Direct answer: collaborative robot dispensing succeeds when part location, nozzle pose, motion, and material delivery stay inside the same process window. Lock the fixture datum and bead limits, verify standoff and tool angle, coordinate speed with pressure or flow and valve timing, inspect the deposited material rather than only robot completion, and define how an interrupted bead is held, cleaned, reworked, or rejected. A repeatable path cannot compensate for unstable material.
Dispensing is a coupled process. The robot supplies position and velocity; the delivery system supplies conditioned material through a valve and nozzle; the fixture controls where the part sits; and the downstream operation may depend on open time, cure, compression, or assembly timing. Review those elements as a datum-path-flow ledger so a fault can be traced to the correct layer.
Define the required bead as a measurable output
Start from the drawing or process specification for bead location, width, height or volume, continuity, start and stop limits, corner behavior, and any permitted gaps. Add the functional purpose: sealing, bonding, potting, lubrication, thermal interface, or another use. That purpose helps the responsible process team choose the material and acceptance method; it should not be guessed from a demonstration.
Mark surfaces that establish the part datum and surfaces that receive material. The fixture must repeat the relationship between those surfaces and the robot frame. A part-present sensor may show that something is loaded while leaving the working face offset or tilted. Use locating features, seating evidence, or measurement appropriate to the tolerance and consequence.
Create a ledger for each recipe:
| Ledger layer | Controlled values | Evidence at release |
|---|---|---|
| part | identity, orientation, datum, surface state | seated and correct workpiece |
| tool | nozzle type, pose, standoff, wear or cleanliness | active tool record and check |
| motion | path revision, speed zones, corner rule, approach and withdrawal | completed approved trajectory |
| flow | material identity, conditioning, pressure or flow, valve timing | delivery state and monitored response |
| result | bead location, continuity, dimensions or functional inspection | disposition linked to the part |
Establish nozzle pose for collaborative robot dispensing
Standoff and tool angle affect where material lands and how it forms. Validate them across the part tolerance, fixture repeatability, robot posture, and nozzle length. Include the nozzle body, hoses, dress pack, nearby clamps, and the deposited bead in collision and contamination checks. A path that clears a dry part may contact a raised bead on a return move.
Define approach and withdrawal to prevent drips or strings from crossing an accepted surface. If the process uses a purge location, drip tray, tip wipe, or calibration target, include those movements in the cell model. The service station must remain reachable with the same tool and safeguarding conditions as production.
Corners deserve a separate motion rule. Robot speed can decrease because of path blending or curvature, while material delivery may continue at the previous state. That mismatch can create excess deposition. Conversely, an aggressive blend can cut inside the required corner or increase standoff. Choose corner speed, path tolerance, and flow response together, then inspect the physical result.
Treat the material as a changing process input
Record material identity, batch where required, storage condition, temperature, mixing or conditioning, pot life or working life, viscosity indicators where available, supply container, and the approved time in the delivery system. For two-component materials, include ratio evidence and what happens after a pause. For materials sensitive to air, moisture, or temperature, define the controls and the point at which a prepared supply must be discarded or reconditioned.
The material route includes pump or pressure vessel, regulators, hoses, valves, nozzle, and sometimes heating or metering equipment. Volume at the part can lag a command because the system has compliance, restriction, or compressibility. Measure start delay, stop behavior, and response to speed changes with the actual delivery path. A valve-open signal reports a command or component state; it does not automatically prove that the required material reached the workpiece.
Universal Robots’ dispensing application page highlights the relationship among programmed path, robot speed, pressure, and material application. Use that relationship as an engineering prompt, not a universal recipe. The material supplier’s data, process-owner requirements, delivery equipment, part surface, and application trials determine the valid operating window.
Synchronize starts, straights, corners, and stops
Break each bead into events. The robot approaches a lead-in point, material flow establishes, the tool crosses the defined start, speed reaches its nominal range, corners or curves modify motion, the end condition is reached, flow stops, and the tool withdraws. Record whether the process uses time compensation, distance-based events, pressure or flow feedback, or a measured bead response.
Use an event table during trials:
| Event | Possible mismatch | Test observation |
|---|---|---|
| start | motion begins before material reaches the nozzle | gap or tapered bead at the entry |
| acceleration | flow remains high while speed changes | local excess after the start |
| corner | robot decelerates without matching delivery response | bead grows or shifts at curvature |
| stop | valve command ends but stored pressure continues delivery | tail, drool, or contamination |
| pause | material condition changes in the stationary nozzle | cured tip, skin, separation, or unexpected restart behavior |
The goal is not to eliminate all transient behavior by assumption. It is to define where transients may occur, measure their effect, and keep the accepted bead inside the project limits.
Inspect the deposition, not the path completion bit
Select inspection from the function and consequence. It may include visual confirmation, bead-presence sensing, profile measurement, mass or volume correlation, camera inspection, pressure or flow signatures, downstream leak or bond tests, or a sampling plan. State which defects each method can detect. A camera that sees continuity may not establish internal mix or adhesion; pressure monitoring may reveal supply loss but not exact bead placement.
Bind the result to the part and recipe. Store enough information to answer which material, tool, program revision, fixture state, and interruption history produced the bead. If a station only reports “robot complete,” downstream equipment cannot distinguish an accepted path from a cycle that completed after a material fault.
An unknown result should move to a hold route. Define whether the part may be reinspected, cleaned and redispensed, manually repaired under an approved process, or rejected. Consider surface damage and timing; rework can be invalid after material cures or contaminates the substrate.
Plan cleaning, replenishment, and changeover
Production stability often depends on tasks outside the automatic path. Document container change, material warm-up or conditioning, purge quantity, nozzle cleaning, tip replacement, hose maintenance, fixture cleaning, waste handling, and the validation required after each task. Include consumables and cleaning time in the capacity estimate.
For recipe changeover, identify which parameters and physical components change: part fixture, datum, nozzle, material supply, pressure or flow settings, path, inspection limits, and downstream timing. Use positive recipe and material confirmation where a mismatch would create a hidden defect. A saved robot program name should not be the only link between the part and delivery setup.
Design the interrupted-bead response before production. On a protective stop, supply fault, or robot fault, decide whether flow is shut off, pressure is relieved, the tool may withdraw, the part remains in place, and the bead is automatically unacceptable. Restart should not resume at an arbitrary path point unless the process owner has validated overlap and surface conditions.
Apply collaborative safety to the complete application
The word collaborative does not make a dispensing cell safe by itself. ISO 10218-2 and ISO 12100 require application-level risk assessment and risk reduction. Evaluate robot and fixture motion, pinch points, sharp or hot tooling, pressurized material, chemicals, unexpected discharge, hoses, stored energy, and access during teaching, cleaning, replenishment, nozzle service, inspection, and fault recovery.
Read the material safety data and site requirements for handling, ventilation, skin or eye exposure, waste, and spills. Determine which work may occur with energy available and what protective measures, modes, speeds, force limits, separation monitoring, guards, interlocks, or procedures are required. Validate the implemented functions on the real cell.
If people share space with the robot during an intended task, define the task posture and contact possibilities with the dispensing tool and material present. A force-limited arm does not reduce chemical, injection, sharp-tip, or unexpected-flow hazards. The nozzle and supply system can determine the risk-reduction concept.
Commission across the process window
Use representative parts at the limits of datum and surface condition. Test fresh and aged-but-permitted material, low and high allowed supply conditions, straight paths, tight curves, starts, stops, direction changes, and the longest pause. Measure the bead and the synchronized process signals. Repeat after nozzle service and material replenishment.
Force faults: no material, low supply, blocked or worn nozzle, incorrect part seating, wrong recipe, valve or robot timing disagreement, lost inspection, and a stop mid-bead. Confirm that the part enters the correct disposition and that restart requires the intended checks. Record the cleanup burden and time; exception handling is part of the production cycle.
NIST’s performance-assessment approach supports observable requirements and repeatable test methods. Report bead results, event timing, fault response, inspection coverage, rework rate during trials, and complete cycle segments. Do not substitute nominal robot repeatability or an isolated path time for dispensing performance.
Inputs for a dispensing-cell review
- part drawings, datum scheme, surface preparation, bead geometry and functional acceptance criteria, representative samples, and fixture concept
- material technical and safety data, storage and conditioning, working-life constraints, supply and metering equipment, hose and nozzle configuration
- path and tool frames, standoff and angle limits, speed and corner rules, valve timing, pressure or flow signals, purge and cleaning stations
- inspection method, result identity, hold and rework route, downstream open-time or press requirements, traceability, and target cycle distribution
- operator tasks, access, collaborative or guarded concept, risk-assessment inputs, maintenance plan, waste handling, and changeover frequency
How this guide was prepared
In practice, the editorial review checked the retained dispensing sequence at 00:02.4, 00:14.4, and 00:24.0. The first view shows a light-blue robot and a slender process tool above an open gearbox housing; the second shows the tool near a different section of the housing perimeter; the third shows the tool displaced again while the located housing remains in view. These visible changes support discussion of datum, nozzle path, access, and motion-to-process coordination. They do not establish material identity, bead quality, collaboration safety, production rate, customer identity, or acceptance. Official robot-cell, machinery-risk, dispensing, and measurement sources define the broader review boundary.
Citation-ready statements
- According to Universal Robots’ dispensing application guidance, dispensing automation combines programmed robot motion with a process tool. EVST addresses this by coordinating datum, nozzle pose, path speed, valve timing, material delivery, and inspection.
- According to ISO 12100:2010, machinery risk reduction follows hazard identification and risk evaluation. EVST addresses this by assessing the complete application, including the nozzle, pressurized material, fixtures, stored energy, cleaning, and operator access.
- According to the NIST Performance Assessment Framework for Robotic Systems, performance assessment needs measurable procedures and metrics. EVST addresses this by recording part, material, nozzle, program, interruption, bead measurement, and disposition together.
About the editorial team
The organization integrates industrial and collaborative robots with fixtures, process equipment, controls, sensing, and safeguarding for manufacturing applications. Its engineering content separates robot motion from process acceptance and asks project teams to close material, interface, inspection, maintenance, and recovery conditions with application-specific evidence.
The application team can review a declared part family, bead specification, material-delivery system, tool path, inspection concept, and risk-reduction tasks as one application. Final material settings and acceptance remain with the responsible project, material, process, safety, and quality functions.
Related engineering resources
- Glue-dispensing application equipment
- Appliance glue cell: bead, datum, and press window
- Collaborative robot application planning
Frequently asked questions
Can constant robot speed guarantee a constant bead?
No. Delivery also depends on material condition, pressure or metering, hose and nozzle behavior, valve timing, standoff, part surface, and transient response. Constant speed removes one variable; the deposited result still requires measurement.
How should a corner be programmed?
Choose the geometric tolerance, robot blend, local speed, tool angle, and delivery response together. Trial the tightest approved corners and inspect for inside cutting, bead shift, excess material, or starvation instead of applying one universal corner setting.
What happens to a part after a mid-bead stop?
Place it on hold and apply the defined process disposition. Confirm material and tool condition, time since deposition, surface state, and whether cleaning or rework is allowed. Do not resume from a controller line unless the overlap and resulting bead have been validated.
Is a collaborative arm enough to allow open access?
No. The complete application must be assessed, including the nozzle, pressurized or hazardous material, fixtures, stored energy, operator tasks, and possible contact. The selected protective measures must be implemented and validated for the real cell.