PCB | January 31, 2006
Report from the EIPC Winter Conference
The EIPC Winter Conference was held on January 26th & 27th, Budapest, Hungary, with nearly 100 delegates from all over Europe. Read Mr. John Ling's report.
Day 1 – January 26th Paul Waldner, EIPC Chairman, welcomed all the many delegates to Budapest, and said that whilst he could see many familiar faces, it was very good to see some new ones as well, many of them, understandably, from the host country. He expressed his particular thanks to the many companies who had kindly joined together in sponsoring this event, and who were taking part through their Table Top Exhibition being held adjacently. EIPC's excellent new Managing Director Frank Smulders chaired the first session, and Augusto Meozzi, President Europe, Isola Group gave the keynote speech, based on the development of business in Europe, with a look at the economic challenges and the technology trends. As if we needed it, Augusto reminded us that between 2000-2005 the USA PCB industry went down by 62%, Europe by 37%, Japan only 15% down, but of course Asia and the ROW saw their industry rise from $13 billion dollars to $22 billion in the same period. Asia now accounts for 46% of the utilisation of semiconductors. In Europe the number of PCB companies has dropped from 561, to 330. Getting down to the industry closest to his heart, he said that whilst in 2001 there were 13 laminate suppliers, now there are just 7. Back in 1990 there were 23. Ah well, nothing stays the same. In 2004 China accounted for 20.8% of PCB production. Yes, but what about 2005? Augusto emphasised that it is the industrial electronics sector that is growing, the OEMs sourcing automotive products now is down to 30% in Europe, and in the PC world it is only 25%. How do we keep our industry attractive for European OEMs? Well, you can improve your pace of innovation; you can minimise internal cost structures. It is always worth noting that China has its own not inconsiderable downsides, which include high labour turnover, a very high reject rate in production, and the cost of employment of expatriates in China is way ahead of, say, Eastern Europe. Bear in mind the cost of shipping as well. Let alone the political situation – democracy is not to be deferred in China, however they might tinker with Google. So what has Isola done to stay competitive? They have undertaken massive restructuring, closed down 4 sites, implemented a lean production concept, acquired their own glass weaving company, and made productivity improvements of 20%. Isola had to meet the demands of RoHS implementation, lead-free soldering, leading to increased thermal reliability, and CAF resistance. Augusto presented details of how Isola had developed products with greatly improved thermal reliability, giving good results with lead-free soldering. Worth noting that HF materials will be needed in adaptive automotive cruise controls, etc. There had been no growth globally in PCB production between 2000 and 2005, and European PCB production had declined 35% in the same period. However the European PCB market is still worth nearly $4 billion, and much of the migration to Asia seems to have leveled out. Europe can optimise its structure, adjust capacities to demand, deliver quality and service, and fight back hard by being innovative. Tarja Rapala-Virtanen from Aspocomp in Finland had an ambitious paper that merited a couple of hours, but she only had 20 minutes. As her philosophy is 'No risk = no fun!' one understood, but still felt cheated. She looked at some new challenges – embedded passive and active components, design indicating smaller, lighter, with miniaturisation, higher speed and higher reliability, and more function. The results of this trend are around us, in things we use on an everyday basis. Tarja wanted to tell us about some of the PCB solutions available: -Embedded components, active and passive. Devices are getting smaller so routing and I/O densities areincreasing. By embedding components we can add value to the chain by creating more functionality. Tarja explained in some detail how this is accomplished, and her paper was accompanied by a number of illustrations that explained how. The whole PCB structure is used for component assembly and this leads to good reliability, makes the total cost competitive - but critical to yield, Rigid-Flex – drivers are space limitation, assembly efficiency, improvement in reliability, the demand for dynamic flex and cost efficiency. Today nearly all mobile phones, digital cameras, LCD's and plasma flat panel display screen are using HDI PCB, but the need is to combine the stacked via and flex technology by a bendable HDI build-up layer. This can all be done using existing standard PCB production equipment. Optoelectronics – embedded optics; the challenges of optics on PCBs is where alignment accuracy is needed. The Ultimate Wish List? Printed electronics as an enabler, that can be used for RFID tags, identification tags, smart packaging. But printed electronics needs a joint collaboration consortium, working with universities, suppliers of materials, substrates and ink manufacturers, Aspocomp is certainly an innovative company, and one that is fully aware of the demands of their end-of-line customers. Professor Martin Goosey of Rohm & Haas Electronic Materials came to tell us about TAZDIS, a collaborative project. PCB manufacturing isn't a very sustainable process, is it? It needs lots of water, lots of chemicals, and produces lots of effluent. What we use today must not preclude its use in the future. The main aim of the TAZDIS project is to make PCB manufacture more sustainable. Here new Ion Exchange systems, high efficiency electroplating and advanced oxidation systems have been extensively studied and integrated into an industrial scale demonstrator. Funded by the UK Department of Trade & Industry, it involved Graphic PLC, Rohm & Haas, Printed Wiring Technologies Ltd., Giga Solutions, and the UK electronics industry association Intellect. Organic Capture Technology is where new ion exchange resins have been used to capture and concentrate organic contaminants from process rinse waters. Saturated resins are then regenerated for continued use, and the regenerated liquor is transferred to a subsequent electrochemical organic destruction unit. Here work has focused on the use of a UV and ozone advanced oxidation process, where the recirculation of regenerant from organic captured media is passed through the advanced oxidation system, and once the organics have been destroyed and the metals recovered, the treated rinse water can be reused. Ion exchange has been investigated for the capture and concentration of copper, nickel and gold from rinse waters. In the field of metal recovery, they have evaluated two types of electrochemical process – the Chemelec Cell, and the Renocell. The former has an enhanced fluidised bed recovery system; the latter has three dimensional cathode metal recovery system. A demonstrator system of TAZDIS is now running at Printed Wiring Technologies Limited. Well worth a visit. Birgit Henss is a multi-lingual project director at Cimnet Systems. Having good software in place is crucial for the success of a company. You need a good information system, which should include quote management; rules based engineering; ERP solution integrating all relevant business processes; and web based collaborative applications. Cimnet have been working in partnership with Coretec, of Canada, who have used the Cimnet System so they can write about it. In many PCB companies there was a lack of structured approach to communication, no integration; systems should be integrated, bringing commercial, technical and financial aspects of PCB manufacture together rather than side by side. Birgit explained the Cimnet RFQ System and their Rules Based engineering system that is somewhere between CAM and ERP. Here engineering, shop floor and business data is integrated to allow the engineering team to produce PCBs consistently, accurately and optimally. Rules based engineering systems capture and transform the knowledge and expertise of an engineering department into a strategic corporate asset. Cimnet have an industry specific-ERP system called Navipoint. Coretec are an end-user of Cimnet Systems, and Birgit was able to illustrate in her paper how this PCB manufacturer had made very significant cost savings, greater efficiencies, and enormously improved yields by the effective use of this management tool. This is a very specialist subject, perhaps not always fully understood by those who have not taken a degree in information technology and software design. However, as a communication system for PCB manufacture you can not doubt its contribution. Gordon Clark is Global Support Director for Koki Compay Ltd. His paper was entitled Lead Free Interconnection, and here he bravely entered a huge subject in a small amount of time. In the UK, said Gordon, the major problems with lead-free are voiding in solder joints. What are the likely causes – was it flux content, at 11% or 12%? No. Was it microvias? No. They tried filling the microvia with 2-pass printing, and the voids were still there. However, if you used lead-free solder with lead-free components you get better results. Dual printing does reduce it a little but it's still there. Anything to do with the reflow process? Leadfree was better, here. Oxides, possibly? Gordon compared linear v saddle profiling with tin-lead, with leadfree there is still some voiding, so this told us little. A combination of tin-lead ball and lead free solder paste will cause massive voiding. So the exact cause of voiding is still in question. But one conclusion is – don't mix lead and lead-free, it doesn't work. He looked at the pillow phenomenon, which can be caused by misplacement; accuracy is much more important than before, and lift more pronounced with lead-free. Gordon packed a lot of useful information into the time available, with an invaluable summary of the pros and cons of various surface finishes. Dieter Walz from Atotech Deutschland presented a paper looking at the interactions of immersion tin with lead-free soldering. He kicked off by asking what is different in lead-free soldering? It's the temperature - the liquidus is different. There are higher temperatures and longer times, there is stronger oxidation of the solder paste and surface finish, and there is increased thickness of the intermetallic compound. At Atotech they have done a lot of work, using three different profiles through a reflow oven. They compared solder flux, solder alloys, (eutectic tin-lead v lead-free) and tested fluxes including an IPCflux, Stannol 500-6B, and Litton-Kester 950E3. And they looked at their interaction with various finishes in wetting times. A great number of tests were made, and their conclusions were that the wetting balance strongly depended on the flux. But you can have comparable wetting times using a nitrogen atmosphere. They got an equivalent result with lead-free in a nitrogen area. They concluded that immersion tin is a suitable final finish when it comes to leadfree soldering, the formation of tin whiskers can be suppressed by additives and should no longer be an issue. It should also be borne in mind that a higher tin thickness is needed due to the increased formation of intermetallic compound by the new lead-free reflow profiles. Gerjan Diepstraten looks after the lead-free soldering sector at Vitronics Soltec BV, in the Netherlands, and is busy selling the process as well as the systems. He discussed the prevention of lead-free solder defects in different soldering processes. He graphically illustrated problems with lead-free solder joints, such as blowholes on laminates, cracks in copper barrels, a multiplicity of defects. Good design is important, and a terrifying number of things can go wrong. Gerjan looked at most of them. Solder balling; you get solder balling in lead-free due to higher temperatures. Getting sufficient amount of solder into the holes; here 100% filling is preferred. Another company who have gone to exhaustive lengths to look at all the potential problems with lead-free and all the potential solutions. He particularly mentioned that for lead-free, cast-iron solder pots are good, as are titanium. Selection of a good flux, that is able to withstand high temperatures and which can prevent bridging and improve through-hole penetration, is obviously vital; Gerjan also repeated the need for component placement accuracy. Michael Läntzsch spoke eloquently about the use of SN100C in lead-free HASL and for board assembly for Balver Zinn. Balver Zinn is based near Dortmund, and make all their own alloys, both for the galvanic industry and for the electronics industry. Balver Zinn feel that the physical and metallurgical background of lead-free soldering is exactly the same as is the last 80 years with tin-lead, namely a reaction with the copper surface – with a clean copper surface good wetting is always possible. Balver Zinn have produced two alloys, SN100C, excellent for wave and selective soldering, SN100CL (SnCu0.7Ni) has been successfully applied in HASL machines, notably on a Cemco Alchemy system. Balver Zinn it would seem have very stable product that has won praise from companies no less than Boeing as being the best lead-free solder in wave soldering performance. Dirk Hillebrand is the Marketing Director of Heidelberg Instruments, and came to talk about LSA (Large Scale Aligner) – PCB exposure with distortion correction. The exposure process is an off-contact exposure between the mask and the PCB. He gave a detailed LSA technology, where a distortion corrected projection of the artwork onto the PCB is made. The system incorporates multi-fiducial alignment, which means that 5 microns lines and spaces may be imaged successfully on a positive photo resist. It offers focal depth, so that a homogeneous line width is created even on uneven circuit boards, e.g. flex/rigid, flex and outer layer/solder mask. LSA is the best of both worlds, its patented synergy of conventional exposure machines with great projection accuracy. Would be good to see a machine in action somewhere, sometime. Bill Wilson of Du Pont Electronic Technologies UK wanted us to get wet!. Wet film lamination techniques go back to 1984, used for laminating dry film photoresists using the flood approach. However, the original system was plagued with leaks and resist wrinkles, so Du Pont redesigned it, and made considerable improvements. They reintroduced WetLam in 2000. Now it's going well, offers considerably improved resist adhesion, and customers are using this system for conventional inner-layers. It can be used for flexible layers, sub-via layers, and photochemical milling. The limits have pretty much now been reached. Bill illustrated the new wet/lam applications, which include flexible/rigid panels, HDI outer layers, It can reduce manufacturing costs, and increase yields, and improves resist conformation to surface topography and imperfections. So get wet, he said. But he didn't bring a coat with him to Budapest, and cheerfully froze in the arctic night air of the city later that evening. Uwe Altman from Orbotech asked - would LDI meet the needs of HDI and packaging manufacturers? Given that there are now at least two companies manufacturing solder mask that is compatible with LDI (Coates, Taiyo), and given that designs are creating tighter registration tolerance, (less than 25m annular ring), together with more production demands, LDI has found its place. Features down to 25um can be exposed on LDI, and fine features below 20um have been successfully exposed in lab. tests. LDI machines capable of 80 panels per hour, with a 4W-laser technology, and a new system with 8W laser has been announced. Orbotech have their new Paragon with automatic panel handling for high utilisation of laser direct imaging, which is already popular in Asia, where resolution of less than 25 micron annular ring is being achieved. Features for HDI of 25-30um, and for packaging of 15um are available to-day, and throughput is seeing still more improvements, up to160 prints per hour. Erik Biehl came from Multek in Germany to talk about how to select base material for lead-free production. (What happened to Sylvia Ehrler, one wonders?) There are some unknowns – via reliability/crackingdelamination resistance – decomposition resistance – electrical degradation. Phase 1 of their evaluation programme was selection of materials, then a repeated reflow evaluation, looking at reliability testing and selection of preferred materials. Capability testing is Phase 5, which is where they are now having started at the beginning of January 2005. Test vehicle was a 25-layer board, 3.175mm thick. In a detailed paper he concluded that out of 16 samples, 4 were just fine, 4 needed some improvement, and 4 needed quite a bit of improvement. 4 just were not up to the mark. High Tg material not automatically the answer – CTE values have a significant influence on lead-free compatibility and performance, and the process window for key processes are narrower. Dr. Bernd Hoevel, of Dow Deutschland GmbH, gave a paper on the influence of the binder system on current and future demands for base materials. They are concerned with the effect of higher soldering temperatures, understandably enough. Dr. Hoevel gave a detailed technical paper that demonstrated a thorough examination of the problem. This included tests to show thermal degradation of the binder system, delamination tests, the influence of CTE mismatch on through hole reliability, filled materials and their impact on drilling performance (LF150 filler superior to non-modified brominated novolac system). Finally, the influence of moisture pick-up under lead-free soldering. Their conclusions were that whilst FR-4 laminates will continue to be the most effective base materials for relatively simple applications, more complex applications involving multiple reflow will require more thermally resistant laminates, and the use of the correct filler (such as LF150 or LF170) may be essential. Far and away the prettiest speaker of the afternoon was Veronique Streukers, who told us all about EBFRIP/Albermarle. EBFRIP is the European Brominated Flame Retardent Industry Panel. She also spoke about Tetrabromobisphenol A (TBBPA for short) which is a reactive flame-retardent used in laminates. 170,000 metric tonnes produced per annum, all in the USA, and over 50% is used in laminates. Veronique informed the conference that the EU Risk Assessment Process, started in 1993, has declared initially that TBBPA is not a persistent bioaccumulative toxic chemical, although finalisation is expected by the end of 2006. TBBPA's use in PCBs is fully compatible with all existing ecolabels regardless of the classification. Human Health reports indicated no risk; no human health hazard of concern was notified. Veronique would like an EIPC member to sit in on their monthly conference call, as they don't have anyone who actually uses copper clad laminate to make PCBs. By the way, if you want your circuit boards scrapped, send them to Umicore in Flanders; they will smelt them down to the precious metal content quite readily. Day 2 – January 27th Thomas Kunz owns Schmoll Maschinen, Germany, therefore he spoke from the heart. They are surviving it all rather well, are satisfied with their development, and are producing 500 machines a year. Whilst their customers are all over the world, they will always manufacture in Germany, as that's where they have their core competence. Some facts & figures: * World market share: Hitachi have 38%, Schmoll/Posalux 27%. * World demand for drilling machines in 2005: 2500 machines. Schmoll Production: * Flex they make 100 machines a year, * Standard 0.35mm machine: 220 per annum, 95% for China. * HDI machines – 220 per annum, 60% to China, 10% to Europe, 10% ROW. * Substrate machines: 60 machines a year, 80% for Taiwan, Korea, Japan, 10% China and 10% ROW. * World demand for routers is about 600 machines. Europe is recovering well and they are seeing increased demand, not only 0.075mm now being drilled in the market, but 0.1mm is now pretty standard. Thomas has seen a panel with 600,000 holes, so you begin to see the problem - as hole counts increase, so speed becomes important. To ensure that their machines perform well, they carry out a number of test including a static drill test, a dynamic test, and a heat dynamic drill test. Cooling concept LM6 gives much better drill accuracy, as machines build up a lot of heat. Linear Z axis is a must for accuracy, so too is a short drill stroke, down to 1.8mm from 3.3mm. Spindle technology today is moving quickly, now there are 300,000rpm spindles for 0.1mm hole. One day one million rpm? Who knows! Their PS300 spindle, 3.175mm shank, is working well against the more (50%) expensive and less versatile Hitachi 2mm-shank spindle. Schmoll use linear drive in the x, y & z-axis, have noted that higher speeds increase tool life, and that 300,000 rpm drilling speeds also allows you to increase the stack height. By 2009, who knows, maybe 0.04mm-tool diameter. Uwe Lenz runs the family firm of Ernst Lenz Maschinebau, also in Germany. They made their first drilling machines back in 1939, and the first one for the PCB industry in 1968. They have been at it successfully ever since. 50% of their machines go to Asia, and the rest to the ROW. He talked about demands in Asia & Europe. Asia needs machines for long runs, manual loading, single purpose machines, micro-drilling down to 75 microns, with separate machines for drilling and routing. In Europe it is rather different, usually single or double spindle machines, with automatic loading and unloading, on short runs needing fast set-up time, and drilling and routing on the same machine is often demanded. Like Schmoll, they too prefer the linear motor for the z-axis drive, as well as the y-axis and x-axis. Lenz have a micro-drilling pressure foot that ensures the most accurate drilling of microfine holes, drilling down to 0.05mm drill bit diameter. In Europe they need machines with a spindle switch-over, they need contact depth control (±12 microns) and broken drill bit detection < 0.1mm bits detectable). A compact machine, with either single or dial heads, usually suffices. Tony Bannan is the Technical Director at Westwind Air Bearings UK, who gave us an introduction to the process capacity factors in micro-hole drilling of BGA board materials. When he spoke about micro, he gave us a diametric view of a human hair, and this looked large compared to some of the drill holes required. He looked at the design criteria that permitted such high accuracies to be achieved in drilling, the problems with collet stability, with dynamic runout and vibration traces, and all considerations in the search for perfect stability. Spindle related sources of drilling error include shaft vibration due to imbalance; shaft orbit due to imbalance and bearing whirl; collet runout due to manufacturing tolerances; motor/drive induced shaft vibration etc. There are also machine-related errors, tooling related errors, including drill tip geometry, surface speed etc. High cost, high developments, but at Westwind high expertise. Jürgen Scrypczinski is the famous and distinguished face of HAM, Germany. Jürgen gave a wide-ranging paper, in which he covered micro drills -still in the range of 0.15mm; the drill material - it is the cobalt in the tungsten that makes it solid carbide, and fast spindle speeds need this hardness; choosing the right grade of carbide is important. HAM still working on the 4-facet point drill head, and have introduced a variable web thickness which gives longer drill life, the back taper element reduces friction, and therefore temperature. Jürgen highlighted the factors effecting positional accuracy, which are several, and he also looked at the subjects of depth routing, routing thick copper, high-speed routing, routing half-plated holes, and the use of chamfering tools. He concluded with a glimpse at backing materials, and their importance how tool life can be prolonged. You can have the best drilling machines, but they are impotent without good tools. Good tools come from HAM, quite obviously. A paper from Dr. Bernd Hartmann talked about the development of a recycling process for an acid cupric/ferric chloride etchant, the work being done at his company Ilfa Feinstleitertechnik in Germany. Such new etching media are required not only for very fine line resolution, but also for etching very thick copper layers for power electronic applications, but should also be part of an environmentally friendly process where materials can be regenerated or recycled, or even reused. Using a Beizomat measuring instrument, they had worked with a great variety of etchants, such as ammoniacal, cupric chloride, ferric chloride, as well as persulfate and hydrogen peroxide/sulphuric acid solutions. They found that mixtures of the cupric and ferric chlorides showed a faster etch rate than the pure systems, and proceeded to develop an innovative system of etchant recycling using two closed loop systems, and one from which the pure copper can be collected as scrap and resold. Here was a perfect example of a PCB manufacturer and an engineering company working together to a good result. Attach Deutschland had sent along Bernd Roles to inform the delegates about the work done at Attach on blind MicroVAX filling in conveyorised equipment. This is a field in which Attach are strong, and work on their horizontal plating lines (Uniplate Inpulse) was carried out as well as on vertical systems. They used insoluble anodes together with a copper replenishment system based on an iron redox process, and they found that this combination gave a much improved electrolyte lifetime in comparison to existing filling systems. The use of insoluble anodes also gives productivity and surface distribution advantages that are in fact critical for filled vias. Atotech are now aiming to enable the filling of through holes with electrolytic copper plating, to replace conventional resin plugging of through holes. Jean Rasmussen dropped in from Enthone to join in the discussion about microvia filling, which he saw as a challenge for process quality. Enthone have produced a new product called Cuprostar CVF1, which is capable of both filling and through-hole metallisation, and this was clearly illustrated in his paper, with pictures of 145mm x 75mm, 95mm x 75mm, and 130mm x 145mm microvias all perfectly plugged. Where both mechanical and laser drilling can leave rough walls, these are not a deterrent to 100% filling. He went on to talk about Envision DMS-E, a sound alternative to electroless copper metallisation. Here a conductive polymer is used as surface treatment, a few 100 Angstrom thickness suffices, and this eliminates the risk of hole wall pull-away caused by surface contamination. EIPC have, yet again, organised a superb technical conference, massively informative, with papers presented by people with a level of expertise that one can only marvel at. That the PCB industry has this professionalism at its disposal in Europe should be a matter of some great satisfaction, and should be used to advantage. Delegates came from France, Germany, Italy, Hungary, Sweden, Denmark, Finland, Luxembourg, Switzerland, Belgium, Holland, Ireland, Israel, Scotland, England and the USA, and you do not need to wonder why. It may have been -16°C outside, but inside the Béke Hotel in Budapest it was warmly welcoming and maybe the glow of satisfaction of the delegates had something to do with it. EIPC hosted a dinner for the delegates and their wives at La Citadella, which sits atop the hills in Buda and overlooks the snow-flecked city. Interesting Hungarian wines accompanied a delicious four-course dinner, and cheerful but impudent musicians would play their violins immediately adjacent to one's left ear until a donation of 200 HUF brought some relief. All very jolly. John Ling for EIPC