Selective Soldering: Design, Process Challenges and Practical Solutions
Although SMT is today’s predominant electronics assembly technology and the proportion of through-hole components continues to decline, the need for some through-hole assembly will remain for many years. As component density and board complexity increase, hand-soldering can no longer be relied upon to give acceptable or repeatable results and selective soldering techniques are growing in popularity. What is the most reliable and cost effective solution to through hole soldering in tin-lead and lead-free environments? What are the design and process challenges and what are the practical solutions?
A SMART Group workshop at the Bromsgrove, UK, premises of equipment and process materials distributor APP Electronics, set out to provide the answers, with a programme combining technical presentations, live demonstrations and hands-on sessions, introduced and moderated by Nigel Burtt, senior electronics manufacturing engineer at Renishaw and SMART Technical Committee Chairman.
Burtt launched the proceedings with a general overview of selective soldering processes, focusing on multi-point dip-nozzle soldering and single point mini-wave nozzle soldering, and described two main classes of machine, where either a robot brought the PCB assembly to a stationary soldering nozzle, or the nozzle moved to a stationary PCB assembly on a conveyor, showing examples of each type.
In principle, the objective was to mimic wave-soldering, but in localised areas. Flux was applied selectively to specific component leads, the PCB was pre-heated to drive off solvents and activate the flux, and then the component leads were brought into contact with the solder wave. A continuous flow of nitrogen around the solder nozzle was essential in selective soldering, to keep the nozzle free from dross, and could be used as a means of additional local pre-heating. The type and size of soldering nozzle was chosen to suit components and PCBs. Single point mini-wave nozzles could be used for dip soldering single leads or drag soldering multiple leads; smaller sizes gave less heat transfer and larger sizes enabled higher speeds. Burtt showed a series of video illustrations of various actual soldering operations.
Solder pot temperatures were generally higher than would be used in wave soldering, placing more constraints on the performance of the flux and potentially more local thermal stress on components and PCB materials. Additionally, there was a greater tendency for copper to be dissolved from component leads and PCB pads, especially with lead-free alloys, and Burtt recommended reference to the Good Practice Guide MAT 26, available free-of-charge on the NPL website. Layout design rules needed to take into account the additional clearances required to accommodate selective soldering, and it was recommended to remove solder mask between pads to eliminate solder balls.
Success in selective soldering is critically dependent on choosing a suitable flux, and placing it exactly where it is required. This was the essence of the first of two presentations from Wim Schouten, regional sales manager for Vitronics Soltec. He explained that although a strong flux was required to clean oxidised metal surfaces and support wetting, there were potential risks to reliability if the right formulation was not chosen.
He advised against using the same flux as used for regular wave soldering because the conditions were significantly different: in a wave soldering application all of the flux came into contact with liquid solder, the high temperature of which de-activated the flux chemistry. In selective soldering not all flux made contact with solder, and residual activators could cause reliability problems. Reliability considerations aside, selective soldering temperatures tended to be significantly higher than in regular wave soldering, so the flux might not be strong enough to work effectively and result in soldering defects. Furthermore, wave soldering flux formulations contained surfactants to promote spreading to cover the complete solder side of the PCB. In selective soldering, excessive spreading of was undesirable although the flux needed to be capable of penetrating the barrels of plated-through holes. Alcohol-based fluxes were preferred to water-based formulations because rapid solvent evaporation resulted in less spreading and higher flux solids per unit area.
Not just the formulation of the flux, but the surface energy of the PCB became significant in the control of spreading of the flux and it was recommended to use solder masks with a lower surface energy for selective soldering applications. Schouten showed design-of-experiment results illustrating remarkable differences in flux spreading for different solder masks.
For precise flux application, dropjet dispensing techniques had been developed. Dropjet could suffer from satellite effects analogous to those experienced in inkjet technology, resulting in extraneous flux droplets. New high-frequency dropjets had recently been introduced, with internal pressure compensation to eliminate satellites, and these had increased the robustness of the fluxing process and increased yields.
Moving from conference room to demonstration shop, Wim Schouten’s team gave a comprehensive description and practical demonstration of selective soldering using the Vitronics Soltec ZEVAm machine and its off-line SmartTeach programming software as an example of a flexible entry-level system. Workshop delegates had plenty of opportunity to learn hands-on how to programme, set-up and operate the machine, which featured high frequency fluxing, inline pre-heating, a movable electromagnetic solder pot and automatic wave height and solder height measurement.
Returning to the conference room, Charles Cawthorne, electronics manufacturing technologist with MBDA Missile Systems, and SMART Technical Committee member gave a user-view of selective soldering implementation and process optimisation in a defence-manufacturing environment, sharing his experiences gained during the progressive transition from wave-soldered through-hole assembly technology in the early 1980s, where typical layer count was six, the finest component lead pitch was 2.5 mm and the maximum number of leads on a package was 16, through the mid-2000s, with a typical layer count of 24, surface-mounted components with lead pitch down to 0.5 mm and lead count per package as high as 728.
In the early days of surface mount technology, hand-soldering had been a practicable option for those components still in through-hole format. But as assemblies became increasingly dense and complex, hand-soldering no longer gave reliable or repeatable results, particularly in meeting the IPC-A-610 Class 3 barrel-fill requirement of 75%. Consequently, MBDA Missile Systems had installed a selective soldering system in the late 2000s. Cawthorne described the factors that had influenced the choice of selective over conventional wave soldering and commented on considerations of process set-up and optimisation, process yield and board design, illustrated with case histories.
In choosing their system, MBDA needed the flexibility to assemble PTH connectors on both sides of the PCB, as well as bottom terminated components and BGAs on both sides. They required the ability to apply topside heating during PTH soldering and to enable different solder dwell times on individual PTH joints to offset thermal sinking effects of ground and power planes. Obvious limitations of selective soldering were the impact on process time and the clearance requirements to accommodate selective nozzles.
Cawthorne discussed process set-up details, including fluxing, thermal profiling, nozzle drag rates, point soldering dwell times, nozzle size selection and the number of passes over particular components. Regarding space constraints and thermal relief on ground and power planes, he emphasised the importance of design-for-manufacture and involving the PCB fabricator and assembler as early as possible during the design phase. Flux choice and application technique were crucial to yield, as was accurate thermal characterisation of the PCB. For lead-free assembly, copper dissolution effects were a significant consideration and he recommended reading NPL Report MAT 26, as previously referred to by Nigel Burtt.
Returning to the subject of fluxes for selective soldering, Ross Bennett, global key account manager for Kester GmbH, described the development of a no-clean flux designed specifically for the needs of the selective soldering process. Echoing the earlier comments of Wim Schouten, he re-emphasised that the characteristics of a normal wave soldering flux were not generally suitable because of reliability issues with unspent flux and corrosive residues, poor connector barrel-fill, and the challenges of sustaining flux activity on thick boards during long periods in the soldering process and avoiding flux spread to adjacent components.
Kester’s development goals were to pass IPC SIR testing in the raw or partially activated state and to control spread and sustain activity over long dwell times to enable good barrel filling in challenging applications, together with a wide process window and no clogging of dropjet heads.
The main factors governing hole filling had been observed to be the quantity of flux deposited, the top side board temperature and the combination of solder dwell time and solder pot temperature. It was recommended that design-of-experiment be carried out to optimise the settings for a given assembly, taking into account variables such as board design, board thickness, components, and soldering equipment. Extensive development and testing had resulted in a zero-halogen, no-clean formulation that gave excellent soldering performance whilst minimising jet clogging and leaving no tacky residues.
Just as important as the attainment of acceptable hole filling was the avoidance of solder bridging between adjacent pins, particularly on fine-pitch components. In his second presentation, Wim Schouten covered this topic in detail, beginning by defining “fine pitch” by reference to the IPC road-map. In 2011, 2 mm was considered fine pitch. Currently it was 1.27 mm and trending to 1 mm by 2021.
He described three different options for mini-wave selective soldering: wettable nozzle, non-wettable nozzle and non-wettable nozzle with the work angled from the horizontal. Wettable nozzles were useful for soldering close to other components and, because solder flowed in all directions, no rotation was required. Solder bridging was also reduced, although lead protrusion length was a critical parameter. Non-wettable nozzles, with solder flowing in a single direction, could give more consistent results, but there was the consequent need for board or solder pot rotation. Bridging was avoided by a directed nitrogen stream heated to above the alloy melting point, and lead protrusion length was then less critical. Best results on pitches as fine as 1 mm were achieved by keeping the protrusion length as small as possible and tilting the work at an angle of 7–10 degrees.
For high volume work, short cycle times could be achieved by dip-soldering using a multiple-wave system, with dedicated tooling plates and laser-cut screens to avoid bridging. Soldering of fine-pitch connectors was feasible, but required higher pump speeds and some form of hold-down unit to prevent components from lifting. Care was needed if using a multiple-dip process, to avoid hairline bridging due to oxide buildup.
This SMART Group workshop very positively achieved its objectives: an excellent learning opportunity with a good balance between technical explanation and practical involvement, which generated plenty of interactive discussion and sharing of experiences from delegates and presenters. Nigel Burtt wrapped up the proceedings with thanks to delegates for their attention, to Vitronics Soltec, Kester, MBDA and Renishaw for providing expert speakers, and to APP Electronics for generously offering the use of their conference and demonstration facilities.