As makers of large and complex 3D parts looked toward UV curing, they found the established techniques of placing fixed lamps end-to-end to be fraught with technical problems and high price tags. Many projects, where the coating was developed and proven, stalled when the price quotation for a system of 10, 15, or 20 UV lamps was presented.This article describes recent developments using robotically actuated UV lamps to cure large and complex parts. Both the technical and economic benefits of this approach are described and compared to the traditional approach of using large fixed-lamp arrays.
UV coatings remain attractive because of their scratch and mar resistant characteristics, rapid process speed, and the environmental friendliness of UV technology.
In describing the attributes of UV coating on their Model U concept car, Ford Motor Co. observed that “environmental concerns in manufacturing are also addressed with a new UV-cure clearcoat system developed by AkzoNobel. Clearcoat is the topmost layer of a vehicle’s paint. It gives a vehicle its shine and protects the paint from damage. During the clearcoat cure, the Model U was exposed to ultraviolet light rather than to the high temperatures that are used traditionally. This system provides a harder finish and means the Model U will be more resistant to scratches than most cars and trucks. The process eliminates the need for a bake oven and uses less energy and solvents than traditional systems.”1
The UV curing industry has evolved over the last 20 years from predominantly flat, geometrically simple and symmetric applications (like paper, floor tiles, wood panels, optical fiber, and DVDs) to complex, three-dimensional shapes (such as UV cure composites and automotive refinish primer/surfacers). This evolution requires a fresh approach to what equipment is most appropriate and how to best cure these non-traditional parts.
Since a 100% UV cure mechanism depends on each facet of the part receiving equal exposure to the UV light source, the challenges of curing something as big and complex as, say, a car body, are formidable.
Historical Approach to UV Curing of Large Parts
The usual method of curing proposed to date has been to use a large number of fixed-position lamps. The lamp positions are pre-set to provide uniform illumination over the entire part surface.
I employed this technique as director of the 2001 Team UV project, which produced a UV-coated racecar (see Figure 1).2
In order to achieve the required uniform exposure, a common procedure involves positioning several radiometers on the parts’ complex surface and making a series of iterative trials, and fine tuning the position of lamps after comparing the radiometric data after each trial run.
To reduce the time of trial-and-error, a method has recently been proposed that relies on a sophisticated computer simulation to model the exposure of many fixed-position lamps needed to create uniformity.3,4
While this approach may expedite the painstaking process of empirically determining lamp position with a radiometer, the model is extremely complex and does not appear to take into account all factors, such as reflections, advance curing as the part moves into the curing tunnel, and other subtle effects that are difficult to model mathematically. Since this approach “imputes” lamp positioning by considering the combined effects of “a thousand points of light,” it does not provide much help if the actual measurements do not coincide with the model. The user, faced with the practical problem of what to change, is back to an empirical, iterative, solution.
Fixed Lamp Arrays
The proposal to cure large surfaces using many fixed position lamps is attractive to the lamp supplier but presents difficulties for the user.
Alignment of lamps: Since each fixed lamp has a finite, linear footprint, only a “best fit” can be hoped for on a complex curvilinear surface. The tradeoff is obvious. If the footprint of the lamp footprint is made smaller, the more fixed lamps that are required, but the better fit can be achieved.
Figures 2 and 3 are examples of utilizing a ray-tracing program, illustrating how an array of fixed lamps can produce relatively uniform irradiation of a curvilinear surface.
While this method is satisfactory for parts where the target distance to the lamp is constant, using fixed lamps is made more difficult by the fact that usually the part is being conveyed through a tunnel of lamps. For example, in Figure 4, the car body does not remain at a fixed target distance to the lamps. And since the entire body must pass before all lamps in the tunnel by conveyor, there can be significant variation.
This means that not only must a “best fit” be developed for any given surface of the body, but an overall best fit must be achieved for the entire body as it is processed. Clearly this is a challenging problem.
Then, assuming such a fit can be achieved for a given part configuration, it will be necessary to derive an entirely different arrangement for another body style, making setup an enormous undertaking. Some users have expressed concern over how to cure components which face away from the lamps (the underside, a difficult top edge, the interior surfaces).
Perhaps a simple “gut-check” in considering whether the idea of fixed lamp curing of a coating makes sense is to ask whether it is worthwhile to apply the coating using a similar arrangement of fixed applicators. In most cases, large and complex parts are coated by hand or with automated guns on reciprocators or paint robots.
Capital cost: The use of multiple lamps carries a financial burden due to redundancy and inefficiency. Typically each lamp requires its own power source, cooling apparatus, mounting fixture, controls, etc. One way to mitigate this problem is to use individual lamps with as large a radiant footprint as practical. This approach is only possible with electrode (or arc) lamp technology since microwave UV lamps are currently restricted to 10” or less in footprint. But using larger arc lamps reduces the lamp lifetime and UV uniformity. It also entails larger power supplies and cooling systems, including water cooled lamp modules.
It appears that a “one (lamp) size fits all” approach is not the best solution for auto bodies. Some surfaces can be treated very effectively with large lamp lengths, while others might require smaller sized lamps to accommodate rapidly changing curvatures. Having one size tool is therefore inefficient.
System maintenance and SPC considerations: Another undesirable aspect of using a large array of lamps is the challenge of maintaining and monitoring a large number of discrete devices. What is the proper procedure when a single lamp degrades or fails due to aging? If a new lamp is installed on an ad hoc basis each time, then there will eventually be varying intensity levels among the irradiators in the array.
By analogy, what should a car owner do when the first spark plug wears to the point of replacement—replace the entire set or just the deficient plug?
This raises a further question of whether to individually monitor the output of each lamp module. It is possible for a single lamp to fail and potentially go unnoticed—producing parts that may not have adequate cure. Of course, the technology exists to monitor and even close-loop control lamp modules to maintain consistent output, but the cost of such monitoring and control for very large arrays of lamps may be expensive.
As was touched on previously, lamp maintenance will necessarily disturb the position of lamps, which must be put right again. It has been suggested that the lamps could be mounted on small, motorized micro positioners, but the control and capital cost of implementing this on large arrays may not be practicable.
Robotic UV Curing
For many years, numerous attempts to use robots to manipulate UV lamps have been attempted with varying success. As pointed out by one lamp supplier, “there are issues that need to be considered when using a robot. First, the lamps must be sufficiently robust to withstand the acceleration and de-acceleration swings of the robot arm, and the lamp must be able to operate efficiently and reliably in a variety of different positions. Finally, the robot must be programmed to ensure that it delivers the correct UV energy to all parts."5
Recent investigations by Daimler-Chrysler into the use of robotic UV curing for automotive coatings have correctly identified significant challenges related to the process cure window, noting that “if UV technology is to be transferred to the production process of a vehicle painting line, then one should be able to calculate the hardening lines and the movements of the hardening movements. Simulation tools are needed for this purpose."6
In 2004, a group of companies formed the North American Automotive UV Consortium to develop these and other missing tools and techniques to advance robotic UV curing. The group initially developed the following “roadmap” to guide the team’s development efforts:
- Development of UV sources suitable for robotic use;
- characterization of the output of these sources (i.e. the radiant “footprint”);
- development of off-line programming simulation tools for light path programming;
- development of tools and techniques for on-line validation of simulations;
- cure test studies with coating suppliers; and
- collaboration with car makers on pilot and production scale programs.
Development of robotic UV sources: Many UV sources are not ideal for robotic applications. They are too complex, unstable, heavy, or require too many interconnections to be mounted on a fast-moving robot arm.
A compact arc lamp source was developed for robotic applications. This unit weighs approximately 18 pounds, which makes it suitable for use on a wide range of industrial robots with capacities in the <10 kg (22 pounds) range, thereby keeping the cost of the robot to a minimum while offering a broad selection of units to choose from.
The lamp contains few electronic components that are susceptible to damage or variation during rapid acceleration. A shutter is provided for both the safety of the operators and to provide full powder to the part within a few milliseconds of electronic shutter triggering, thus allowing the car body to be in position before beginning exposure to UV energy. A minimum of hoses and electrical connections makes mounting of the lamp to the robot simple and keeps interconnections from becoming accidentally twisted or entangled during lamp articulation.
A second source was used for testing consisting of a UV LED array (Figure 5). While the UV output of the LED array is somewhat lower than traditional arc lamp sources (maximum of approximately 2 W/cm²), UV LED technology is rapidly developing. The advantages of the LED array are its extremely long lifetime (>30,000 lamp hours), the instant on/off capability of the device (2 ms from off to full power), and that the array emits no direct heat to the target. The output of the array is a narrow bandwidth, falling from 385 to 405 nm. One advantage of robotic manipulation of the UV LED array is that extremely close (~1.0 in.) target distances can be maintained, which provides higher average peak irradiance than could be achieved with fixed positioning of UV LED arrays. The results obtained in lab trials are very encouraging.
Characterization of the UV Source
The radiant energy profile of the UV arc lamp source was “mapped” to accurately determine the footprint of the lamp. This footprint allows a UV robot “tool” to be created for the off-line simulation software.
A model of the lamp output can be described quantitatively in the x, y, and z axes. Thus, the proper orientation and target distance of the lamp can be used in the robot off-line simulation. Proper rotation of the lamp can also be programmed so the lamp is kept normal to the tangent of the surface at all times—a capability that is not possible without articulation of the lamp.
Off-Line Simulation Tools
An ongoing effort of the consortium is the development of simulation software for off-line light path development and analysis. This will permit car makers to develop and fine-tune curing paths without interrupting production.
Modeling the UV lamp tool allows for programming paths with proper overlaps to minimize potential “striping” of the part while achieving maximum uniformity in the fastest production cycle time.
The program also includes the ability to track the conveyor in real time. This allows paths that minimize the effects of “mapping” or pre-curing of coating due to advanced exposure to UV light.
On-Line UV Trials
A number of on-line trials have been conducted at the Fanuc Robotics facility in Toledo, Ohio. The goal of these trials was to evaluate the performance of the UV source, accuracy of the model, and to evaluate the effects of robot arm speed, conveyor speed, part presentation, and other variables.
Radiometric data were collected using a novel device. The multi-sensor data acquisition unit allowed the consortium to collect UV data from various locations on a complex surface. Sensors were embedded into locations that were predicted to be difficult to cure with fixed lamps.
The robot program was then fine-tuned to achieve uniform peak irradiance on an automotive door panel moving at a line speed of 12 rpm. Once uniform peak irradiance was established, robot variables were tuned to achieve equal UV dose. Power output of the lamp was kept constant for all testing. The unit is capable of producing 500 W/in at full power.
Another advantage of the robotic technique is the target distance to the part can be set and maintained at the optimum distance for the reflector design. For many lamp units, especially those using elliptical reflectors designed to focus to a line, the focal lengths are relatively close (typically around 2 inches from the face of the lamp). This means the lamp must be operated out-of-focus (in what some refer to as the “far field”). While this is common practice, it is also inefficient as the power falls off rapidly in the far field.
Figure 6 shows the process of fine tuning the light path for consistent peak irradiance during one of the earliest line trials. The total time to tune the system so peak irradiance is kept within a narrow range is estimated to be less than one hour.
Of the six steps outlined in the roadmap developed by the consortium, solid progress with encouraging results have been obtained from efforts on the first four steps.
- The UV lamp designed for robotic use is a sucessful development. The unit is lightweight and agile and, therefore, posed no obvious problems in use. The shutter system was an important safety feature for frequent trials.
- Several improvements will be implemented in the next generation of lamp design. There were also several ideas for improvement in how to integrate the lamp unit to the robot.
- Off-line programming work is underway and already yielding positive results. Lab trials identified many features, which can be added to the programming.
- On-line data collection using the 3DCURE unit was successful in allowing the team to rapidly develop paths that yielded uniform peak irradiance and energy density. Several improvements to the data collection system are being implemented to make higher speed data collection easier.
- The radiometric data indicate sufficient peak irradiance and dose can be achieved with the line speed (12 fpm) and robot arm speed (600 mm/sec.) that were used during the trials to affect proper cure of commercial formulations. (Based on baseline cure data provided by coating formulators). This opens the doors to steps five and six of the roadmap that will involve curing of coatings at production cycle times.
Cost Model Development: Robotic versus Fixed-Lamp Systems
The North American Automotive UV Consortium has developed an interactive cost model that provides comparative capital and operating cost data needed by manufacturers. Figure 7 is a summary of cost model results for replacing variable power, 600 W/in-lamps with a similar output robotic curing cell.
The results presented here are for a simplified model, where capital cost is based on list costs of all equipment under study, published energy consumption, and replacement parts costs. The model doesn’t attempt to quantify “soft costs” involved in equipment setup times, floor space consumption, downtime, etc., which appear to favor a robotic approach. The capital cost comparison is sensitive to the cost of the robot, since this is a relatively expensive component.
The following comparison anticipates a $60K robot and associated hardware needed for integration. The model is also sensitive to the cost of fixed lamps. The comparison below was computed using microwave powered 10-in lamps rated at 600 W/in with variable power supplies since the literature suggests that variable power may be necessary to achieve the required uniformity and to provide various monitoring features.
While numerous scenarios have been evaluated, a few trends are already clear.
First, the operating cost of a single robot lamp is always less expensive than a multi-lamp array. This is due to the lower parts replacement requirements and the lower energy consumption.
Second, the capital cost of eliminating fixed lamps with a robotic cure system is higher until a threshold number of fixed lamps are replaced. In the model presented here, the robotic system has a lower capital cost once an array of five of more fixed lamps is replaced. If more exotic equipment is anticipated (such as lamp monitoring, or micro positioners for fixed lamps), then the robotic system may offer capital savings compared to even smaller arrays. Conversely, if exotic robotic equipment is installed, it may make the robotic system more costly to install.
While good progress has been made leading to several improvements in the tools and technology for robotic UV curing, there is a still a steep development curve to be tackled:
- Continued refinement of the UV lamps sources, including UV LED developments;
- improvements in the off-line simulation software to include subtle variables, observed in on-line trials, so that off-line simulation and real world cure experience are related as closely as possible;
- improved data acquisition tools will allow us to more accurately program and measure UV irradiation to further define the process window;
- refinements to the first four steps of the six-step roadmap will lead to future expansion of the testing to include actual curing of coated parts under simulated production conditions (e.g., cycle time);
- one outcome of the work to date has been the formation of a new company (UV Robotics, LLC) that will specialize in the integration of UV lamps, robots, and other process equipment and controls for end-use applications7,8; and
- the group anticipates expanding its membership to include coating suppliers and tier-one and OEM automotive partners.
Growing List of Potential Uses
While the goal of this work is to provide car makers with a set of tools that enable use of UV coatings, the work clearly has implications for the tier-one producers and other industrial processes.
In one example (Figure 8), a single robotic lamp was able to cure a clear coating on an automotive headlight lens that is currently cured using 12 fixed-position microwave lamps. The UV energy required for the task was significantly reduced (from nearly 7.0 to 2.5 Joules). This dramatic efficiency is possible since the light is used more effectively.
In another popular application (Figure 9), new PVD metallization system for alloy wheels requires that UV be cured in the difficult recesses of the wheel.
The ornate cutouts, popular with high-end customers, are difficult to cure with fixed lamps since many shadow problem areas exist. The robotic UV approach solves this problem nicely.
As UV coatings continue to find their place alongside traditional coatings for a wider array of parts, the need for economical curing systems will become more prevalent. Robotic UV curing makes sense both from a technical and an economic standpoint and is likely to grow in popularity, especially for those with larger and more complex parts that cannot justify large fixed lamp designs.
- Ford Motor Co., 2003 Detroit Auto Show Model U press release, www.ford.com.
- Mills, P., “Team UV: Leading Edge Technology for the Automotive Industry,” Proceedings, Radtech Europe, Basel, Switzerland, 2001.
- Joesel, K., “UV Curing of Automotive Clearcoats–3D UV Curing Simulation,” SURCAR 2003.
- Schneider, M., Klein, W., and Schroder, C., “Optimized Position of 3D Lamps for the Treatment of 3D Work Surfaces,” Fraunhofer IPA, Proceedings, RadTech Europe Conference and Exhibition; 1999.
- Skinner, D., “3-Dimensional Curing: A New Role for UV Curing in the 21st Century.”
- Raith, T., Bischof, M., Deger, M., Gemmler, E., “3-D UV Technology for OEM Coatings,” RadTech Report; November-December 2001.
- North American Automotive UV Consortium, Consortium Report, Volume 1; Winter 2005.
- North American Automotive UV Consortium, Consortium Report, Volume 2; Winter 2006.
This article is based on the work of a consortium; a team of talented and dedicated specialists. I wish to acknowledge the contributions of the following companies and individuals:
Oliver Treichel and Oliver Starzmann of IST Metz GmbH; Dennis Kaminski of IST America; Mark Owen, Tom Molamphy, Jon Marson, and Alex Schreiner of Phoseon Technology; Keith Torp, Jerry Perez, and Ed Walczak of Fanuc Robotics, North America; David Snyder and Kyle Bostian of EIT Instrument Markets; and John McDonough and Renny Wolfson of UV Robotics LLC.
Paul Mills is the presdient of UV Robotics, LLC, in Cleveland, Ohio. He may be contacted for additional information via e-mail or at (440) 570-5228.