NPL/SMART Group Conformal Coating & Cleaning Experience
By far the best-attended seminar-workshop event at an excellent National Electronics Week (NEW) show was the NPL/SMART Group Conformal Coating and Cleaning Experience, seamlessly organised, coordinated, and managed by SMART Group Technical Director and showman extraordinaire Bob Willis.
Willis combined the exhibition and demonstration area with the seminar theatre to offer an integrated learning experience dedicated to cleaning and coating. The area was a bustle of activity for the three days of the show, several visitors having travelled from mainland Europe and Eastern Europe with attendance at this NPL/SMART event as their primary objective, and several UK assemblers I spoke to were looking to gain a clear understanding of the technology and available options before embarking upon in-house conformal coating.
An hourly series of short, punchy presentations from industry experts, interspersed with opportunities to ask questions and to see equipment and materials in real-time action, maintained a consistent momentum. In Willis' words: “It’s about getting them interested to start with!”
Getting people interested in design considerations was a task expertly undertaken by David Greenman from Humiseal. Appropriate and adequate conformal coating drawings were rarely provided, and conformal coatings were commonly specified for historical reasons rather than properly selected to suit the anticipated operating environment. It was important to check the compatibility of a coating with the components of the assembly, and particularly to determine the wettability of the solder mask to ensure good adhesion. “No clean is no option” was his comment about surface cleanliness prior to coating and he recommended SIR testing as a means of qualifying the complete assembly process.
From a design point of view, it was necessary to carefully consider component layout, especially if dip coating was specified, to avoid features such as open through-hole vias, and to add tooling holes appropriate to the application method. The essential elements of a conformal coating drawing were definitions of: cleanliness requirements, areas to be masked and method of masking, conformal coating type, application method, thickness specification and inspection standards.
A good crowd gathered to hear Phil Atkinson of Fraser Technologies put his perspective on the need for cleaning before conformal coating. Having outlined the general benefits offered by conformal coating: Protection against moisture intrusion, humidity, shock, vibration, and corrosion, as well as electrical insulation, he summarised the properties and relative merits of the range of coatings available: Acrylics, urethanes, epoxies, silicones, and polyparaxylenes. There was no universal solution; it was a matter of intelligent selection in respect of performance requirements such as operating temperature, moisture resistance, chemical resistance, and whether formal qualification was required.
Why clean? He described the surface of a PCB assembly as a “chemical residue soup” which, before lead-free, did not generally present a significant problem and most conformal coatings could be successfully applied directly without cleaning. However, as technology had progressed with tighter packing densities, higher operating frequencies, and higher power, coating without cleaning was not recommended. Lead-free had been the real game-changer, with higher process temperatures and more tenacious residues, and he listed a whole range of contaminants that could be present--originating from the PCB and components as well as from the assembly process itself--and gave examples of failures that they could cause. In summary, he advocated that cleaning before coating was good practice, and recommended ionic cleanliness testing to prove its effectiveness. An appropriate cleaning process could promote optimum coating adhesion, whereas a poor cleaning process was the root cause of many coating defects.
The third presentation came from Willis himself, giving advice on selecting a coating based on product and environment. Why conformal coat? He listed several purposes: To prevent premature failure, to improve the performance of specific components, to improve environmental protection, as a possible alternative to a sealed enclosure to reduce the cost of the end product, and to ruggedize a product design. He then summarised the characteristics of available coatings and typical coating thicknesses. A golden rule was never to use conformal coating to overcome process problems by attempting to cover them up--it was vital to have a properly engineered and controlled assembly process in place before considering any conformal coating operation. He showed many examples of process and field failures, explained their causes, and recommended consulting the NPL Process Defect Database--freely available--as a point of reference to help identify and resolve problems.
How to rework and repair conformally coated PCB assemblies was the topic of the presentation by Humiseal Technical Support Specialist Jon Anderson. He cited IPC 7711/7721 as a comprehensive work of reference for training engineers in rework and repair procedures for conventional and surface mount components as well as procedures for the removal and replacement of coatings and the soldering and de-soldering of different component types.
IPC 7711/7721 gave guidance on identifying coating types and suggestions as to which removal methods were the most appropriate for which coatings. Although it was possible to rework through coatings without removing them, it was not recommended to do so because of the risk of contamination of assemblies and tools with breakdown products.
Selective coating removal could be achieved by solvent swelling, and solvent pens were available from the suppliers of coatings. Alternatively, mechanical methods could be used, both for selective removal and for clean-up after rework. Typical systems used micro abrasive blasting, with dry media selected to suit the application; natural materials, plastic or mineral abrasives were available in a range of grit sizes. Anderson showed many examples of selective coating removal he had carried out at SMART Group workshops. Once the rework operation had been completed, the conformal coating could be reinstated by brushing or manual dipping.
Willis took the stage again for the day’s final presentation: A discussion of process and field failures, beginning by showing the results of a SMART Group on-line survey which indicated that the most common conformal coating defects were de-wetting (36%), bubbles (27%) and missing coating (13%). He used video clips to illustrate how such defects occurred, emphasising the value of assessing the surface energy of solder masks, and the importance of controlling viscosity and hold times in the coating process. Coating specifications should define where coating was required, where it was not required, and particularly where it did not matter, to avoid unnecessary re-work.
Dendrite formation was the most common field-failure mode in his experience, a consequence of ionic contamination in the presence of moisture and electrical bias, and once again his real-time video examples clearly illustrated how it occurred.
Regarding re-work and repair, his advice was to remove conformal coating before de-soldering components, to minimise the damage, to keep it simple and to clean the tools afterwards!
To dip or to spray? The second day’s seminar session opened with a discussion of the relative merits of different conformal coating techniques from Jacques Mycke of Nordson Asymtech.
There were many choices to be made, depending on quantities, sizes and dry film properties: Automatic or manual, total or selective coverage, solvent based, non solvent based or water based, drying or curing, with heat, moisture or UV?
Brush coating was by far the cheapest and simplest method, but slow and heavily dependent on operator skill. Dip coating was a simple mechanical technique, but entry and withdrawal speeds were critical factors and there would always be thickness differences top-to-bottom, and a risk of air entrapment. And dip-coating was not suitable for moisture-cured or UV-cured materials. Masking was time-consuming and costly, and the subsequent removal and disposal of masking material added to time and cost.
Needle dispensing and jetting were useful techniques for selective coating of small areas, and spray coating, manual or automated, was the preferred method for overall coverage. Selective spraying was flexible and repeatable, but could require masking because of overspray unless additional needle or bead dispensing was used to contain or hide feathering.
Film coating was a specialised, patented, airless technique for selectively applying lower viscosity materials, which were dispensed as a controlled film rather than an atomized spray. It was a high speed process with high transfer efficiency and single pass coverage, and the need for masking was virtually eliminated. Film coating could be combined with jet coating in a single machine, with the film coater covering the larger areas and the jet covering highly selective areas, and a gel applicator could be added to make dams and plug vias.
The most theatrical presentation came the following hour, from IPC’s Lars Wallin, entitled, “IPC, ESA, and NASA Inspection Requirements for Coating,” with a sub-title: “Is it clean for coating?” Scandinavian eccentricity demanded the attention of the audience. “Do you understand my English?” he asked delegates from Spain and Estonia, in the manner of a stern schoolmaster, but with more than a hint of mischief in his voice. “Are you an expert?”
Continuing with quick-fire questions: “How many IPC standards are there?” “How many refer to cleaning and coating?” and demonstrating their relevance and applicability with the aid of an assembly-shop layout drawing, he arrived at the leading question: “Is it clean for coating?” to which he gave a short answer: “Probably” and a slightly longer answer: “It depends…” backed up by a secure answer: “You can't take the risk of not cleaning” and a cheap answer: “It costs too much to clean!” Then the hard questions came: “Do you have your cleaning process under control?” “Are you sure that your process chemistries are electro-chemically compatible?” “Do you know how clean is clean?” “Does a ROSE or SEC test tell you whether your PCBAs are clean?
Nominating flux as the root cause of most cleaning issues, he speculated on how many different flux chemistries there might be on the market, arriving at a figure of approximately 150, and asked: “Do all these fluxes demand the same cleaning?” Likening the nomination of a single cleaning process to a game of Russian Roulette, he explained the significance of J-STD-001E requirements for soldered electrical and electronic assemblies, and its associated standards for solder, flux and paste, and the differences between “Acceptable,” “Process Indicator,” and “Defect,” with a reminder that an assembler who does not have these standards included in his own specification cannot claim to comply with them.
Finally, he reviewed IPC-610: “It won’t solve the problem--it will cost you money!” and the workmanship standards for polymeric coatings defined in NASA-STD-8739. His closing words were: “The winner has a plan, the loser has an excuse!”
Nordson Yestech AOI Specialist Andy Bonner talked about techniques for automated conformal coating inspection. He reviewed systems with top-down and side-viewing cameras and a choice of white, red, or UV illumination, and algorithms specific to conformal coat inspection and parts inspection. “Look for the blue bits, and a few other things…” giving examples of bubble defect detection, coating or no coating on pads and through hole pins, coating inside connectors, and cracks and delamination that could be detected. Side angle camera inspection gave the opportunity to inspect coverage on hidden coating areas and review tools enabled operators to determine whether true defects existed.
Detection of defects was only part of the story: It was also possible to record all of the defect data and generate detailed reports on assemblies, works orders, individual boards, systems and operators, and to seamlessly drill down from an assembly to an individual defect within the report viewer.
Next hour, attention turned to the cleaning of PCB assemblies for coating, with a presentation from Graham Fraser of Fraser Technologies on high density interconnects, assembly process cleaning integration and assembly cleaning options. He made it clear that although electronics cleaning had made substantial progress in the past 40 years, lead-free soldering and low stand-off components in high-reliability applications meant it had never been more important to ensure that cleaning was both effective and measurable. And as the complexity of electronic assemblies continued to advance, the removal of process residues became an increasingly complex issue.
He discussed the cause-effect relationships between board design, paste or flux, cleaning agent, and cleaning machine to be considered in achieving an ideal cleaning condition.
When no-clean fluxes were first introduced, spacing between conductors was larger, residues posed minimal reliability risks, and, in many applications, cleaning was not required. As distances between conductors narrowed and Z-axis clearances reduced, residues of non-reacted flux activators under component gaps posed an increasing reliability risk.
“Flux residues are not created equally!” Today’s soils tended to be harder to clean, and cleaning agents had evolved to match the soil properties. Increased wash times were needed as gaps became smaller. Wash temperature was dependent on the nature of the soil: For some soils, higher wash temperatures were preferred whereas on some others, lower wash temperatures gave better results. Wash concentration needed to be optimised to match the cleaning agent to the soil, and an increase over optimum gave no technical benefit but increased process cost. And getting the cleaning chemistry out of the assembly at the end of the process was an equally important consideration--all aqueous formulations would leave residues if not completely washed out.
Reviewing the choice of cleaning equipment available, he stressed that the importance of cleaning needed to be considered as a priority in setting up an assembly line: “Don’t spend all the budget on the latest pick-and-place machine then find you have to cut corners on the cleaning process! Talk to your suppliers and get your evaluations done up-front” was the message.
Willis rounded off the second day’s programme with a review of available literature on cleaning and practical advice on confirming component compatibility with cleaning materials, performing simple tests like weight change after immersion, stability of marking, and degradation of wrapping materials. Some component issues were very obvious, but care needed to be taken with assessment to avoid short- or long-term reliability problems. A contentious issue was the presumption that ultrasonic cleaning was potentially damaging to components with wire-bonded interconnects. In his long experience, he had never seen a failure attributable to ultrasonics, even in instances of grossly excessive exposure.
Bob Willis introduced the final day’s seminar and workshop sessions with a discussion of the advantages and disadvantages of performing the conformal coating operation in-house or contracting it out. He outlined he requirements for a production facility: A dedicated area of the factory with temperature and humidity control, ventilation and extraction, facilities for material storage and material disposal, as well as coating, drying, inspection, and rework equipment, and associated health and safety considerations. The benefits of using a contract service included access to considerable expertise in coating and no capital equipment or additional personnel costs. Weighed against these were the considerations of lead times and work scheduling, costs of coating service plus packaging and transport, and the need for vendor assessment and approval. To illustrate cost, Willis did calculations based on price per hour for labour, overheads, and profit, together with material costs for conformal coating and masking, arriving at figures between £4.00 to £6.00 per assembly depending on coating method.
His next topic was PCB contamination testing options, and he reviewed commonly-used techniques and documented standards. Visual inspection could give valuable information, especially if UV tracers were present in the flux, particularly in identifying possible traps for cleaning and drying residues. As a process control tool, ROSE testing was frequently used for cleanliness monitoring to IPC, IEC, and MIL specs, using 50/50 or 75/25 IPA/water to extract ionic material then measuring the resistivity of the extract.
Meaningful measurement of contamination on a board required consideration of the total time and the rate at which the contamination rose over time. Ionic contamination could originate from soldering materials, the printed board surfaces, fabrication materials, and other sources. Although many people were happy with a result less than the specification limit, much more useful information could be gained by observing the shape of the graphs, and Willis showed examples of results from different PCB solderable finishes and soldering methods. For actual surface analysis or comparison of residues with known process materials he recommended having infra-red spectrometry or ion chromatography carried out by a specialist laboratory.
SIR testing could be carried out on bare PCB coupons, or coupons with components, and could be used at any stage of the process. But he recommended that simpler tests be done first to get the process as well under control as possible before committing to the time and expense of SIR measurement.
His third session was a review of test methods for conformal coating reliability, with reference to IPC-CC-830B; Qualification and Performance of Electrical Insulation Compound for PCBA, IPC-HDBK-830A: Guidelines for Design, Selection and Application of Conformal Coating, and IPC-TM-650: Test Methods Manual.
The conformal coating and cleaning experience concluded with a discussion of conformal coatings defects, causes, and cures. Stressing again the importance of starting with clean surfaces before contemplating any conformal coating process, Willis reviewed the assembly operation with regard to the compatibility of PCB and component finishes, fluxes, soldering conditions, and cleaning chemistry, then gave examples of coating and cleaning defects.
Minor de-wetting or bubbles on non-critical areas such as substrate or package surfaces should not be a reason for rejection, but the occurrence of such defects between terminations could have a serious impact on reliability if potentially corrosive materials were not fully encapsulated. Electrochemical reliability problems were caused either by simple metallic corrosion of solder, PCB conductor or component termination by contaminants from PCB production processes, flux residues or harsh environments, or by electrochemical effects in the presence of bias voltage. The most important failures of this type were dendrite formation and conductive anodic filamentation (CAF) and Willis showed several examples from the NPL Process Defect Database.
The NPL/SMART Group Conformal Coating & Cleaning Experience delivered a wealth of free practical help and advice to engineers wishing to gain an appreciation of the realities of coating, cleaning, and testing techniques and guidance on material selection to suit particular service conditions. Seven delegates diligently attended every single session, and were additionally rewarded for their attendance with packs of IPC standards, presented by organiser Bob Willis.