Major Changes?

The process consists of two interlinked aspects. On the one hand we have to deal with the physical reality, i. e. everything that happens around the production line. We have to add to this what normally is called ‘know-how’, technical information and experience related to materials, equipment, process parameters and the effects they have on the final product. But all this would be handing in thin air without ordering, delivery, storage and organization of components, printed circuit boards and material such as solders and fluxes, cleaners and solder iron tips, spare parts, nitrogen and all the other nitty-gritty stuff that is required to make production hum. We have a pretty new word for all this: logistics.

But companies is people and we have to make sure that they remain on a level with the technical requirements. When we switch from lead-bearing to lead-free solders keeping our personnel informed is a major task and of utmost importance. The latest level of information has to be brought across.

Such a system of distribution is well established at your company and we hope that it can be kept, largely unchanged, for the new technology. However, there is a danger if the system is not flexible enough because the clear and straight forward organization we are accustomed to will be lost when several solder alloys have to be introduced.

The law allows for a parallel usage of lead-bearing and lead-free solders (e. g. for export markets into non-EU countries) and for repair under warranty for previously SnPb soldered products. And it is entirely possible that cost pressure will force many production units to employ different lead-free alloys simultaneously. Add to this stringent requirements for components and PCBs and things can become quite messy.

However, there is also good news: near the beginning of the process the placement equipment does not seem to be affected by the type of alloy used, although the task of placing ever smaller components (‘bird-feed’) does pose a great challenge. Where we will ses a major impact is during the different soldering processes. And it is not entirely unlikely that we will regress into situations where every single type of product will require its own profile once more.

 

The Process

Processes can be visualized in many different ways. A fishbone diagram (Ishikawa-Diagram) can depict which steps have to be performed and which routes must be taken in order to ensure a proper working of a production line. Perhaps it would be good to take such a brainstorming session a little further, as many cost factors that impact production derive from such points as design and layout or purchasing. These sources of problems may become more marked as we enter into lead-free processes. Although we do not know whether we have to adjust our pad configuration in order to optimize manufacturability and reliability (at this moment in time the indication seems to be that changes are not really required), there are certain measures that must be taken: e. g. it is highly recommended that assemblies exhibit better thermal equilibrium between the different parts.

When purchasing components it will no longer suffice to define their values and perhaps their tolerances and dimensions. Pre-describing the metallization may have to be entered into the contract. Another important item will be the thermal loading during processing and thus we may have to specify as well those components should be able to endure. At least production has to be aware of those details.

Because we still lack proper standardization for printed circuit boards the properties of the laminate must be described very clearly otherwise the vendor may ship material that is not compatible with the process. Once we are at it, we might as well define the cleanliness of the PWB on delivery. We expect an increased amount of residue after soldering and thus cleanliness of material entering into it will be valuable. Only then will we be able to maintain a ‘no-clean’ process in the future. [We propose to request at least 50 % of the old MIL-value.]

So, do not hesitate to extend the fishbone diagram into areas such as purchasing, design and layout.

 

Procedures

We start with the actual situation and describe – before the backdrop of the physical presence of the line – all the relevant parameters used in the present production environment. As a second step, we go a little further and include the reception of goods. Finally, in a general sweep, we embrace purchasing and engineers and layout etc.

Once we set down the collection of parameters in use it should become clear why we are using precisely those (e. g. conveyor speed for optimal through-put, or requested by customer or because of sensitive components etc.) and at the same time we will find out where significant variance will occur from the present settings and values.

Once you are at it you may update the documentation to reflect the present situation. Check whether the ISO-9000 papers from your last certification effort still show the actual values that you find displayed on the equipment. It is certainly not rare to find that the machine operator has – in a very positive frame of mind in order to improve the process and to get better results – changed the settings unbeknownst to others. The question may be allowed whether his good intentions are based on a complete insight into the consequences of his actions.

We  have identified all the soldering processes in our factory and hence it is a simple matter now to identify all the sets of parameters in use. For some companies it will only be one set of values, for others each product will have its own. Sometimes products are collected into ‘families’ and then sets of parameters are identified with products in these families. In most cases these values can be found in the storage medium or stored in the actual machine’s control system. Add to the values all necessary comments and information that will allow you to make decisions later. Check with your design department, the layouters and purchasing whether they can add anything to this data.

 

Quality Assurance

Maybe we also should visit your department of quality assurance? Many of the pieces of information that you need should be available there. In particular you will find out what happens to goods received. Ideally, the understanding of quality is resident at the vendor. Regular meetings, discussions and quality audits at the vendor’s place are one way of ensuring a high level of product quality.

What does QA really test? Is it the value of components, their dimension or do they extend their actions to important process aspects as solderability, absorption of humidity in components, thermal properties or even actual tests of the surface finish of boards and components? Perhaps you are lucky and confirm that, indeed, all your requirements are already met. If not, there may be a lot of work heading towards you as a production that is interested in a high level of reliability and a low level of process defects will need to know much of the above mentioned information.

 

Purchasing

It is a common mistake, not only in the electronics industry, that purchasing has been reduced to only mercantile thinking: everything as cheap as possible. This ‚Cost Center’-idea on the principle ‘competition is good’ can cost a company a lot of money. It is quite reasonable to try to keep individual cost down, however it is more important to keep the cost of the end-product as low as possible. If low quality is purchased cheaply the end-product may become expensive due to the necessary repair processes. Saving a fraction of a Cent € for each component purchased may at first look smart, however if one has to spend tens or hundreds of € even to repair only 10% of the rotten product, this ‘smartness’ evaporates quickly. And we have not even considered the cost for potential warranty repair.

 

Soldering Processes

We distinguish in general two alloys: Bi-based solders with a melting point around the 138ºC of its eutectic with tin and hi-melting alloys that melt between 217 and 230ºC.

1. Wave Soldering

Table 1 shows typical parameters for wave soldering with different solder alloys.

Whenever it is possible to use bismuth solders (and it is clear that this is not possible for all products, however it may offer an possible solution for about 40 – 60 % of applications) there are a number of features that make a change-over to these alloys rather attractive. The low melting point and the fact that the tin content will not rise substantially make the purchase of new equipment unnecessary. Neither the traditional material used for the pot nor the pump will be attacked at these temperatures and at these levels of tin.

The difference between the specific density of SnPb versus that of BiSn is small and hence changing the conveyor angle will hardly be required.

However, there are three clear pre-conditions that have to be met:

a)       A strict ‘no-lead’ policy has to be introduced: no lead on components and no lead on the PCB can be permitted. No other process step that possibly can contaminate the joint with lead can be tolerated. Pot and pump may have to be exchanged if they cannot be cleaned sufficiently well to get rid of all Pb-residues.

b)       The maximum temperature at which the resulting product is to be used must be below 100ºC.

c)       Soldering must take place under inert conditions

As the solder joints are somewhat less malleable, mechanical shock of the joints should be avoided.

The eutectic BiSn-solder can be modified by adding small percentages of other metals, which we will deal with in one of the following articles.

Hi-melting solders can be seen in three classifications: SnCu+X; SnAg+X and SnAgCu+X – where X stands for Sb, or Bi or others. Their melting points lie between 217 and 227ºC and thus about 40 – 60 K above the melting point of the SnPb eutectic.

Again and again one hears that such solder are used at pot temperatures of 255ºC. This may be marginally acceptable for a product that is thermally not very demanding, however, as soon as complicated or heavy assemblies must be processed, such temperatures are unrealistic, which is testified by a number of internationally published articles. Whoever is running these low-temperature experiments tries to fulfill some wishful thinking, sidestepping the problems that can arise because of the high application temperatures. Naturally, when capping the temperature at 255ºC, leaching of 316 or even 305 steels is much less and the amount of dross generated will also be lower. It will also not be required to raise the pre-heat temperature to the same degree as if one has to enter into a 285ºC solder flow. As any process engineer can testify: many things are possible in a laboratory environment that will never work in actual production situations. Thus it is not surprising that quite a number of such experiments have taken place and reported that the results with 255ºC solders are acceptable or even ‘good’. However, in the real world we can not expect that such ‘fiddling’ will be successful and that one would obtain acceptable defect rates in the process and at the same time the required reliability for demanding products. In our opinion, a high level of touch-up and repair will automatically put an end to this type of daydreaming or intentional whitewashing.

Some of the parameters of the process, such as e. g. the cooling rate, will achieve greater prominence.

 

2. Reflow processes

In table 2 typical parameters for reflow soldering, again adapted to the different solder alloys, are shown.

Hi-melting pastes will narrow down to the two families SnAg+X and SnAgCu+X as the melting point of SnCu+X will be considered too high. Their thermal process profiles will shift to higher temperatures (see pic. 2 – reflow soldering profile for hi-melting solder pastes).

The peak temperature ranges at about melting point of the alloy plus 30-40 K, hence approx. 245-260°C. Peak temperature of about 235ºC abound in the literature and they may be quite feasible under excellent nitrogen inertion. However, due to solderability problems encountered on components and boards and the general lack of good wetting properties for the lead-free alloys, we anticipate a slow adjustment of the peak temperatures in the area mentioned above.

Choosing the best „plateau-“temperature - as close as possible to the melting point of the solder (approx. 220ºC), however not so high that the activators of the paste will react pre-maturely – most likely is one of the most difficult conditions to meet. The activators in the paste (Di-carbon series of acids) disassociate at about 160ºC and thus substantially lower than we would wish, i. e. at temperatures of about 245 – 260ºC minus 60 K = 185 – 200ºC. Profiles without plateau will become more general especially wherever either a better thermal design of the assembly or better equipment prevent the development of a major ΔT. On the other hand, we do not yet have reliable information about the effect of an increased thickness of the intermetallic layer on the reliability of the lead-free joint. Thus we do not even know whether a large ΔT will have a negative impact on the reliability as in the case of SnPb. Perhaps we do not have to react to large ΔTs (except perhaps for components) as before and we might be able to use the pyramidical profile without paying the Δt-price?

 

Hand soldering

Hand soldering is hardly used for entire assemblies any more in Europe. Most of its application has been limited to touch-up, repair and selective soldering processes. In each one of these cases it is of utmost importance that the alloy – and if at all possible the flux – matches the solder in the main processes. Which means, that for unleaded solder processes that we have to distinguish even more clearly which material is used for those secondary treatments.

Setting the temperature of the solder iron tip depended always on melting point of the solder:

Melting point of the solder + ‘super heat’ for the required heat transfer and flow of the molten metal + approx. 70 K for the temperature loss experienced during contact of the tip with the joint. If one wanted to work at about the same temperatures as used during flow soldering, this yielded the following calculation for eutectic SnPb-solder:

183ºC + 70 K super-heat + 70 K temperature loss ≈ 340 – 350ºC (with a certain ‚margin of safety) and not the 450ºC (as far as one can turn the knob) as regularly observed in practice.

Let’s transfer this to the lead-free model:

Lo-melting solder:

138ºC + 70 K + 70 K ≈ 270ºC

Hi-melting solder:

220ºC + 70 K + 70 K ≈ 360 – 380ºC

But perhaps we can use hotter soldering irons in future if components accept this temperature too. If the thickness of the intermetallic layer is no longer a concern with regard to reliability then we can accept a hotter iron and the resulting thicker layer.

Other solder processes hark back to the related process discussed above.

Although disturbingly many of the important facts required to establish the ‘ultimate’ profile for wave or reflow processes have not yet been investigated we should be able to take the ones presented above as a starting point for such an optimization procedure. Add to this the requirements of components, printed wiring board, equipment and other process material and we should be able to ‘make do’.