AMS is a company in transition, and managing that transition will be important. Before you read Unit 10, we strongly advise that you think about the range of products made by AMS (Scenario 1), and consider the options for making the conversion to lead-free equivalents.
Unit 10 focuses on the technology transition, but inevitably you will pick up many insights into the full scope of the tasks awaiting the AMS management team.
So far in the module you will have struggled with an enormous amount of information to assimilate, and realized that the task of managing the transition to lead-free involves constant updating as the industry moves towards July 2006. You will also have discovered that the transition is not clear-cut, and that the best solution for a company will depend very much on the range of products made, the resources available, and the attitudes and prejudices of the end-customers. At the same time, you will probably have gathered an impression of the size of the task and thought about how the transition might be implemented. In the final three units of the module, we are looking at the process of transition, and the implications for the company and the way it operates, and returning to the diagram we used in the introduction to describe the implementation process (Figure 1).
In Units 1 and 2 and in Assignment 1, the focus was on the implications of the lead-free transition for the company and its customers; in Units 3 to 8 and in Assignment 2, our attention moved to the planning area at the bottom of Figure 1, in which we examined the options for going lead-free and exposed some of the issues. The task that formed part of that second assignment, where you looked at the implementation issues for the components on one particular circuit, will have at least scoped the task of transition. In Assignment 3, which you will be considering as you read these next three units, you will be asked to complete what in the real world would have been the end of the preliminary audit – starting with a brief statement of the implications for the company and the options available, and then looking at practical ways in which the company might make the lead-free transition and avoid or overcome some of the likely obstacles.
In this unit, we are concentrating on the task of managing the transition to lead-free, and looking in more detail at the transition process itself. Our starting point is the result of the audit, as this will have identified which of the current products need to be migrated to lead-free, indicated the likely migration path, and estimated the likely response from the end-customer. Although it is important never to take the situation for granted, typically a company will have some customers that are more proactive and positive towards lead-free than others, and there may even be some who need to be “dragged along screaming”. This suggests that the phasing of the transition will be dictated by the end-customer, although there may be some variation due to component supply considerations.
For most companies, going lead-free will involve making available lead-free variants of all soldering processes, hand, wave and reflow, which need to be combined in varying ways to make a final lead-free product. So, when developing options and making our materials choices, we need to keep in mind the fact that more than one of these soldering processes may be involved. For example, if we decide to use SnCu for wave soldering, and SnAgCu for reflow, which material do we use for rework? Rework needs either to use both materials, or (more likely) to be carried out with SnAgCu solder, with the decision implied, even if not made explicitly, that the reliability of the resulting mixed-alloy joint will be sufficient. For most customers, this is probably not a issue, but the company (or group within the company) that is carrying out the assembly work needs to have this discussion early on during the process of developing options.
The task of developing the options, and selecting which to evaluate, is one that is often carried out by the “prime mover” within a company – the engineer who has been most “bitten by the lead-free bug”, and is trying (sometimes against opposition) to move forward the company’s response to the lead-free challenge. Whilst pragmatic, this is in general not as good a policy as involving the widest possible range of stakeholders. There are enormous benefits to be gained from identifying the people/functions within the company that will be affected by the lead-free transition and “getting them on board” as early as possible. Whilst one should never assume that the stakeholders will be ignorant of lead-free issues and “talk down” to them, even at this early stage it is good policy to spell out the consequences of non-compliance and give at least an idea of the issues and likely timescales. Although a formal costing of the transition process will not yet be possible, some indication of the order of cost is appropriate when involving senior management, at the very least indicating the level of engineering and management resource likely to be required. At SAMC, we found it useful to articulate the issues and the process in the way shown in this slide show.
As with any problem-solving, developing options and selecting those to evaluate can be approached by both convergent and divergent ways of thinking. Most engineers find they are happiest when searching for data along defined tracks and pulling it together in a structured way, and this is probably a good way of tackling sections of the task such as “Which of the alternative materials shall we use?” However, there is a definite place for “out-of-the-box” thinking when it comes to deciding how to deal with suppliers and customers. In particular, using as many departments as possible to think laterally about the issues, perhaps using a formal brainstorming session, can expose both problem areas and opportunities for harnessing the co-operation of suppliers and customers as well as in-house contributors.
The process of developing options will have selected certain processes to be evaluated, but will not have defined the processes in detail and may not even have selected specific materials. For example, although the appropriate decision for a particular company might be to use an SAC formulation with a VOC-free flux, and to clean parts after soldering because a conformal coating is to be used, there are a number of suppliers of such materials, and some variation in the silver content. That alone will create a long list of potential pastes, which can be whittled down by applying ‘screens’, asking questions such as:
Notice that many of these screens are more commercial in nature than specifically technical. Even if no other stakeholders are involved, the procurement department needs to be happy about the suggested suppliers and committed to work with them to a successful conclusion. Given that lead-free solder will eventually supersede tin-lead solder, with obvious consequences for the supplier, existing vendors will usually support a process engineer looking to make a change. But the change also provides an opportunity to look for an alternative supply partner who might offer commercial or technical advantages. Involve your Purchasing Manager in the decision process!
The evaluation process is simply stated:
Simply stated, and probably what most people do as a “quick and dirty” implementation of a process change. So what is missing?
The first element missing is a formal definition as to what constitutes success, made before the product is prototyped. For example, is good visual appearance sufficient? For some customers it may be, but others will require confidence that the resulting joint will not “fall apart” when temperature cycled. Another perspective comes from the production manager, who will be keenly interested in the process yield, as an increase in defects even at the ppm level can affect yield and costs.
In an ideal world, the company will already have well-defined standards, previously-developed procedures for product evaluation, and SPC charts monitoring the processes. This makes it easy to compare the new lead-free product, with existing materials and processes. But, for many small companies, such procedures are less than fully-developed, and the data correspondingly scanty. Knowing the potential for the eventual comeback “Look at all the problems we’ve had since we went lead-free!”, in some situations it may be advisable to include the current process/material as one of the options being evaluation.
Evaluating several material options under a range of process conditions, but being unable for economic reasons to make large numbers of parts, brings the danger that happens with any small sample sizes of trying to draw sufficient conclusions from insufficient data. Rather than just let the experiments develop (“a tweak here; a tweak there”), there is much to be said for using a formal Design of Experiments approach to running the tests and analyzing the results. Detailed consideration of this technique is outside the scope of this module, but is definitely worth considering – if you look at almost all the optimisation work reported in the literature, DoE has been used to improve the validity of the data and contain the cost of collecting it.
Design of Experiments uses statistical methods, and we can’t leave this topic without some mention of Six Sigma, as this approach (pioneered by Motorola) is commonly used within the electronics industry. Whilst the name refers to statistics concepts, and there are links to ways of getting defects below 1ppm, the Six Sigma technique is better described as a structured way of solving problems.
Six Sigma works by defining the problem, finding out from the problem owners what will constitute a successful outcome, generating a range of possible solutions, evaluating these options, implementing the preferred option and evaluating it against the original criteria. Use is made of statistical methods where appropriate, and there is an emphasis on the “voice of the customer”. Although very much a “big company procedure”, and one where such companies have invested heavily in training their employees to take a regimented approach to its implementation, Six Sigma has the merit of creating a framework for problem-solving that is accepted throughout a company. If you are a Six Sigma company, lead-free implementation could be your “black belt” project . . .
In their Stencil Printer Optimization Study, John Stevenson and Derek Drabenstadt used controlled experiments to optimize the printing of solder paste. Whilst the paper refers to Six Sigma, it is more a good example of experimental design, using the power of statistics to assess which of the 39 (!) possible process variables are really important for control. And the outcome, which was to define material handling and equipment better and install control charts, has resulted in a dramatic decrease in defects and rework.
Another example of good design of experiments is shown in the papers on Total process control and management by Paul Wang and his colleagues that appeared in the April/May 2004 issues of Circuit Assembly. In this case, a wider range of work was undertaken, including computer modelling.
So far, not only have we selected from a range of options, but the list has been further refined to one/two materials and the company has developed a process that is capable of giving the best results with the chosen material. The next stage, apart from a sanity check that the methods and materials meet the initial customer requirements and are compatible with the intended equipment, is to be certain that the end-product meets the customer requirements.
Here we have a genuine choice, of testing real products or carrying out the same testing on specially-designed boards, normally referred to as “test vehicles”. The advantage of a test vehicle is that it is designed to be tested, in the same way as the test coupons on a printed circuit board. Knowing the most likely problem causes, the test vehicle can be designed to be as sensitive as possible to these potential failure modes. Our example of this relates to a printed circuit board. In the past there have been problems with the ductility, elongation and crack resistance of the copper plating, leading to via holes going open-circuit on temperature cycling. So a typical test coupon contains vias that are “daisy-chained”, so that a single measurement can verify the continuity of perhaps many hundred of vias. In this case, checking the test coupon is clearly much more effective than trying to assess the integrity of individual vias within a working circuit.
At the same time, the test vehicle can avoid the use of expensive components, because the parts used only need to be physically the same. A final advantage of a test vehicle is that it can be designed to be representative of a wide range of products, so that reliability testing of just one design might be equivalent to proving the reliability of many different products.
This was the idea behind the “Capability Qualifying Circuit” originally devised for BS9450 (the rules for capability approval of electronic components are now defined in BS CECC 00114-3:1993). However, the CQC idea received a lot of criticism:
All these problems add up to increased cost and lack of commercial viability. The result has been that, except for specialist applications, the idea of using a formal qualification procedure to test only representative circuits has not been generally adopted. And the brave attempt to qualify a whole technology by similar means (BS CECC 210000:1995) has sunk without trace.
However, although this is not always realised, similar concepts lie behind normal commercial practice. This is to test a limited number of parts of a representative assembly, and then make a judgement as whether the data collected for that assembly and the confidence gained in its performance are sufficiently representative to allow the findings of the test to be extrapolated to other designs.
Note that testing circuits always involves some degree of risk, however small. Even if the parts tested were identical to the ones whose performance we wished to assure, there would still be a risk that the test conditions were not representative of life, especially if we had selected an accelerated test regime in order to reduce the test time. The further we depart from exact equivalence of test conditions/module under test and the “real thing”, the greater becomes that risk.
Fortunately, when it comes to “general causes” of failure, failure rates will be relatively independent of the structure. For example, a 224-pin BGA on one board is likely to have a similar reliability to a 300-pin BGA on another design, assuming that the BGAs are of the same construction. Where we need to be careful is when it comes to what reliability engineers refer to as “special causes” of failure, those where failure is not random and low-level. The example of a “special cause” that springs to mind is a set of chip ceramic capacitors near the edge of a board that might have been expected to have the same failure mode and rate as those at the centre, except that they were subjected to excessive strain during the process of depanelling! However, making the reasonable assumption that such special causes are design-dependent, so will vary little between lead-free and eutectic tin-lead solder, we are tolerably safe (for most applications) to choose a real circuit for our test vehicle, at least as regards run-of-the-mill components. This allows us to work with the customer for that product to develop the process and evaluate the end result.
But prudence suggests that we should look critically at the assembly, and then look at the components used on other modules to try and identify any features that were not included in the initial test and thus might need further evaluation during the lead-free implementation.
Having indicated some of the issues, review some typical published solder joint or lead-free evaluation programmes. We have suggested a few which may be of interest. How might that work be mirrored in your own company? Or at AMS?
Review some typical published solder joint or lead-free evaluation programmes, to form a judgement as to what is needed in an implementation programme, and how your company might carry out an equivalent to the work reported.
As you do this, try and summarise the extent to which each programme was successful in delivering a reliable product on the intended timescale, and try and identify the reasons for any success. Making notes on this will help prepare you for Unit 11.
Search for "lead-free soldering" +implementation +program, or "lead-free soldering" +evaluation +implementation, or look at some of the resources below.
Celestica’s June 2003 presentation Development and implementation of lead-free assembly has insights into the extent of the work involved in making a managed move to lead-free. Rob Horsley describes how preliminary work grew into two stages of formal evaluation, and indicates how the programme was managed.
As part of earlier studies, you may have come across Epson’s Road to Lead-free Soldering, which considers in detail both reflow and manual soldering processes using SnAgCu solder and wave soldering using SnCu solder. But the paper also looks at the reliability of different materials, and uses Taguchi techniques to show that SnAgCu solder gave a higher bond reliability than eutectic tin-lead.
Look at the controlled implementation of lead-free in the Motorola QuickStart project. There is a short description in Lead-free solution at your fingertips (EM&T, May 2002), but lots more information at http://www.pbfree.com/. Based on their experience, Ron Lasky and Tim Jensen of Indium Corporation have also written How to implement Pb-free assembly at your factory: 2004
Reading the information on these published evaluation programmes, and on the implementation activities of these larger companies, can be very off-putting for the smaller manufacturer. The question it raises is “Can we afford £1M?”, with an implied negative answer. However, we have to remember that larger companies not only have larger budgets but are also more vulnerable should things go wrong, and so perhaps need to take a more cautious attitude to risk.
However, with a small company, even a small loss or relatively minor problem could have disastrous consequences, so no company can afford to be cavalier in its attitude to risk. In our view, it is important that the risks should be explicitly identified at an early stage. However, the overall expense of the evaluation programme can be pared down, firstly by taking advantage of the experience of larger players, and secondly by involving customers and materials/equipment suppliers in the programme.
“While the work of lead-free industry giants will lessen your task, it will not take it away.”
Ron Lasky, Surface Mount Technology (SMT), August 2004
The final stage in the transition process is to implement the preferred method. Not only must the process itself be changed, but the procedures need to be documented and operators trained, making appropriate changes to inspection and test and the methods of monitoring. Your reading around lead-free issues should have convinced you that process controls need to be tightened; if SPC is not already used formally, then there is an opportunity to introduce appropriate controls at the same time as making the transition to lead-free.
In Unit 8 we used John Lau’s model of the alternative routes by which a lead-containing assembly could be converted to totally lead-free. As a reminder, read this link, which makes it clear that the choice of whether to convert materials or components first will depend on the company’s situation.
Managing the transition process means managing every aspect of the transition, and this means identifying the problems that the company is likely to face. This is something Assignment 3 will be asking you to tackle for AMS. One potential problem area is commented upon by Engent in their Lead-free proposal, which indicates that in many high volume applications companies are being forced to deal with lead-free solder termination on commodity components much earlier than the road map suggests. As a result, “ a number of significant solderability and yield problems can occur in even the most robust processes based on eutectic tin-lead solder”, and they therefore recommend early work on the solderability of lead-free components. This is certainly something that needs to be borne in mind by companies of all sizes should they decide to take Route C to lead-free.
For most companies, the transition will involve coping concurrently with several different technologies, and the implications of this affect both equipment and materials. For equipment, the choice is whether to dedicate a particular machine to tin-lead or lead-free, or change between the two. Depending on the type of equipment, this will involve varying amounts of changeover and setting up of the new conditions, and may have some risk of cross-contamination. Certainly care has to be taken that the status of equipment is identified, and that materials are very clearly marked. Look for examples of “mistake proofing” used in the industry, and think about how you might ensure that the right materials are always used with the right processes on the right assemblies – not as easy as it seems!
In Lead-free Implementation: the other 90%, Phil Zarrow and his colleagues from ITM Consulting stress early in the presentation that a decision must be made either to go 100% lead-free or to maintain two processes, and make it clear that implementation and logistics involve significant effort, which is what they refer to as “the other 90%”. They quote Kim Hyland (Director of Process Integration at Solectron) as saying “Logistics are the most complicated part” and reflect NEMI’s comments that individual companies will have unique requirements, and that management of these issues will be specific to the company.
For both boards and components it becomes crucially important to know their lead-free status: Do they have lead-free terminations? Are they lead-free compatible? There are two aspects which we would like you to research, and form your own views about.
Two suggestions for starting points:
A related issue concerns what is sometimes referred to as “materials declaration”. In other words, how do we indicate within a legal framework that our product is RoHS-compliant, and state the amount of (allowed) lead in the product? [When introducing the End-of-Life Vehicles Directive in Unit 2 we mentioned the requirement, at least for some applications, for the assembler to declare the content of prohibited materials, and specifically lead]
In their report Possible compliance approaches for Directive 2002/95/EC, made to DTI in April 2004, ERA reported on different potential approaches to giving producers and enforcement authorities a simple low-cost procedure to ensure that no banned substances were present. [Of course, at the time the work was done, no agreement had been reached on the interpretation of “homogenous material”, although it now seems that the tightest of the possible interpretations has been adopted, where the banned materials cannot be used even as thin coatings]
ERA’s conclusion (lots of background, because this is a heavyweight report) was that producers should take responsibility for RoHS compliance, but simply make declarations to this effect, having obtained assurances from their suppliers that materials, components and equipments did not contain the banned substances. This recommendation was made on cost grounds, because ERA felt it unreasonable to expect producers to analyze incoming materials.
This is fortunate! It is a far from trivial task to check components on arrival, as most companies will not have easy access to an analytical laboratory, and there is in any case a high cost associated with that kind of service.
One of the practical difficulties associated with the lead-free conversion is that it is not easy to be certain whether or not a particular batch of components has lead-free terminations, let alone whether the parts contain lead internally. As Bob Willis asks in Global SMT & Packaging (April 2004) “How do engineers determine if the surface is free of lead?”. We are aware of three techniques available for testing that surfaces are lead-free:
Only the last offers any potential for in-house use by most companies, but care needs to be taken to make sure that the results are valid. Bob Willis reported that, in order to give reliable results, leads need to be cleaned with alcohol before testing.
Although actual testing is not required, ERA recommend that producers should still be expected to take “reasonable steps” to comply. When one considers the historical background, it is highly likely that, at least to start with, ‘compliance’ will be based solely on maintaining a ‘paper trail’ of compliance statements or their electronic equivalents. Unfortunately, standard formats for transmitting such statements have not yet been fully developed, although a number of options are being considered including the “Substance Declaration of Conformity” (SDoC) that is being worked on by IEC.
The only encouragement from ERA’s conclusion is their recommendation that “Markings on components providing a readily visible indication that the banned substances are absent would also be suitable. This approach would minimize the need to maintain paper or electronic compliance records”. This does depend on standardized marking and the legislation allowing this to be used. Rather than say “watch this space!”, perhaps we would be better to remind you that there is a good up-to-date resource on the DTI web site.
Compliance is an area that attracts lawyers, so you should not be surprised that there are proposals for a statement with wider scope contained in this IPC Material composition declaration guide for electronic products, which was developed by representatives from industry associations. This source is worth noting by those with a longer view than just lead-free, because it refers to a much greater range of controlled substances. The cost of compliance may be expected to be monotonically increasing!
In the medium term, and certainly within the transition period, coping with multiple technologies is a likely scenario for most companies. In order both to protect invoiced sales, which is vital to the survival of a company, and to minimise the cost of stock made obsolete by the change, the move to lead-free cannot be implemented “over the weekend”. Most consultants would recommend phasing the transition with care, starting with an initial pilot activity on a suitable representative assembly. The lead-free implementation on this sample would be carefully monitored, with feedback to improve both process issues and any logistic problems.
Solectron’s April 2004 presentation on Lead-free production contains some useful thoughts on compatibility and the transition process. A parallel presentation by Jasbir Bath on Lead-free manufacturing issues and transition to lead-free soldering contains addtional information.
Ideally one should have completed the pilot before starting work on converting other designs, but pressure of time from the July 2006 deadline means pressure for parallel development activity to convert all designs. As we discuss in this linked section on concurrent engineering, parallel development is an activity that needs “rich communication” between all those who are involved. And this includes the customer – ideally the product made should be “for real”, in the sense that parts made will be evaluated in a field situation, whether that customer is another part of the same company or an entirely separate organisation, as in the case of a contract manufacturer such as AMS.
Running a number of experimental activities in parallel puts substantial demands on the management of a company – hence the stress on “rich communication” in concurrent engineering. For most companies this will mean pulling together a lead-free implementation team, and keeping the situation under continual review. The review process is conceptually similar to that used in ISO14001 for assessing the status of a company in relation to its environmental response. As with an environmental audit, actually scoring the company is one way of ascertaining how far along the path to lead-free the company has gone, and making everyone aware of the points of weakness.
“Moving from a lead-based to a lead-free electronics assembly process is like travelling on a journey. In either case it is critical to know three things: where you are now, where you are going, and the conditions that will occur in between”.
Ron Lasky in Surface Mount Technology (SMT), August 2004
In his article Ron Lasky promotes a lead-free status assessment, as this helps clarify where the company is at any point during its progress towards lead-free. Not only does this give a base-line, but it also indicates the resource areas that are needed by the company.
Ron was involved in the development of the one tool we found that helps answer the question “How ready is the company?”. Produced by Indium and Motorola as a result of their QuickStart programme, “Pb-Free Readiness Assessment©” is “an online tool to help you identify your company’s strengths and weaknesses in implementing a comprehensive Pb-Free workflow.” Before you move on, take a look at this questionnaire, and insert the answers that are appropriate to your company.
Maybe you don’t agree with all the questions, but at least they cover most of the issues. When you have completed each of the four sections, you will get feedback from the site as to how your company measures up against other companies involved in the lead-free transition. When we did this, we were surprised at the relative level of “maturity” of the industry on some elements, and its comparative state of unpreparedness in other areas. Bear in mind that this is a live site, so that the results you get are influenced by all the respondents. Then reflect on how AMS would stack up against the industry average on these various parameters – it will help you identify the management activities that you need to comment upon in Assignment 3.