Intellect Components and Manufacturing Technical Workshop

Citation

Ling, J. (2005), "Intellect Components and Manufacturing Technical Workshop", Circuit World, Vol. 31 No. 2. https://doi.org/10.1108/cw.2005.21731bac.002

Intellect Components and Manufacturing Technical Workshop

Intellect Components and Manufacturing Technical Workshop

Keywords: Electronics industry, Electronic equipment and components Yield improvements in PCB manufacture

The workshop held in the offices of Rohm & Haas Electronics Materials Europe Ltd, in Coventry on Wednesday 15 September.

Peter Starkey of STS chaired the meeting in his customary affable style and in welcoming the many delegates acknowledged the generosity of Rohm & Haas (formerly known as Shipley) in providing the facilities and an excellent buffet lunch.

The first speaker was David Wayness, European Marketing Manager at Rohm & Haas who covered “An overview of digital imaging for advanced interconnect application”.

Digital imaging was the creation of an image composed of pixels of discrete values that can be read and manipulated by a computer. There are three main types – laser direct imaging, laser direct structuring, and ink-jet printing. Laser direct imaging, which will expose a direct image on to photoresist, for whether it be etching resist, plating resist and or solder mask. LDS – is effectively a Siemens system that is used to selectively ablate immersion tin from a copper surface, the remaining tin acting as an etch resist. Inkjet – is the means for creating a direct image for etching, solder mask and legend ink. He listed the manufacturers and moved on to discuss the advantages of digital imaging.

• •

  Elimination of photo-masks, faster job turn, reduction of work in progress, better resolution, and higher yields;

• •

  Image compensation – the system allows some image compensation panel by panel. There is a potential to change the design rules as well.

• •

  For the inner layer process, a comparison shows that ink jet imaging only needs five steps, which compares well to the nine steps needed for photoresist.

David took a look at ink jet printing systems. There are many considerations, such as:

1. 1.

   system level parameters, including substrate size, throughput; reliability;

2. 2.

   print head level parameters – drop formation, the nozzle itself, material compatibility, the operating temperature range, and operating frequency.

3. 3.

   the pattern formation accuracy, variables effecting drop accuracy, and how nozzles need to be tuned.

4. 4.

   Material level parameters – surface tension, molecular weight fractions, stability, drying requirements, solvent or water borne inks vs 100 per cent solids;

5. 5.

   what do you want it to do? And, more importantly, will it?

In discussing a roadmap of print head performance, David went on to say that Rohm & Haas have a committed global programme on inkjet, and are studying the potential for yield improvement. Inkjet, he said, can lead to savings, through the elimination of tooling, and a reduction in process steps.

Frank Coultard asked if any sectors of the industry were picking up on this technology? David said that yes, they were, but like all “early” companies they were experiencing the hard learning curve of practical application that follows such crusading moves.

Neal Wood, Product Manager at Atotech UK LTD, wanted to go back to basics on yield improvement. He was greatly concerned that against a background of considerable change, the industry had lost a lot of valuable experience, which coincided, with a period of great technical innovation. This made for a difficult scenario, with lower yields leading to lower profits in a dynamic market, where competition is fierce, and prices are falling. Here one has to maximise performance, and reduce costs which include materials, manpower, utilities, and time in re- work, costs of poor customer service; with the reduction in the number of supply house now reaching critical proportion, partnerships are vitally important. That way it is easier to introduce innovative processes, and improve yields, said Neal.

He gave some examples of how collaboration between supply houses and their customers, or partners, could lead to the collection of correct analysis of yield data, a better understanding of failure modes, better use of raw data. Currently the quality of data available is poor, so is interpretation, and there is a need to change attitudes.

If machines are down, there is a big production bottleneck. If maintenance suffers, how does this effect the condition of pre-clean, of developers, of resist strippers? If there is bad handling, where panels are handled all too frequently, and there is a lack of thought or care what happens? Remember that resist is soft, and shows finger marking, and, with finer track and gap and spacing, is easier to damage. He showed some examples of defects caused by various means.

In this day and age you have to maximise profitability to survive, and if you can increase prices, well, that would be good, but you do have to improve yields, and for this there are some fundamental points to consider which include: accurate data collection, correct analysis, correct categorisation of scrap, and the need to identify the key areas for action.

Steve Jones, Chairman of Circatex joined in here to say that he extended an open invitation to visit Circatex, who should be seen as a professional colleague in the industry. In the UK, he said, we have all got a place together and we need to share knowledge to ensure mutual survivability. There is a need for UK industry to get together on this topic, and to promote collectively the interaction between manufacturers and suppliers.

Steve Kerr is the business improvement manager at Invotec, and his business has improved quite a lot. How? well, said Steve, it had a lot to do with robust design.

He began with a look at the PCB industry today, the product trends. True, high volume had been lost to the Far East, also technology was moving eastwards, eastern rates of pay and hours worked would be illegal in the West, but OEMs and CEMs like to procure locally, and the defence industry likes a local supplier too, for strategic reasons.

We need to focus on being lean, and having robust designs for technically advanced products and services. Components and their footprints will get smaller, electronics will get faster and holes will get smaller. Smaller, smarter, faster, all led to why micro-vias are needed. Fine pitch components put strains on the board's real-estate, and high levels of interconnect call for reduced pad sizes and reduced hole sizes, and are cost drivers with the possibility of lower layer count. Steve covered the RCC vs Prepreg saga, the pros and cons, and looked at sequential microvia construction. He then went on to cover design rule drivers which incorporate 11 major headings, which include an increase in functionality per square centimeter. HDI's board need laser drilled microvias, so material stability becomes important too.

Complex designs tend to suffer poorer yields so how we handle raw materials does not help much. Good design for manufacture (DFM) reduces opportunity for errors (OFEs) and reduces overall design costs. At Tamworth they have an Engineering Guidelines Book for designers, which explains to them what they (the designer) can do to work effectively with the manufacturers. He gave some values for DFM, and explained how at Tamworth they communicate through a training course run in-house to train designers, and their Best Practice Design Rules are there to assist designers for the latest electronic equipment. Steve was enthusiastic about enhancing customer trust (ECT) which sat nicely alongside the DFMs and OFEs with HDIs for OEMs!

Finally, Dr Steve Jones came down from Circatex to tell all about yield improvement on large panels. It was very important to work closely with customers, who are the drivers towards large panels. Yields are in fact not a major concern for him, as we could never compete with the Far East even with a 100 per cent yield. So what are the product variations that lead to scrap? Is it the process, is it the materials? Dr Jones looked at the way forward.

First of all you need a yield prediction model. If you do not measure things you do not; know about things, and as Dr Jones is a mathematician by instinct, he showed the model he had created and explained how it worked. There is balance between yield on the one hand and process utilisation on the other, the yield bit is what he sells, and the process utilisation is what it is costing him. If you can eliminate the variables in materials, processes, then costs come down. Simple, really.

At Circatex they investigated multilayer presses, prepreg style, warp direction, compared Thermount to different glass styles, and materials such as filled RCC vs unfilled RCC. They made a predictive stretch model which allowed them to pre-plan product engineering. Their predictive model showed that of typical 100 µ gap/track jobs, 90 per cent are within the design rule. A highly detailed presentation indicated that study of yield improvement has been close to Dr Jones' heart for some considerable period of time, and it showed. It showed in the software that has been designed which allows for accurate prediction of yield for any one job. As Dr Jones said, it is all to do with measuring things, otherwise you do not know about them.

Summary

Here was Intellect putting on a yet another very good technical seminar during which some very professional people were showing how to do it, so that all could freely learn. But all were not there. But maybe they know anyway, so that is all right then. The future of PCB manufacture in the UK is assured. Phew.

John LingAssociate Editor