Project Intention
The idea is to explore different ways of colouring and creating graphics directly on a finished volume object, using spray paint (gun + compressor or airless to be defined). The heart of the project would lie in exploring all the games that can be played when spraying between a volume object and a graphic: playing with angles, edges, undercuts, nooks and crannies, offsets, superimpositions, inaccessible parts, interstices, the distance or angle of projection, mixing sprayed colours. On the borderline between graphic design and set design.

The use of a robotic arm to paint a volume may not seem original; we all have in mind the image of robots colouring bodywork on a car assembly line. But whereas these robots seek the perfection of a uniform colour, here the aim is to automate imperfections, to control the repetition of a variety of effects that could be created randomly by hand.

Although our origins lie in the world of objects and space, in recent years we have begun to explore the subjects of colour, graphics and printing processes. Confronting this research with the work of volume is a question that now challenges us.

Invited Designers
Pinaffo & Pluvinage

Project Title
Volumic Spray

Tools
UR10, Spraypaint, Stepper Motor, Arduino

Medium
UR10, Spraypaint, Wood

Software
Custom ad-hoc Software UR SVG Planner, programming via teach pendant.

Techniques
A custom 3D-printed tool was developed to actuate a standard spray can via the robot’s IOs. The tool is equipped with a stepper motor, which can be controlled via Arduino. The Arduino is connected to the robot’s IO pins through a relay. When a signal is issued on the IO, the Arduino causes the spray can to be pressed down. The movements were primarily programmed via teach pendant and free-drive mode of the robot. Some experiments also involved the custom software UR SVG Planner, which translates paths from SVG files into robot movements. The rotation of the stepper is translated into a linear movement that pushes down the cap of the spray can. Different spray can caps ensure different line widths.

Both graduates of ENSCI – Les Ateliers, Marion Pinaffo and Raphaël Pluvinage have been working together as a duo since 2015. Straddling the line between design and visual art, their collaboration produces a unique approach that spans multiple fields, including toys, interactive installations, objects, and scenography. Regardless of scale or context, their work always revolves around exploring aesthetics of manipulation and interaction. For the last four years, they have been observing, manipulating and experimenting with the technologies that surround us. Guided by curiosity and intuition, they analyze the physical systems and forms that operate in the background of the virtual world and explore the interactions that can take place between the mathematical logic of automata and human creative sensibility. Their graphic sensitivities, preference for intuitive uses, and commitment to economical means lead them to create a physical reality that is both surprising and accessible to all. Their work has been exhibited in France and internationally: Gaîté Lyrique, Paris; Frac Champagne-Ardenne, Reims; Frac Picardie, Amiens; Madd, Bordeaux; Mad, Paris; Tripostal, Lille; Fotokino, Marseille; Studio 13-16, Centre Pompidou; Design Museum, Holon, Israel; Triennale di Milano, Italy; Gent Design Museum, Belgium; CIVA, Brussels, Belgium; Mudac, Lausanne, Switzerland) and is part of the national collection of contemporary art (CNAP) and Mudac. They live and work in Toulouse, France.

What was the idea you originally brought to “A Third Hand”?

Our project involves using spray paint to colour and create graphics directly on an object. At heart of the project lies the exploration of all the games that can be played when spraying between a 3D object and a 2D graphic: playing with angles, edges, undercuts, nooks and crannies, offsets, superimpositions, inaccessible parts, interstices, the distance or angle of projection, mixing sprayed colours. The project is on the border between graphic design and set design.

The use of a robotic arm to paint a volume may not seem original – we all know the image of robots colouring bodywork on a car assembly line – but where these robots seek the perfection of a uniform colour, the aim here is instead to control and automate the imperfections, mixtures and effects inherent to spray colouring.

The aim of the workshop session was to produce as many samples as possible, so that a collection of objects could then be drawn (conceptually) based on the results of these samples.

To make the most of the reduced time available with the robotic arm, the project was carried out in 2 main stages: manual experiments (without a robot) in order to control as many parameters as possible beforehand, and experimentations with the robotic arm. In the first stage, we tested different brands of spray paint and the cleat structure. We tried around fifteen colours, looking for combinations, superimposing, etc. We also looked at about 10 different caps (the paint dispensing tips), at different distances and speeds, on a flat surface (cardboard) to produce a book of 70 samples on which to base future work. We then produced a few manual tests of spraying different lines, at different angles, onto the edges of the strips. At the end of all these tests, we carried out conceptual research into shapes and patterns that could be tested in the presence of the robot: around sixty potential experiments.

In the second stage, when we had access to the robotic arm, we began by getting to grips with the point-by-point programming mode. We carried out extremely simple tests (a line on a cleat) to refine our mastery of projection (taking in account variations in speed, acceleration, distance, angle and heading) within the limits of the project’s capacity. This produced about 30 small samples. When then produced complex samples with volume projection (combination of colours/patterns/etc. etc.), which resulted in about 40 samples.

Getting to grips with making simple movements was extremely quick (thanks to the point-by-point programming). And by projecting a simple straight line at a precise speed and distance, we were already able to continue the tests we’d started beforehand, which in reality were impossible for our human arms to carry out. Any artistic process that requires a movement that is too slow, too fast or too precise to be conducted by a human is worth exploring with such a robotic arm.

On the other hand, we hadn’t thought at all about the difficulty of programming an elaborate but precise movement in space. We had initially (wrongly) imagined that we would be able to export curves from MIMIC, for example, in order to control the robot and the layout of a volume pattern. This was not the case, and we ended up using a mix of point-by-point programming, or 2D tracing with the processing app, to simulate the different parts, which we could have done with 3D movements. Mastering the perfect coordination between the robot’s movements and the triggering of the paint spray was sometimes a bit technical. A delay of a tenth of a second or so between the signal being sent and the actual spraying of the paint can represent a big delay if the arm is moving at high speed. To sum up, there were no unpleasant surprises, but although you can quickly get to grips with the arm and experiment with it, you quickly realise that it’s a different kettle of fish to finalise the project completely and make a complete object, with all the details.

From an effects and results point of view, they were close to what we had hoped for in our conceptual research, but we were able to refine the patterns that worked best, combine them and so on. Drawing in 2D was also exceedingly easy, thanks to the processing programme that had already been designed.

Programming the arm so that it shakes the spray can before the drawing commences – so that you don't have to take it apart to do it manually – really made us smile. Joke aside, some of the effects of shading, mixing patterns, the way a 2D line is projected onto a volume shape, etc., were still difficult to imagine in detail. Each experiment was always full of suspense.

The tool for holding and activating spray paint cans opens is immediately useful, including for 2D use, especially because these cans are often standardised. It could therefore be useful with other types of paint (fabric paint, glass paint, and so on), but also treatments (varnish, effect inks, etc.). Beyond our formal experimentations, we can foresee semi-craft applications for spraying objects (including single-colour) in small series.

Later down the line, if and when a programming tool is developed for a 3D equivalent of the current Processing tool (which transcribes a 2D vector path into robot movement), it would have countless uses. It would enable quick testing, both for marking on a volume (spray painting, but also using a brush, even machining a block) and for other applications: drawing a curve in 3D software that the robot can follow would open new possibilities for many projects, and in some cases be much more efficient than point-by-point programming.

We’re continuing the project to create a collection of objects based on the results of these samples, which we could potentially fabricate in front of a live audience. This project is both in line with our practice of exploring industrial colouring and graphic design processes, and at the same time new: it allows us to move from 2D objects to 3D graphics on the one hand, and on the other it would allow us to set up a form of self-production (mini-series).