Gaggione, optician, French plastics manufacturer, specialised in producing high quality optics associated with LEDs, is one of the world’s top three companies in the sector, if not the very best, thanks to over half a century’s experience in plastics manufacturing, its staff and its R&D tool resources.
It has a long history of working with Ayrton, thanks to which we have obtained exceptional permission to meet engineers and visit the plant.
And we can measure just how many factors might not only degrade dispersion, but also flow and color mixing throughout the collimator design and manufacturing process.
Let’s head for Plastic Vallée in the Ain administrative department along with Yvan Peard, Ayrton director, where we are met by David Veryser, sales director of the optics division.
Our aim is to find out more about the company that often, if not systematically, remains in its customers’ shadow, like a closely guarded secret, with the exception of Ayrton with whom they have formed a reliable partnership over time, based on trust.
Subsequently we’ll visit the plant and take photos under close surveillance because, in 60 years of plastics manufacturing experience, manufacturing secrets have to be kept under lock and key.
A little history lesson to start with:
[private]
The history of Gaggione
We are in the Jura Mountains, between Lyon and Geneva in the famous Oyonnax valley. This is home to plastics manufacturing in France that reached its peak injecting plastic for the car industry until low cost countries came on the scene with lower prices. They say that there used to be more Ferraris in this valley than anywhere else in France!
Pierre Gaggione was a mould maker. Born in Italy, he came to France after World War Two and set up a mould manufacturing company in the back of a garage. At the time, plastics manufacturing broke down into two professions: people who made moulds and people who injected plastic into the mould.
Over the years, Gaggione became a company specialising in injecting very thick transparent plastic, meeting high demands from the cosmetics and luxury market: gift boxes, perfume bottle stoppers… and working with thick transparent plastic led them towards light optics.
The company was then family run by Pierre Gaggione before being taken over in 1999 by Babylone, a holding company.
Just part of the Gaggione team of course as the plant works 3×8.
Gaggione, all around you for the last 65 years
Do you wear perfume by Chanel, Nina Ricci, Versace? You’ll certainly have handled a Gaggione packing. The pretty famous coffee capsule box, smooth and transparent like glass, is also a good illustration of this industry’s know-how. Today, Gaggione limits its field to optics and packaging. It progressively left the luxury market, also recognising that they had forged their experience in plastics manufacturing.
A Plasticase mould.
Packaging involves custom-built cases, made for a great French brand tooling, for example who guarantee their tools for life. Gaggione ensures that the plastic hinges can be opened 100,000 times without deteriorating.
Under the e-plasticase brand, Gaggione has also developed a range of standard cases that can be personalised with customer logos. These are communication tools intended for professionals from all fields of activities : tooling, measuring instruments, medical, first aid, auto…
In optics, dedicated exclusively to lighting, Gaggione provides custom-built solutions used indoors, outdoors, for architecture, shows, public outdoor lighting, in industry and on railways. We find them on high speed trains (reading light lenses), on public roadways (traffic lights), on the motorway (safety panels), on aeronautical, in the medical field (lights in operating rooms).
But let’s not get carried away…
Stages focussing the company
1996, 1st Lumileds LED, injection of the world’s 1st collimator
In the 1990s, Gaggione was a plastics manufacturer, as explained by David Veryser, sales director of the optical division.
David Veryser: “Back in 1996, the first power LED in the world was put on the market by Lumileds. It was called Barracuda and then Luxeon. Philips designed an optic called a collimator to focus the light from this source that emits over a half sphere and they asked us to inject it.
So it was in 1997 that Gaggione brought out the world’s first collimator. Working from a study, we knew how to control the shapes with very fine tolerances. We knew how to inject by controlling the shrinkage. We became a plastics manufacturer working in optics and we focus our strategy towards the development of this promising market.”
1999, Gaggione was taken over by the Babylone Holding Company
Babylone owns 3 companies: Gaggione for engineering and thermoplastic injection, Surcotec in Geneva is specialised in engineering and surface treatment, and Quadratec in Montreal Quebec in thermoplastic injection.
David Veryser: ”Babylone bought Quadratec a year ago to get their foot in the door in North America and save some time on deliveries.
It’s a small company, employing around 15 people, evolving on a ruined market. For them, optics gave them the chance to develop a niche sector and for us, it gave us a foot in the door locally. They kept their own production, separate from optics.
The engineering is done in France and they have two injection machines kept aside for optics.
The hybrid reflector LEDnLIGHT, born of a collaboration between Gaggione and Surcotec. (Photo Daniel Gilet)
Surcotec in Geneva carries out metallization. They work in clock making, medical, luxury and optical. We found a common product in a hybrid reflector. We provide the central optic design for a wide diameter LED and they provide the reflector metallization. Surcotec has developed silver metallization that ensures 95% yield compared to 85% with aluminium.”
2005, bringing optical engineers on board
However, investing in the world of optics implies bringing scientists into the company, principally optical engineers.
David Veryser: ”The transformation came about from 2005 when we recruited Jean-Pierre Lauret to design optics. From this moment on, Gaggione became an optician that was going to use plastics manufacturing to make products.
That also means monitoring LED technology, remaining in close contact with manufacturers; this involves understanding how the product is going to interact in a complete system that has thermal, electronic, binding and colour mix problems”.
2006, naissance du catalogue LEDnLIGHT
2006, arrival of the LEDnLIGHT catalogue
Part of the 500 references in the LednLight range. (Photo Daniel Gilet)
David Veryser: ”When I was at Philips (Before joining Gaggione, David Veryser worked for Philips Lighting for 14 years), we commissioned Gaggione to develop custom-built solutions. But for some urgent projects, Philips reluctantly had to work with their competition. The time required for an optical study (3 to 6 weeks), making a mould (8 to 12 weeks), testing the first parts, classification….
Sometimes you have to go fast and take standard optics off the shelf. This was what pushed Gaggione to develop their first range of five standard optics.
Today, the LEDnLIGHT catalogue offers over 500 different references and continues to develop more.
SLU : Therefore, for each LED reference, Lumiled or Seoul at the time, there was a collimator?
David Veryser: Yes, that was our decision, for each LED there would be an adapted optic or a mecanical interface called a Holder, a support that refocuses the optic in line with the LED unit. Today, the chips are almost all on ceramic bases with the same thickness therefore when we develop an optic, it is generally compatible with most LEDs from the same market segment.”
Yvan Peard: It was at this time that we adopted the first Gaggione optic for the Seoul P4 in Ayrton projectors (Easy color, Moduled…) because Seoul was clever enough to bring out the P4 with a K2 compatible unit.
2007: Diamond machining
It was the arrival of the diamond machine that would allow Gaggione to reach almost unheard-of quality levels.
David Veryser : ”In optics, we talk about tolerances in terms of micrometers, hundreds of nanometers in visible light.
To be respectable in the optics industry, a mould impression must be machined to these scales. The mould impression is the part of the mould that replicates the shape of the product. Diamond machining ensures roughness accuracy between 1 and 10 nm. The mould does not require polishing because this might deform it. It is already smooth. The template mould is close to perfection. Afterwards, we have to know how to control the shrinkage of the material during the injection phase.
Back in 2006, under friendly pressure from clients, the Gaggione CEO was ready to invest in diamond machining. We’re talking about 1 million dollars. That’s the price of a Formula 1 racing car! But before putting in the order, they had to find a driver, their very own Schumacher. This was David Gluchowski who sought us out (Gaggione has a good reputation). He was using this technique for a confrere. He sent us his CV almost with the quote for the digital control machine (he laughs). It came from Moore Nanotechnology Systems in USA (Nanotech 350FG 4 axis).”
The Ayrton/Gaggione partnership
The new NandoBeam 302 sophisticated optic designed by and only for Ayrton.
Ayrton was one of the Gaggione’s first collimator customers, more precisely ever since the Luxeon: that creates strong ties.
The P5 20 mm optic will follow because, with the appearance of multi-chips to which standard optics did not adapt, Ayrton developed a sophisticated optic intended for colour mixing from 4 diodes with separate power. The part is shaped like the Eiffel Tower and Gaggione will be the firm that injects it. However, draconian tolerances imposed for positioning diodes are not an industrial solution and despite many attempts and corrections, the product is not coming together.
Yvan Peard: “There is always this phase when you invent things. It is certainly risky, but if you don’t go through it, you’ll never invent anything.
This Eiffel Tower has cost us a lot of money and time, but it also helped us to consequently develop the 45 mm, in partnership with Gaggione, adapted to colour mixing for a 4 chip model. It has helped us to find the route for the Arcaline elliptic, the Ice Color narrow and the Wildsun zoom following an Ayrton technical specification. This collimator was the precursor in 2011 and it must have been a success having seen the number of copies (with varying degrees of accuracy) made throughout the world.
With Gaggione we reason in terms of the whole product, always considering the LED’s geometric spread with regards to the beam.
We have undertaken a partnership driven by research; we meet up several times a year and we try to dream up tomorrow’s light, always focussing on shows and displays
We share information, some developments and that’s not always easy with a manufacturer who might sell your product to the competition. The idea of trust is therefore essential.”
The Ayrton 90 mm
Tracking down the Youkounkoun
The LEDnLIGHT 90 mm, result of a partnership between Ayrton and Gaggione (Photo Daniel Gilet)
Our plant visit will follow how an enormous collimator with a 90 mm diameter, 45 mm Height, was developed for Ayrton. A plump, 220g optic, made of perfect plastic, christened the Youkounkoun by Yvan Peard, who laughs as he explains it comes from a comedy road movie, Le Corniaud, revolving around the biggest diamond in the world, hidden in a Cadillac’s battery.
The Optics Technical Office
This 90 mm is the result of a partnership between Ayrton and Gaggione, starting with a feasibility discussion. Wide optic, narrow beam, colour mixing and zoom, the technical specification is a tricky equation to solve between optics and plastics manufacturing. After a year of studies and shared prototyping costs, this optic is ready to incorporate two new Ayrton light points, the Wildsun 702 and the Rollapix 402.
We met Jean-Pierre Lauret, optical engineer, in charge of the optic design office. Jean-Pierre has vast experience in optical design along with perfect understanding of constraints involved in producing moulds and injection meaning that he can anticipate them at the design stage.
Tight angle and colour mixture
Squaring the circle
SLU : How did this 90 mm come about?
Jean-Pierre Lauret: ”The 90 mm collimator is the technical pinnacle of a concept that has matured over time plus the development of preceding collimators. It started small and represents several years of nurturing.
The basic problem was that the RGBW multi-chip LED chips are juxtaposed. A collimation system is going to have a strong tendency to project the image of the emitting surfaces.
The LEDnLIGHT 90 mm (on the right) has grown slightly as a result of previous developments and experience. (Photo Daniel Gilet)
As a final result, what we get with classic optics is 4 juxtaposed spots. Therefore we have to get these 4 chips images perfectly superimposed in the design.
The other way of seeing things is to light up a single chip, therefore to work from an off-center source for the optic and recover a centred beam. Regardless of the chip’s position, the beam must always have the same centred projection.
Much of the work must be done by the reflective surface and the remainder, working in direct transmission, must do as little as possible.
Combining the two gives the resulting beam.
To get a good colour mixing, much of the work must be done by the reflective surface.
In colour, this raises a problem because everything that is going to work in direct transmission is going to project the image of the chip like a video projector and everything that is going to work as a reflection has a natural tendency to mix colours.
The secret is to maximise what is going to enter the parabola and minimise what is in the center.
In a second stage, we have controlled what is happening directly.
In a third stage, we developed a special output surface to control the light diffusion.
This is the concept applied successfully to the 45 mm.
Design for the LEDnLIGHT 90 mm optic zoom in the Difsys software.
In addition, we finished off the 90mm generation by adding a function that means the disk is perfectly clear. Once again, we obtained the mixture by maximising the size of the reflecting surface to focus perfectly and we controlled what is happening in the center with a set of lenses. We also created a diffusive structure on output allowing us to go on mixing in the small residual colour faults and obtain a high quality colour mixture, taking the specific shape of a rosette.
Experience taught us that centring and the chips position on the unit are critical elements.
If the chips are offset by just 1 or 2/10e mm this can upset the colour mixture. If the collimator is decentered over the LED, we do not get the chosen beam. The spot with the 4 lit diodes must be white.
When we decentre it in the wrong direction, we can get a pink spot with a blue crown around it, or a green spot with a purple halo.
The more we tighten the beam, the more sensitive it becomes.
The real problem lies in obtaining the narrow angle and the colour mixture.
With a narrow angle, we need to have low direct diffusion to not get a too wide beam. The diffusion is however needed to obtain a good color mixing. This 90 mm is a challenge and we correct the faults on it using other means.”
Pro trace to simulate the optical performance of the 90 mm with a chip. We can see on the intensity curve that the LEDnLIGHT 90 mm used by Ayrton with the chip used for measurement would have an angle under 8º at I/2.
The mechanical design office
The LEDnLIGHT 90 mm collimator created in 3D using Top Solid software. It is used to make the mould
Whilst Jean-Pierre Lauret designed the optic and its specific shapes, it was Stéphane Locatelli, Project Manager, who finalizes the 3D part using Top Solid® software by adding details such as the injection point, the demoulding angles, ejectors, etc.
This software can produce the CAD model that will be firstly used for the optical simulation and then for the mould design and machining, whether this is traditional with simple shapes, using electro-erosion, wire erosion and finally using diamond machining for complex shapes such as the rosette of our Youkounkoun. Not easy to copy!
This technical office is run by Joseph Busi who has been with Gaggione since the early days.
Diamond machining
One of the mold impression of the 90 mm LEDnLIGHT collimator.
The 90 mm, and particularly the impression for the diffusion lens and its associate the zoom lens, lead us to the air conditioned white room dominated by the famous digitally controlled diamond machine operated by David Gluchowski.
Dynamic control over 4 axes (x, y ,z, c), optical adjustments that are accurate to the nearest 34 pm to guarantee the 30 nm displacement tolerance over the 350 mm rails, linear motors, therefore with no gearings or bearings that might generate vibrations. Constant hydraulic pressure guarantees no friction… To keep the machine stable, it stands on marble, itself laid on pressurised platforms installed on a plate uncoupled from the rest of the slab because stability guarantees no shape faults ( (
In machining, the diamond tower is accurate to the nearest nanometer.
Alignment of the mold impression before machining operation.
The pin turns at 0 to 10,000 rpm, driven by a linear engine. Parts are held by vacuum. This machine’s description all sounds like fabulous mechanics but it comes at an equally amazing price. Nanometer accuracy definitively does not come cheap.
The machining tool is not to be outdone. Whether it is made of diamond or polycrystalline (although David prefers the more reliable natural material), it comes in different ranges and profiles, the smallest measuring 1 micrometer will set you back €2500!
There is a diamond at the end of the tool. It is either glued or welded and sharpened with under 2 microns shape fault.
The diamond, seen by the camera used for machine axis alignment with the centrer of the tool.
SLU : are you the only ones using this tool?
David Gluchowski: ”Here, yes and in France there are around fifteen of us, as training takes an extremely long time and to get good results, it is not enough to just put in a CAD file to drive it. There is an entire context of adjustments, analysis, anticipation, knowledge of mechanical effects generated that might modify the impression in the end.
We register the peak-to-valley that means the differences of shape between the mold impression and the injected part. If it is not fully compliant to our quality standard we have to analyse why, to modify the machine parameters and generate a new file to do it again until we get the final result.
SLU : What material is used to make the impression of a mould?
David Gluchowski: Generally we use steel as it is a robust material ensuring that the mould lasts a long time. Except that steel contains carbon, and so does diamond. If a diamond touches the carbon, it explodes.
For diamond machining, we therefore use copper alloys. If we want to make a mould that’s going to last a very long time, we make a pre-shape out of steel and send it to specialized company to get nickel grown over it by electrolysis. This is a very long stage. It takes several weeks to get a thin 500 µm layer that can be diamond machined.”.
Image of the surface condition for the LEDnLIGHT 90mm collimator measured using a CCI Lite (to the nearest Angström).
David then showed us his control machine collection. Form Talysurf FTS Series 2 from Ametek – Taylor Hobson is a (truly!) diamond tipped mechanical profilometer / roughness meter to control the roughness down to the nearest 5 nm. CCI Lite from Ametek –Taylor Hobson Precision, an interferential microscope, analyses and maps the surface conditions with 0.1 Angström resolution. And finally, the ZIP 300 Smartscope from OGP, opto-mechanical 3D measurement apparatus, uses a camera alongside a ruby detector to scan surfaces and measure shape faults.
Thanks to an OGP opto-mechanical measurement, technicians obtain a 3D view of injected parts to verify their tolerance.
Checking the shape fault on the LEDnLIGHT 90 mm using a profilometer / Ruby detector. (Photo Daniel Gilet)
Plastic injection
The only electric machine (300 tonnes) intended to inject the great thickness optics.
31 machines take care of the injection and correspond to two different types of technology: German piston hydraulics for force, Japanese electric machines with steppers to work on fine adjustments. They are ultra accurate in terms of movement, speed and position. Last but not least, we have an hydraulic equipment with an equivalent to 350 tonnes for its mould closing force, in order to injects large diameter optics. That is exactly the one we are interested in.
David Veryser: ”The material arrives in granules that first of all go through a desiccator to remove any humidity content from the material.
Then it is taken to the injection nozzle through an endless screw throughout which there are heating trays that melt the material, bringing it to the right temperature. The material should not be allowed to degrade due to the risk of creating black spots. It should be melted progressively without burning the material. This is experimental cookery, hard-won and valuable experience.
The mould is made up of two parts. When the mould is closed we can inject the material. Temperature probes can control the mould’s thermal adjustments.
After an injection series, the mould impressions undergo stripping in the expert hands of J.J Grisard who uses different abrasive pastes
After injecting the material, we can start the cooling. The larger the part is, the longer the cooling time as the core must be cooled whilst maintaining important pressure so that the material does not set immediately at the mould entrance.
It must be pushed to hold its shape in order to avoid the materials’ formidable physical shrinkage as it cools.
Then the mould opens, the ejector push the optic to break it away on one side and a robot comes to grasp it and put it delicately on a belt.”
Example of shrinkage on an optical part for two different cooling times.
A technical comment on roughness.
SLU : The injection point remains, like a little umbilical cord. How do you deal with it Yvan
”Yvan Peard: It’s actually pretty big…. But when we’re talking about a zoom, in this case, we can make use of this carrot as part of our positioning requirement. It gives us the perfect position for the zoom lens against the collimator.”
SLU : What will the fault tolerance be for this optic?
David Veryser : ”The max acceptable shape fault for this optic is < 100 microns; beyond that we start to see the beam break down.
If you take the beam for an individual chip, the ideal geometrical shape is a perfectly round disk; if you add a shape fault, the beam becomes potato-shaped, the spot is decentred a little and the beams are no longer superimposed and so the colour mix deteriorates.
PMMA granules, primary material for Ayrton collimators.
SLU : What is the material used to inject the collimators?
David Veryser: PMMA (poly methyl methacrylate) better known as Plexiglas® or Diakon® and polycarbonate are the two polymers most used in optics. We will choose one or the other depending on use, as they have different characteristics. PMMA is rigid and brittle but it is pretty scratch resistant. It has very good optical capacities, only 10% loss on a thickness of 62 mm. It withstands 90º which is only just acceptable as LEDs are moving more and more towards high temperatures.
Comparison between PMMA and polycarbonate. Green is good and red is critical.
David Veryser : On the other hand, polycarbonate is flexible and elastic. Impact resistant, it can withstand physical constraints but it is sensitive to scratching. It performs badly when exposed to UV rays, oxidises easily, yellows and finally becomes brittle.
But it withstands 135º temperatures and above all it performs well in fire as it is self-extinguishing whilst PMMA burns in the presence of flames and projects flaming droplets that propagate fire.”
SLU : And for Ayrton?
Yvan Peard : ”Yvan Peard: The collimators are made of PMMA and all our projectors have a polycarbonate output lens.”
Adjustment
Plastics manufacturing is a thankless technique. All the accuracy implemented in designing and machining the mould can be ruined by shape faults, without even mentioning black spots. This is because after injecting a large optic like the Ayrton 90 mm, the plastic has to be cooled. The outer skin solidifies firstly, whilst the core remains hot and liquid. Towards the end of the cooling cycle, the core shrinks and causes the outer layer to deform which spoils the collimator’s optical function requested.
David Veryser : ”When beginning production, the adjusters produce tuning files like recipe cards. The adjusters fit the mould and set the injection molding. Production begins and when the parts seem to be satisfactory visually, the controller checks several mechanical and optical points before approving production. Then the operators visually check over each part. Regularly, a controller verifies that the parts are still compliant, therefore confirming there has been no drift in the process.”.
The photometric laboratory
It is then the technicians’ turn to run the Product Audit. They take the parts to the lab for the next step in photometric control using two goniophotometers, a talysurf Imaging (acquisition with camera ProMectric®) and a LEDGON 100 from Instrument Systems.
Jean-Pierre Lauret, David Veryser and far right Régis Chaplain (photometrical technician)
Test of the 90 mm with a LED… it looks promising!
Photometric measurement of the 90 mm collimator mounted on a goniophotometer Radiant Imaging. We get the light distribution.
Beware of imitations
On the left, the Gaggione 45 mm reference LLC49N and then an European copy and two Asian copies.
The 45 mm collimator is often imitated which is both flattering and really disturbing. The technical office team have managed to get hold of 3 copies: one compliant European copy and two Asian copies with visual faults. They could not resist putting them through the goniometer, associated with the same LED, to compare them with the LLC49N original.
Collimators curves. The LEDnLIGHT is mallow, the European copie is red, and the chinese copies are turquoise and green
Curves shows that you will need from two to three times more LEDs with copies to equal one original Gaggione optical system coupled with the same LED. Copies generate additional cost because you need to use more LED that means more cooling issues and so on.
When Gaggione carries out a custom study, the customer can be sure that they will obtain a final result complying with their technical specification. All the company’s energy goes into this. We have been given the idea during this visit and these discussions that low cost production is a dirty word around here. On the contrary, the words we hear most often are tolerances and reliability. This reassures customers that when they fit their optic into their light point, it will send out a lot of clean, beautiful light. We will see in Frankfurt whether the Wildsun 702 and Rollapix 402 can win the 90 mm challenge.
[/private]