Placement Vacuum Pressure Verification: Nozzles, Sensors and Missed Picks
A component is held on a placement machine by nothing more than a pressure difference, and that pressure difference is the one parameter in the process that has no visual confirmation. Vacuum pressure verification is what stops a placement head from producing a missed pick that is attributed to a feeder.
Why Vacuum Pressure Has to Be Measured, Not Assumed
A placement head lifts a component by drawing air through a nozzle, and the holding force is the pressure difference multiplied by the area of the nozzle opening. If the vacuum pressure falls, the holding force falls with it, and the component either fails to be picked at all or is dropped somewhere between the feeder and the board. Both outcomes appear in the defect data as placement faults, and both are usually blamed on the feeder until the vacuum is measured.
The measurement is simple and rarely done. A calibrated pressure gauge or a digital vacuum sensor teed into the nozzle line gives the working figure, and comparing that figure against the machine specification takes a few minutes per head. Doing it once per shift on a rotating basis means every head is checked within a defined period without stopping production to check them all at once. The reason the parameter drifts is that it depends on a chain of small restrictions: the nozzle bore, the seal between the nozzle and the spindle, the internal tubing, the filter, the ejector or the vacuum pump and the changeover valve. Any one can lose performance without failing completely, and the loss appears as an intermittent fault.
How the Vacuum System Works on a Placement Head
Most SMT machines use a venturi ejector driven by compressed air, which converts a supply pressure into a vacuum at the pickup nozzle. The vacuum level achievable therefore depends on the air supply as much as on the nozzle. A supply that has dropped from 0.5 MPa to 0.4 MPa will reduce the vacuum at the nozzle by a significant fraction, and the machine will not report an error because the supply is still within its own alarm window.
The nozzle is the other half of the system. A nozzle with a bore larger than the component needs will pick the part but hold it with less force per unit area, and it will also pick up neighbouring parts when the feeder presents them close together. A nozzle with a bore that is too small will fail on heavier components and will be sensitive to any contamination in the bore. Measure supply pressure at the machine, then vacuum at the nozzle, then the holding force on a sample of components. The first two are number checks; the third is the one that predicts whether the component will survive the acceleration of the move. A machine can show full vacuum and still drop a part if the nozzle tip does not seal against the component surface, which is why the force measurement matters.
Setting the Vacuum Threshold and the Leak Test
Machines alarm on vacuum only when it falls below a threshold, and that threshold is often left at the factory default. Set it from measurement instead: measure the vacuum on a good nozzle holding the actual component, then set the alarm at roughly 70 percent of that value. A threshold set too low passes a nozzle that is beginning to fail; one set too high produces nuisance alarms on small components that genuinely generate less vacuum.
The leak test is the complementary check and it catches what the vacuum reading alone cannot. Block the nozzle with a plug, apply vacuum and measure the decay over a fixed interval, typically five seconds. A nozzle and spindle assembly that holds its vacuum is sealed; one that decays quickly has a damaged seal, a cracked nozzle or a leaking fitting. Perform the leak test at the same interval as the vacuum check, because a leak that develops gradually will pass the threshold test for weeks before it fails. Record both figures per head. A vacuum that is stable while the leak rate rises is a seal problem; both falling together is a supply or ejector problem. The distinction determines whether the fix is a nozzle, an O ring or a compressor.
Nozzle Condition and Its Effect on Vacuum
Nozzle condition sets the achievable vacuum. The tip must be flat and the bore must be clear; a tip that has been chipped by a collision will not seal against the component, and a bore that has collected flux or dust will restrict the flow. Inspect nozzles under magnification at defined intervals and clean them with the manufacturer’s recommended method rather than with a wire or a drill.
Contamination is the most common cause of a slow vacuum loss and it is invisible from the outside. Solder paste transferred from a board to the tip, dust from cardboard carriers and adhesive from taped reels all build up inside the bore. A nozzle that has been cleaned and shows a rising vacuum is a nozzle that needed cleaning, and the interval was too long. Nozzles also wear at the tip. A nozzle used for hundreds of thousands of picks develops a rounded or polished rim that seals less well than a new one, and the vacuum decays without any visible damage. Keep a spare nozzle of each type and swap rather than compensate, because raising the vacuum setting to mask a worn tip will cause pickup of adjacent components. Replacement intervals should follow the same accuracy discipline used for placement capability checks, so that a nozzle change and an accuracy check happen together.
Vacuum Sensors: Types, Location and Calibration
The vacuum sensor is usually built into the head and is often the least maintained component in the system. A sensor that reads low will cause the machine to compensate by increasing pick time or by rejecting parts, and the fault looks like a machine problem. Verify the sensor against a reference gauge at a defined interval, using a tee fitting at the nozzle rather than at the sensor port, so that the comparison includes the tubing.
Sensor location matters as much as sensor accuracy. A sensor mounted at the ejector reads the vacuum source, not the nozzle, and it will not see a leak in the tubing or a blocked nozzle. Where the machine provides only a source-mounted sensor, the routine check has to be made with an external gauge at the nozzle, and the difference between the two readings becomes the health indicator for the tubing. Calibrate the sensor on a schedule and keep the calibration record with the machine. A sensor that has drifted by 10 percent will shift every threshold on the machine, so the alarm that used to catch a failing nozzle now accepts it. Where the sensor cannot be calibrated in place, replace it on a defined interval rather than waiting for a failure.
Missed Pick and Dropped Part Signatures
A missed pick and a dropped part leave different signatures in the data. A missed pick usually shows as a pickup error at the feeder, and the nozzle returns empty to the vision camera where the machine detects the absence. A dropped part often goes undetected: the nozzle reports a successful pick, the vision check is skipped or passes on a partial image, and the component is simply missing from the board, where an optical inspection system later finds it.
The vacuum trace recorded during the pick and place cycle is the most direct evidence. A correct pick shows a rapid rise to the working vacuum and a stable plateau through the move; a missed pick shows a plateau at a lower level; a part lost in flight shows a sudden step up to free-flow vacuum part way through the move. Where the machine records these traces, they should be retained for a sample of placements and used whenever a placement defect is investigated. Correlating vacuum traces with the optical inspection results closes the loop. A defect that the trace predicts and the camera confirms is a vacuum problem; a defect the trace does not predict is a placement or a solder problem. That single comparison decides where the engineering time goes.
Component Mass, Surface Finish and Pickup Reliability
Vacuum holding force has to exceed the force the component experiences during the move, which is its mass multiplied by the acceleration. Large placement machines can accelerate at several g, so a heavy component needs a proportionally larger nozzle and a higher vacuum. A nozzle sized for a 0402 chip will not reliably move a large inductor, even though it will pick it up when the machine is stationary.
Surface finish changes the seal. A component with a smooth top surface seals against a flat nozzle and needs less vacuum; a component with a rough, textured or domed surface leaks at the tip and needs both a softer tip material and a higher vacuum. Connectors, shields and modules with uneven tops are the usual offenders, and they are the ones that should be checked with the external gauge rather than trusted to the machine reading. Tape and reel packaging, described in our feeder setup notes, affects the pick as well. A component that is sitting low in its pocket, or one whose cover tape has not peeled cleanly, will be presented at an angle, and an angled presentation reduces the sealing area. Where a specific feeder position fails repeatedly despite a good nozzle and good vacuum, check the pocket depth and the peeling force before changing the nozzle.
Maintenance Schedule and Spare Nozzle Control
Build the maintenance schedule from the measurements rather than from the calendar. Define the vacuum and leak limits per head and per nozzle type, check a rotating sample every shift, and replace a nozzle when its leak rate rises above the limit even if its vacuum is still acceptable. Track each nozzle by identification so that one failing twice can be removed from service.
Spare nozzle control matters as much as the schedule. Nozzles of the same nominal size are not interchangeable between manufacturers, and a nozzle that has been mixed into the wrong box will be fitted to the wrong head. Store nozzles by type in marked positions, keep the quantity matched to the machine inventory, and record every swap with the head and the date. Include the air supply in the schedule. Check the supply pressure at the machine, drain the filter and separator, and verify the air quality against the machine requirement. Water and oil in the air supply reach the ejector and the nozzle, and the contamination they cause looks exactly like a worn nozzle.
Records and Correlating Vacuum Data With Defects
Keep the vacuum and leak measurements per head with the date, the nozzle fitted, the supply pressure and the ambient temperature, since all of these shift the reading. When a placement defect rate rises, the record shows whether the vacuum was already at the low end of its window, which turns a recurring defect into a scheduled maintenance action.
Link the vacuum record to the optical inspection results and to the AOI-reported defect locations. A cluster of missing components at one feeder position, with a vacuum trace that predicts each one, is unambiguous. Publish the vacuum trend on the machine so that the operator can see a head drifting rather than discovering it through a batch of defects.
FAQ
What vacuum pressure should a placement nozzle hold? It depends on the nozzle bore and the component, so measure it rather than assuming a figure. Set the machine alarm at roughly 70 percent of the vacuum measured on a good nozzle holding the actual component, and verify the holding force on a sample of parts as well as the pressure.
Why does the machine report a good vacuum but still drop parts? A leaking tip can show full vacuum at the sensor while failing to hold the component, because the sensor is measuring the source rather than the nozzle. Measure with an external gauge at the nozzle and add a vacuum decay, or leak, test to the routine.
How often should placement nozzles be cleaned? Set the interval from the leak test rather than the calendar. A nozzle whose decay rate rises is contaminated or worn, and cleaning should be scheduled before the vacuum threshold is reached. Clean with the manufacturer’s method and never ream the bore with a hard tool.
