order is identify the source, then change the part. The three cause families look different in practice.
| Root cause family | Typical symptom | Distribution pattern | Direction |
|---|---|---|---|
| Nozzle body | Part will not pick, or drops between pick and placement | Concentrated on one nozzle ID | Replace or clean the nozzle |
| Vacuum system | General loss of holding force across a whole column | Appears across a column or on several machines at once | Check generator, tubing, seals, filter |
| Component or material | Nozzle behaves normally, one part number always fails | Concentrated on one reel, package or supplier lot | Check part geometry and incoming material |
The most useful diagnostic on the floor is cheap: swap the failing nozzle with a known good one and put the good nozzle in the failing position. If the failure follows the nozzle, the nozzle is the cause. If it stays with the position, the cause is in the vacuum system or the material. On every line Merex engineers support, this one action narrows the investigation from “this machine” to “this part” in under five minutes.
What the distribution tells you
The distribution in the log is usually more informative than the error message.
- One nozzle ID fails repeatedly – you have a nozzle problem until proven otherwise.
- One whole column of nozzles is weak – look at the shared vacuum distribution, not at the nozzles.
- One part number fails on every machine that runs it – the nozzle is innocent.
- Failures cluster in the first trays after a reel change – look at how the part sits in the pocket, not at the head.
2. The Five-Minute Triage
Step 1 – Read the distribution (about 1 minute)
Export the pick failure log and sort it by nozzle ID, by feeder position and by part number. You are looking for concentration: one nozzle, one column, one part number, or one time window. Random scatter across all three points away from a mechanical cause and toward something environmental – paste mist, a temperature shift, a material change.
Step 2 – Read the vacuum value (about 2 minutes)
Most placement machines expose the vacuum sensor value directly, and how that reading relates to the symptom is the most direct diagnostic there is.
- The vacuum never reaches the set value – the system is leaking or blocked somewhere upstream.
- The vacuum reaches the set value but the part falls in flight – the nozzle bore, the nozzle seat or the sealing surface is the problem.
- The vacuum is fully in spec and failures continue – stop looking at the head and look at the component.
If your machine shows only a pass/fail flag, a cheap external vacuum gauge and a manual nozzle test adapter are worth more than a full diagnostic session. You need a number you can record before and after a change.
Step 3 – Look and feel (about 2 minutes)
Remove the suspect nozzle and inspect it under a lamp.
- Bore wear: compare the inner diameter against a new nozzle of the same type. A bore that has opened up will not hold a light component reliably even when the vacuum reading looks healthy.
- Burrs and deposit: run a fingernail along the bore. It should feel smooth. A rough bore catches the terminations of the component and tilts it during the pick, which is often misdiagnosed as a material problem.
- Rim damage: a chipped rim lets the component shift at the instant of pick-up. Any visible gap is a scrap condition, not a wear condition.
- Seal face: check for flatness, embedded debris and a degraded rubber O-ring if the type uses one.
Hold a finger over the nozzle opening and create a vacuum against it. A healthy nozzle builds and holds negative pressure; a worn or cracked one will not hold it.
3. Failure Modes and Replacement Standards
3.1 The nozzle body
Bore wear on small passives (most common).
Every pick drags the edge of the component across the inside of the bore. On 0402 and 0603 chips the contact is concentrated on a tiny area, and after several million picks the bore grows. The symptom is a part that picks successfully then releases before it reaches the pad. Standard: fit a new nozzle of the same specification and type. If the failure disappears, the bore was the cause.
Burrs and paste residue.
A burr inside the bore hooks the polarity end or a lead of the component, so the part picks up skewed or rotated. Standard: clean first, replace second. Wipe with a lint-free cloth and a dedicated cleaner. Never strike a ceramic nozzle to clear it – ceramic chips, and a chipped tip is scrap on sight.
Rim chipping and deformation.
Dropping a nozzle, loading it into the wrong holder or a head swap at speed all chip the rim, which shifts the part at the instant of pick-up. Standard: any visible rim gap means the nozzle is scrap, regardless of how good the bore looks.
Material fatigue and blackening.
Metal nozzles oxidise and darken under sustained heat and paste mist. Surface roughness rises, friction becomes uneven and the seal face degrades. Standard: a visibly blackened nozzle combined with a rising failure rate means batch replacement, not a single swap.
A high-frequency mistake: ceramic and metal nozzles are not interchangeable. Ceramic nozzles are built for small parts from roughly 0402 to 0802 and hold the component on the flat face of the bore. Point one at a connector or shield can and the part drops every time. Metal nozzles with a wider bore are built for heavy and odd-shaped parts. Using the wrong family raises the failure rate instead of lowering it.
3.2 The vacuum system
When the swap test shows the failure staying with the position, work through these four items in order.
Vacuum generator.
Placement machines commonly use electric vacuum generators, diaphragm or piston driven. The classic signature is a vacuum that builds slowly, will not hold, or a clear change in the operating sound. These are wear parts and the type must match the machine. Voltage and wattage are not a specification – flow rate, holding capacity and response time are what matter.
Vacuum tubing.
PU tubing that has been bent, crushed or kinked for years leaks in ways you cannot see. Check the routing for pinch points, the fittings for looseness, and the tube for whitening and hardening, the visible sign of aged polyurethane. Where you suspect a section, split the line: block the tube at the midpoint and read the vacuum on each side. The side that cannot hold pressure is the leak.
Seals and O-rings.
The O-ring between the nozzle and the suction cup, and the seal ring on the vacuum cup, are the most commonly missed leak points on the machine. Aged rubber hardens and loses elasticity. The part may look perfectly fine while its sealing performance has dropped well below the original. Standard: remove the seal, compress it lightly and release it. A visibly slower rebound means replace it.
Filters.
A vacuum filter loaded with dust, flux mist and paste particulate reduces holding force across an entire column – the classic “it just gets worse over time” failure, and the reason filters are scheduled items rather than reactive ones.
3.3 Do not forget the component side
A meaningful share of pick failures are not spare part failures. Check the material when:
- One reel or package type fails every time – component geometry or an incoming material problem.
- The part is heavy, tall or has an offset centre of gravity – check the bore against part weight and geometry.
- Failures follow a printing problem and a component shift – that is printing and distribution, not the head.
- Failures cluster in the first trays after a reel change – check how the part sits in the pocket.
The test is the same as before: run that part number on a different machine with a different nozzle. If the failure follows the material, it is not a nozzle problem.
4. Nozzle Selection and Stock Planning
Selection is not “buy the same size”. The spec has to line up on five things: machine model, nozzle ID, body material, bore size and type suffix. A nozzle of the correct diameter from the wrong family will not seal properly, or will drop the part.
| Application | Material recommendation | Bore approach | Stock recommendation |
|---|---|---|---|
| 0402 – 0603 resistors and capacitors | Ceramic or high-grade metal | Small bore matched to component width | 3 – 5 units per nozzle online |
| 0805 – 1206 components | Metal | Medium bore | 2 – 3 units per type |
| Connectors, shield cans, odd-shaped parts | Metal with a larger bore | Matched to part outline and centre of gravity | 2 units per type plus 1 odd-shape spare |
| High-speed, high-duty lines | Coated metal or carbide | Matched to component tolerance | 10% of installed count |
| High-temperature, heavy paste mist | Corrosion-resistant metal or coated type | To standard | Evaluate the whole batch on a cycle |
On aftermarket sourcing, one principle covers it: a nozzle is both a sealing part and a precision part, so price alone is the wrong comparison. A supplier should confirm three things – whether every bore is gauged individually, whether material and surface finish match the original, and whether negative-pressure hold is tested and recorded before shipment. A low-cost nozzle with none of those three is not a cheaper nozzle. It is a line that picks badly.
5. Preventive Maintenance and the Acceptance Test
| Item | Suggested interval | Judgment method |
|---|---|---|
| Nozzle bore and rim | Visual check before every shift; replace on any abnormality | Visual plus tactile, cross-checked against the failure log |
| Nozzle vacuum hold | Weekly spot check; mandatory after every replacement | Seal the nozzle, pull to set vacuum, observe the decay |
| Vacuum tubing and fittings | Monthly check for ageing, kinks and strain | Split-line test to locate the leak |
| O-rings and seals | Replace on any loss of elasticity | Compression rebound visibly slower means replace |
| Vacuum filter | Scheduled replacement based on dust load | Open and inspect filter loading |
| Vacuum generator | Replace when vacuum build time increases | Compare build time against the baseline reading |
The acceptance step is the one most shops skip, and it is what makes the system auditable. Every new or refurbished nozzle gets a negative-pressure hold test: seal it, pull the set vacuum, record the decay over a fixed period. That record becomes the baseline for that batch.
A replacement with no acceptance record is not a replacement. Without it, nobody can prove the new part is better than the old one, and the next failure has no baseline to measure against.
Questions we get asked
How high is a pick failure rate before it counts as abnormal?
There is no universal threshold, so measure against your own line. Keep a historical baseline per machine and treat a rate of two to three times the baseline as an abnormal event that goes through the five-minute triage.
Does replacing a nozzle require re-calibration?
The replacement itself does not need a coordinate recalibration, but a first-article check and a position verification are mandatory afterwards. If the bore or overall dimensions differ from the original, nozzle Z height and pick stroke may need a small adjustment.
Can aftermarket nozzles replace the OEM part?
For standard small-component sizes, a good-quality aftermarket nozzle is routinely stable in production. For heavy, odd-shaped and high-duty parts, run a small validation batch first and focus on bore tolerance and vacuum hold. Keeping one OEM nozzle as a reference on critical stations is inexpensive insurance.
A whole column has lost holding force. Nozzle or vacuum first?
Vacuum first. Swapping nozzle positions is the cheapest test available: replace the nozzle only if the failure follows it, and investigate the generator, tubing, seals and filter if it stays with the position.
How often should nozzles be replaced?
There is no fixed life. It depends on component size, pick frequency and line intensity. Manage the failure rate and the vacuum hold value together rather than waiting for a batch of failures to force the issue.
In closing: the quality of pick failure handling is not measured by how fast you can swap a part. It is measured by whether you can narrow the cause down to one component in five minutes. Triage by distribution, verify by swapping, replace against a standard, maintain on a schedule, record a baseline for every change. Run that loop consistently and pick-failure line losses usually drop by a wide margin.
Merex Technologies supplies SMT nozzles, feeders, vacuum components and repair services for placement lines, backed by feeder calibration equipment and CPK testing for dimensional verification. If your line is dealing with rising pick failures, high nozzle consumption cost or short lead times on spares, send the machine model, the nozzle IDs in use and a recent failure log, and our engineers will return a spare parts list and triage sheet matched to your actual configuration.





















