Closed-Loop Coat Weight Control Using Online Thickness Measurement in Hot Melt Coating
Closed-loop coat weight control is the ultimate step in automating hot melt coating quality. The online thickness measurement system (beta or NIR gauge) provides real-time data on the applied adhesive weight. A control algorithm—typically a PID (Proportional-Integral-Derivative) loop—compares the measured weight to the target and adjusts the gear pump speed (for average weight) or die actuators (for cross-web profile) to minimize the error. The benefits are substantial: reduced start-up scrap, consistent product quality across shifts, and the ability to run at higher speeds without sacrificing uniformity. However, implementing closed-loop control requires careful tuning to account for transport delay, measurement noise, and the system’s dynamic response. Without proper tuning, the control loop can become unstable, causing coat weight oscillations that are worse than no control at all. This article provides a practical guide to setting up closed-loop coat weight control for hot melt coating machines, based on industry best practices.
The first step is to identify the transport delay—the time between when adhesive exits the die and when the measurement gauge reads that same point on the web. The delay D = distance from die to gauge (m) / line speed (m/s). For a typical configuration with the gauge placed 10 meters after the die and line speed 3.3 m/s (200 m/min), the delay is about 3 seconds. During this 3-second period, the adhesive is already coated and traveling, but the control system has not yet seen the effect of any pump speed adjustment. This delay limits how fast the loop can respond. To compensate, the controller may use a “Smith predictor” or a feed-forward term based on line speed changes. The PID parameters must be set conservatively: low proportional gain, low integral gain, and a derivative term that anticipates the error. Some control systems include an “adaptive” PID that adjusts parameters based on the measured delay—shorter delays allow higher gains, leading to tighter control. For high-speed lines (over 300 m/min) where the gauge is far from the die, the transport delay can be 2-3 seconds, which may be acceptable if the pump speed can change quickly. For slower lines (30 m/min) with a distant gauge, delays of 20-30 seconds may be too long for effective control; in such cases, the loop is used more for drift correction than for fast response. An alternative is to install the gauge closer to the die, but this may not be possible if cooling is required before measurement. Implementation of Near IR technology to measure hot melt coatings speeds start up, increasing production, reduces scrap and allows for quality data archiving by roll or run.
Hot Melt Coating Machine - Hot Melt Adhesive Coating Machine
The control algorithm typically has two modes: manual (operator sets pump speed) and automatic (pump speed adjusted by PID). In automatic mode, the operator enters a target coat weight (e.g., 20.0 gsm). The PLC calculates the initial pump speed based on a calibration factor (gsm per pump rpm). The gauge measures the actual weight, and the PID calculates a correction. The pump speed is updated every few seconds (the update rate should be slower than the transport delay to avoid instability). The integral term eliminates steady-state error, but too much integral action causes “windup” when the line is stopped (the error persists, and the integral term accumulates, causing a large overshoot when restarting). Anti-windup mechanisms (e.g., clamping the integral term when the pump speed reaches a limit) are essential. For lines with high measurement noise (e.g., from web flutter), a low-pass filter on the gauge signal is used; the filter time constant should be set to avoid filtering out true process variations. The choice of gauge also affects control: beta gauges have less noise than NIR sensors for pigmented adhesives, while NIR sensors respond faster because they can be placed closer to the die (no cooling required). For both, the gauge must be calibrated regularly using offline gravimetric samples. A typical calibration schedule is weekly for medical products, monthly for general tapes.
Profile control—adjusting the cross-web uniformity—is a separate closed-loop system. A scanning gauge provides a profile of coat weight across the web every 30-60 seconds. The controller compares each measurement point to the target profile (typically a flat line at target weight). For manual die bolt adjustment, the system may display the profile and recommend bolt adjustments (e.g., “Bolt 12: tighten 1/8 turn”). For automatic profile control, the system uses motorized actuators on each flexure bolt. The controller calculates the required movement for each actuator using an influence matrix—a mathematical model of how each bolt affects thickness at multiple points. This is a multi-input, multi-output (MIMO) control problem; simple “local” control (adjusting the bolt directly under a thick spot) may cause oscillations because adjacent bolts interact. Instead, the controller uses a decoupling algorithm or an iterative method that adjusts all bolts simultaneously based on the influence matrix. After each adjustment, the system waits for the profile to stabilize (1-2 minutes) before the next iteration. Typically, 2-3 iterations are enough to bring the profile within ±1.5% of target. For high-speed wide-web lines, automatic profile control is essential; manual adjustment would be too slow and less precise. The measurement system also provides data for statistical process control (SPC). The PLC logs the average coat weight, standard deviation, and profile flatness for each roll. Alarms can be set when the weight drifts outside control limits (e.g., ±3 sigma). This data is used for quality certification and for predictive maintenance—for example, a gradual increase in profile waviness may indicate the backup roll needs regrinding.
Practical implementation tips: Always verify the gauge calibration with a gravimetric sample at least once per shift. For NIR gauges, clean the optical windows daily; a thin film of adhesive vapor on the window will cause measurement drift. For beta gauges, perform a “source check” monthly to verify the source is not leaking. The control system should have a “manual mode” that allows operators to override the automatic control for troubleshooting or for coating the first few meters after a web splice (where the coat weight may be unstable anyway). Set up alarm limits: if the measured coat weight deviates from target by more than 10% for more than 10 seconds, trigger an audible alarm and automatically switch to manual mode (to prevent the controller from making large, possibly dangerous adjustments). Some advanced systems include a “feed-forward” term based on line speed: when the line accelerates, the pump speed is increased preemptively to maintain coat weight, compensating for the known relationship between speed and weight (higher speed = lower weight for same pump speed). This feed-forward reduces the error that the PID must correct, improving control. Implementation of near-IR technology to measure hot melt coatings speeds start up, increasing production, reduces scrap and allows for quality data archiving by roll or run. By properly integrating online thickness measurement into closed-loop control, hot melt coating lines achieve unprecedented consistency, enabling manufacturers to meet the tightest specifications while minimizing waste and maximizing throughput. For the hottest melt coating processes, where product quality is paramount and scrap costs are high, closed-loop control is not an option—it is a necessity.
