How to Control PU Filter Gasket Width, Height and Pouring Accuracy

Control PU filter gasket width, height and pouring accuracy as three linked outputs, not as one CNC setting. Width is driven mainly by delivered mass and wetting, height by foam chemistry and cure conditions, and path accuracy by fixture location, nozzle stand-off and speed through corners. Lock the material window first, then tune flow against travel speed, verify a cold start and a warm production run, and release the process only from measured parts.

A PU gasket dispensing machine can repeat a programmed contour, but repeatable motion does not guarantee a repeatable seal. A stable bead requires the metering system, two-component material, filter frame, fixture and inspection method to remain inside one documented window. This distinction matters because a bead can follow the correct path and still be too wide, too low or locally starved.

Start with the assembled seal requirement

The correct target is the gasket condition after cure and compression, not an attractive free-rise profile. Record the mating flange, available groove or land, expected assembly gap, compression direction and allowable interference. Then convert those product requirements into a nominal cured width, cured height and contour position with upper and lower limits. Do not copy a tolerance from an O-ring drawing: ISO 3601-1 covers O-ring dimensions and tolerances, while a dispensed PU foam gasket is formed directly on the filter frame and must be qualified for its own joint.

For each filter family, keep one signed master sample and one controlled drawing. Mark measurement locations before the trial: straight runs, each corner type, the start/stop joint and any change in substrate level. A single average hides the exact defects that cause bypass leakage. The production record should show individual readings as well as the average and range.

Output	Primary influences	Measure after	Typical failure signal
Cured width	Mass per path length, viscosity, wetting, nozzle gap	Defined cure time	Wide on straights, narrow after acceleration
Cured height	Mix ratio, temperature, reaction profile, delivered mass	Defined cure time and conditioning	Correct width but low crown or variable rise
Contour position	Fixture datum, part variation, robot path, nozzle alignment	Immediately and after cure	Bead size is stable but shifted off the sealing land
Start/stop joint	Valve timing, overlap, reflow and cleaning state	Immediately and after cure	Bulge, gap, stringing or unmixed spot

Use one mass-per-length equation to connect flow and speed

Pouring accuracy becomes easier to diagnose when output and table speed are expressed as mass per unit path length. The basic control value is mass per length = mixed output in g/s ÷ path speed in mm/s. If the machine delivers 3.0 g/s while the nozzle travels at 100 mm/s, the theoretical application is 0.030 g/mm. That number is a starting setting, not a promised gasket size, because expansion and wetting are material-dependent.

When width is uniformly high around the entire filter, reduce mass per length in a controlled step; when it is uniformly low, increase it. Change only output or speed during a diagnostic trial, never both at once. The site’s published machine range spans four configurations with stated total output from 1.5–30 g/s and table movement from 2–24 m/min, so the selected metering package must be confirmed against the intended bead rather than chosen from maximum output alone.

Observation	First check	Controlled action	Do not assume
Width high everywhere; height also high	Actual mass per length	Reduce output or increase speed by one documented step	That the nozzle is the only cause
Width high; height low	Viscosity, substrate wetting and cure condition	Restore material and temperature window before motion tuning	That less material will restore height
Narrow bead only after corners	Acceleration and material delivery response	Add corner compensation or smooth the speed transition	That the straight-run flow setting is wrong
Random width variation	Tank level, pump signal, pressure, bubbles and nozzle condition	Remove the unstable cause; repeat a weighed shot	That CNC path editing will fix metering noise

Stabilize the two-component material before path tuning

A path program cannot compensate for drifting viscosity or reaction. Condition both components to the supplier-approved temperature window, circulate or agitate them only as specified, verify that tank levels stay above the validated minimum and remove entrained air before making dimensional decisions. Log actual material temperature near the metering point; room temperature is not a substitute for component temperature.

Check ratio by a supplier-approved method and compare separate component or mixed-shot mass over a fixed interval. The product page lists a designed A:B range of 100:25–35 for the available machine family, but that is an equipment envelope, not a universal recipe. The approved ratio comes from the exact polyol/isocyanate system’s TDS and process validation. Isocyanates also require controlled handling, ventilation and exposure controls; the UK HSE isocyanate guidance explains why measurement and worker protection belong in the production plan.

Calibrate metering with weight, not the HMI value alone

Run a timed shot into a suitable container, weigh the net mass on a calibrated scale and calculate actual g/s. Repeat at least three shots after the machine reaches its normal operating state. The HMI setpoint is a command; the scale result is evidence. A consistent offset suggests calibration error, while a wide shot-to-shot spread points to air, pump wear, pressure fluctuation, valve response or an unstable material state.

Perform the check at the low, normal and high output points that the production recipe will use. A machine can be accurate near the middle of a pump range but unstable near its lower limit. Select the SPU20-LQ1, LQ2, LQ3 or LQ4 configuration so the target bead sits inside a controllable working region. The broader PU machine buyer’s guide explains why output range, component count and application method must be matched before purchase.

Set nozzle stand-off and alignment from a physical datum

Set nozzle height from the actual filter surface or groove datum, not from the empty worktable. Excessive stand-off lets the stream wander and stretch; insufficient stand-off can flatten the wet bead, contaminate the nozzle or make part-height variation visible as width variation. Use a gauge block, teach pin or documented touch-off method, and recheck after a nozzle change or fixture service.

The nozzle should remain centered over the sealing land and normal to the local surface unless the process trial proves a different orientation. Inspect the tip for partial cure, damage and eccentric flow. A smaller or partially blocked opening raises back pressure and changes the stream even if the pump command has not moved. Record nozzle identity and service life in the recipe so a replacement is a controlled change.

Control corners with speed planning and dispensing compensation

Corners expose the timing difference between machine motion and material delivery. If a head decelerates while output stays constant, mass per length rises and creates a corner bulge; if it accelerates faster than the delivery response, the bead becomes narrow after the turn. Use rounded path transitions where the product permits, keep acceleration repeatable and apply output compensation only after confirming the mechanical path.

Sonderhoff reports a three-axis dispensing cell with positioning repeat accuracy of ±0.05 mm and output options from 0.05 to 100 g/s, demonstrating that motion repeatability and metering range are specified separately. Its published process history also states that constant gasket height is important for sealing under compression and can be monitored at defined points. These are useful control principles, not performance claims for a Yongjia configuration.

Separate width corrections from height corrections

Width and height often move together when delivered mass changes, but not always. If both drift in the same direction, check mass per length first. If width remains stable while height moves, investigate component temperature, ratio, moisture, cure environment, material age and the time between pouring and measurement. If width moves while height stays stable, inspect wetting, nozzle stand-off, speed and the substrate surface.

Published FIPFG process information describes the cured foam gasket as sealing by compression between mating parts. It also notes that material selection can influence the width-to-height relationship; one fast-cure PU example forms an approximate 2:1 width-to-height profile. Treat that ratio only as a material-specific example. The production target must come from the selected chemistry, filter geometry and compression test, never from a generic internet ratio.

Fixture the filter so the datum repeats

The robot can repeat perfectly and still dispense in the wrong place if filter frames float inside the fixture. Locate each part from functional datums, constrain it without distortion and add mistake-proofing for mirrored or rotated loading. Check fixture wear with a master part and record the allowable locator movement. Thin plastic frames need support near the seal path so clamping does not change the surface height.

The machine product page states a maximum square work area of 700 × 700 mm and a maximum round diameter of Φ650 mm across the listed configurations. Those values define the published work envelope, not fixture clearance or guaranteed contour accuracy at every part height. Send the real part drawing, fixture stack height and access model before final configuration. The same principle applies when a PU adhesive dispensing machine or a low-pressure PU foaming machine is considered: process output must match the material and geometry.

Measure a defined cross-section after a defined cure time

Use one measurement method for process capability. A vision system can map width and contour position without touching the bead; a laser or optical profile sensor can measure height; a cut cross-section is destructive but useful during validation. Hand calipers can deform soft foam, so use them only if the gauge method has been proven repeatable. State the conditioning time and temperature on every report.

Measure more than the easiest straight section. Include the start/stop joint, the smallest radius, the longest straight, the point immediately after a corner and any height transition. For each location, record nominal, lower limit, upper limit, actual value and pass/fail. Keep raw values: a mean inside tolerance does not rescue one low section that can create an air bypass path.

Diagnose defects with a cause matrix

A useful defect matrix starts from the geometry signature. Uniform oversize points toward output, speed or material expansion. Periodic variation points toward pump pulsation, drive behavior or recurring fixture movement. A single-location error points toward path coordinates, local substrate contamination or a height transition. Start/stop defects point toward valve timing, overlap and cleaning state.

Defect signature	Likely cause family	Evidence to collect	Release check
Bulge at every corner	Deceleration without flow compensation	Speed trace, output command, corner width map	All corner types inside the drawing limits
Gap or thin start/stop joint	Valve delay or insufficient overlap	High-speed video, joint mass, cured section	Continuous joint after cold and warm starts
Bead shifted but dimensionally stable	Fixture or path datum	Part datum position and taught coordinates	Contour location within positional limits
Height falls during a batch	Material condition, ratio or temperature drift	Temperature, ratio, tank level and timed-shot trend	Beginning, middle and end samples pass
Random bubbles or voids	Entrained air, moisture or feed interruption	Material handling log, pressure trend, section images	No void at specified inspection locations

Run an eight-step setup and verification loop

The process should move from product requirement to released evidence in a fixed order. Skipping directly to a visually acceptable bead makes later troubleshooting subjective. Use the same loop for a new filter, material lot, nozzle, fixture repair or major recipe change.

For the chemical reaction sequence behind mixing, rise and cure, use the site’s polyurethane foaming process reference. It provides context, but the approved gasket recipe must still come from the chosen material supplier and the filter validation plan.

Build a FAT around measured filters, not dry motion

A dry-path demonstration proves reach and program execution; it does not prove gasket width, height or pouring accuracy. The factory acceptance test should run the buyer’s representative filter frame, intended or agreed trial material, production nozzle and fixture. Agree measurement points and acceptance limits before the test so neither side judges the bead by appearance alone.

At minimum, include a cold first-off, repeated normal-cycle parts, a planned pause and restart, and a warm last-off sample. Capture timed-shot mass, actual component temperatures, alarm history, cycle identifier and all dimensional results. If final chemistry is unavailable, label the FAT as a mechanical/metering trial and reserve reacted-gasket acceptance for site validation.

Practical RFQ checkpoint: send the filter drawing, sealing-land section, target cured width and height, tolerances, material TDS/SDS, required output, shift pattern and available utilities. Yongjia can then confirm the metering configuration, fixture concept and trial scope instead of guessing from one finished-filter photo. Request a gasket dispensing review.

Lock the recipe and monitor production drift

Once the trial passes, freeze the recipe revision, material identification, ratio, temperatures, output, path file, speeds, acceleration values, valve timing, nozzle, stand-off, fixture and cure conditions. Limit editing rights and require a documented change reason. A backup file without matching material and fixture data is not a complete recipe.

Verify the first part after startup, one part at a defined interval, the first part after a stoppage or refill, and the last part of the run. Trend width, height and shot mass separately. A trend moving toward a limit is actionable before it creates rejects. Link each measurement record to the filter lot and recipe revision so a complaint can be traced to production evidence.

Frequently asked questions

What controls PU gasket width most directly?

Delivered mass per unit path length is the first control variable, calculated from actual mixed output divided by travel speed. Wetting, viscosity, nozzle stand-off and substrate condition then determine how that mass spreads. Use weighed output and a defined cure condition before changing the path.

Why is gasket height unstable when width looks correct?

Stable width with changing height usually points away from contour motion and toward chemistry or cure: component temperature, ratio, moisture, material age, ambient condition or measurement time. Verify those records against the material supplier’s approved window before adjusting speed.

How should pouring accuracy be checked?

Check three separate results: repeated timed-shot mass for metering accuracy, contour position against part datums for motion accuracy, and cured width/height at defined locations for process accuracy. One HMI value or one attractive sample cannot prove all three.

Can a universal width-to-height ratio be used?

No. Published examples may show an approximate profile for one specific foam system, but expansion, wetting and cell structure depend on chemistry and cure. Set the target from the selected material, joint geometry, compression requirement and filter validation.

What should be included in a PU gasket machine FAT?

Use representative parts, fixture, nozzle and agreed material; test cold start, normal cycles, pause/restart and warm production; record timed-shot mass, temperatures, alarms and dimensional readings at corners, straight runs and the start/stop joint. State clearly which final filter tests remain for site validation.