Towards a RoHSy future, the final furlong
07 June 2006
We have been following the journey of JJS Electronics as it has made the transition to Pb-free manufacturing. With the deadline only weeks away, JJS has bedded in its new process. Naim Kapadia summarises the Pb-free project
RoHS is almost upon us and JJS has now successfully concluded its conversion project. It contained the following main elements: Management, Components, PCBs, and Soldering.
The initial aim of the project was defined as: Establish and implement a robust leadfree process, which had to be reliable and on time.
This was to be done by the most direct and simplest route. Time and effort looking at and comparing different suppliers was not in our short-term strategy. On this basis the selection of solders and fluxes and of a supplier was simplified: the best products were those from the supplier with the best support and knowledge. And if there was no difference then the incumbent kept the business. In meetings with our principal solder suppliers it was apparent they had good breadth of knowledge of how alloys and fluxes behaved and could advise us usefully without confusing us without too much theory.
We are confident the trials carried out and results we have obtained have been successful and will be replicated in future builds.
Users have very little control over component suppliers. They work to their own timetables and by their own methods. The knack is to assiduously check and recheck all detail and make sure it is documented. This process is complicated by the rather complete lack of uniformity in each supplier's approach. In the end JJS adopted a minimal stocks policy and built lookup tables etc into the components database. Thus prefixed and suffixed part numbers, or completely different part numbers are easily accommodated. On components with no change in part numbers the situation was a little different, as there is no real way of telling if old or new stock is being used. Again a pragmatic approach was adopted. Minimal stocks means only a small number of components is likely to be doubtful. It can therefore be cheaper to throw these out than to analyse. The loss on higher value components is minimised.
As a contract equipment manufacturer JJS does not have complete freedom on PCB finish. Again a pragmatic approach was adopted. As long as the customer could be persuaded to a lead free finish, the company was "Swiss" on which type. As it happens most of the PCBs by type and by volume are ENIG (electroless nickel immersion gold). ENIG, like all lead free PCB finishes, is compatible with leaded solders. This meant that there is no specific window to be observed on change over. However, this is one area that could be early on the review list once Pb-free processing is established. ENIG has been noted to give low joint strength with SAC alloys and this could be significant for some future customer applications.
The choice of solder, flux and supplier was one of the most critical decisions made. Solder is an integral part of the finished assembly so there were genuine concerns with long term reliability; especially as the documentation and history of lead free alloys is slight compared to over 50 years knowledge and experience amassed with tin/lead. A conservative approach was therefore adopted. Again later it might be that this area is revisited, but the savings available on solder pricing is not so great to make it a high priority.
This is the major process in JJS and extensive discussions took place on this. Our requirements were that:
- the material be as close to a drop in as possible with respect to process
- the successful material would have a consistent performance from as low a reflow temperature as possible and over a wide range of time/temperature combinations. This is desirable to minimise MSL problems on the one hand and allows the paste to accommodate any possible profile requirements imposed by different board and component combinations on the other.
- Work with existing tin/lead stencil library wherever possible to minimise the cost of change. This is a considerable and often overlooked factor.
The material chosen was Indium5.1 with Indalloy 241 solder (SAC387). This is a second generation no-clean paste and a linear descendant from the original SMQ230 introduced over 5 years ago. The SAC387 alloy was chosen as it is close to the theoretical eutectic and our supplier, Indium had the most experience with it. Again this could be revisited following the IPC work on SACs showing no statistical difference in reliability between the major SAC types, and some slight processing advantages for SAC305.
There were two difficult and related decisions here, both intricately related.
In solder paste alloy choice is relatively simple, which SAC? It is also simple to change your mind - just open a different jar.
Wave soldering, ironically almost the oldest and simplest PCB soldering technology, brings some quite complex decision making to alloy choice. It is the cost implications that bring the difficulty. At first the alloy choice looks complex with many different alloy types available from a host of suppliers. The complexity is increased by the fact that high tin solders (virtually all lead free types are known to fairly readily dissolve the materials of construction of most wave soldering machines. So in addition to choosing the correct alloy from a customer point of view our alloy choice could mean we also had to factor in a significant capex for a new soldering machine.
Based on input from our major supplier we considered this area in a slightly different way:
A closer inspection reduces wave solder alloys to two broad families, SAC or variants of tin/copper. Within SAC there are the various types: 387, 305 etc.
In the Sn/Cu there are the plain Sn/Cu types and the doped types. The "dopes" are actually small alloy additions, which modify the solder in some way. Either to make it more like a SAC in in-service performance (the so-called low SACs) or to reduce machine compatibility problems (nickel loaded Sn/Cu).
In the end it was decided that a new wave soldering machine would ultimately be required to take advantage of the latest spray application technology and VOC-free fluxes, this would naturally be lead free compatible machine. This being the case a standard SAC could then be used. So SAC 387 was selected. Again this follows the "conservative choice" dictum ensuring no compatibility problems with the SAC of the solder paste. Flux choice was relatively simple. Our existing supplier had a suitable material so no major evaluations were necessary. Our flux is therefore Hi-Grade 3592 - this had already been checked for compatibility by Indium with its solder paste formulations, so an easy choice.
Both these areas can be revisited if required: the flux simply enough at anytime, its just a matter of changing a bottle. The solder when contamination or alloy imbalance forces pot change. This is likely to be much sooner and more often than for tin/lead.
There are as many as 15 different flux cored solder wires available for hand soldering. This is an astonishing number for such a simple technology. Equally astonishingly there is a wide variety of performance levels amongst the different types. We chose Indium's Core 230 in Indalloy 241 for its compatibility with the other flux types and to maintain consistency in alloy type across our material range. Core 230 has adequate wetting, does not smoke excessively and has slight no-clean residues. Possibly there are other materials available but the Indium material works well and the amount used does not warrant an extensive programme
Managing a dual (leaded/unleaded) site
As much as technical understanding is important, it is equally important that the processes selected to carry out lead-free are managed and controlled in a proper manner. As many CEMS and OEMS have already experienced they have to run two processes to accommodate the lead free and the older leaded products.
When applying a dual process system it is very important to ensure a compliant and controlled system is considered for now and the future without cross-contaminating the process. JJS approach is to use the simple scientific model known as the "The Conscious competence model" as illustrated below which will give a broad understanding of the process.
As lead-free is implemented in the organisation the operators will reach a point where the process will go into the "unconscious competence mode" the skill becomes so practised that it enters the unconscious parts of the brain - hence it becomes 'second nature'.
The awareness of process control will become an important factor in the future to ensure product has been built to RoHS compliant process. Using the technique certain gates have been put in place to clearly distinguish the two processes, and safety gates have also been implemented when in "unconscious competence mode".
JJS has completed and established a robust process and can offer totally RoHS compliant product. It was definitely a learning experience for the team throughout the entire implementation stage. Our experience showed that it was definitely not just an alloy change as a lot of the people speculated at the initial stage of the project.
I would like to thank all our team who have contributed internally and the long hours they have spent to make this happen. In addition on behalf of JJS I would like to especially thank the Indium Corporation who contributed equally to the project.
Naim Kapuda is Manufacturing Engineering Manager at JJS electronics
Unconscious incompetence - Not aware of the existing process
Conscious incompetence - The person becomes aware of the existence and relevance of the Process
Unconscious competence - The person achieves 'conscious competence' in a skill when they can perform it reliably.
The person will need to concentrate and think in order to perform the skill
Conscious competence - The skill becomes so practised that it enters the unconscious parts of the brain - it becomes 'second nature'
Contact Details and Archive...