Remember when surface-mount was the 'next big thing?' Reminisces on the scary "new" technology.

"Why should I care about posterity? What has posterity ever done for me?"

--Groucho Marx

"You live and you learn. At any rate, you live."

--Douglas Adams

Contrary to the beliefs of my colleagues, I undertake the Herculean project of cleaning my office every now and then. The other day, it was snowing and cold as hell outside which I took as nature's way of telling me it was time (that, and not having seen the surface of my desk for several weeks). So I fired up the war-surplus flamethrower and set to work. I admit to being a bit of a pack-rat and there was no way I would part with my collection of conference proceedings that date back to the mid '80s, when I became active on the SMT rubber-chicken seminar circuit. These things are like virtual time capsules: browsing them proved both interesting and a source for this column. Let's journey back to those days of yesteryear ....

[ILLUSTRATION OMITTED]

It was 1985. One track (of 16) at the International Symposium on Microelectronics was devoted to "Surface-Mounted Component Technology." Bear in mind that what we now know as SMT resulted, for the most part, from the convergence of through-hole technology and hybrid microelectronics. Folks from the through-hole world understood epoxy-glass laminates and soldering as it pertained to components and assemblies. Those who emigrated from the hybrid side knew about things like screen-printing materials (e.g., conductive inks), furnaces and very small "stuff." So while the ISHM conference focused mostly on hybrid and thick-and thin-film processes and materials, hybrid packaging and processes, and ceramics, one short session was devoted to an emerging assembly realm. One paper discussed the application of "laser microsoldering to SMDs" using YAG and C[O.sub.2] laser equipment. Diode laser technology had not evolved and the laser equipment was very expensive. Also, since it was a point-to-point methodology, throughput was an issue. It was primarily for this reason that the author concluded that, despite a number of advantages, the technique would have a hard time against vapor-phase soldering, the reigning reflow technology of the day. But, hey, if you had to solder SMDs to PCBAs already bonded to heat sinks (as the author did), laser would be the only way to go. A paper by one of the three IR reflow machine manufacturers at the time presented the advantages of IR (in this case, short-wave and medium-wave) over vapor phase, particularly with respect to the control of the rate of heating and cost. Another presented a novel concept of using solder columns as a means of attaching an IC device to a PCB. Called a chip carrier mounting device (CCMD), the author suggested that it could be used for perimeter or possibly even grid arrays. Apparently one of the key material problems at the time was getting solder paste with good rheological attributes--in other words, one that would print and not slump as the melting temperature was approached. A new development in solder paste formulation to accomplish this was discussed in a paper by a pair of metallurgists from SCM Metals, Lee and Hwang.

Three months later, in January 1986, the IPC held its second Surface Mounting and Reflow Technology Conference. Then technical director Dieter Bergman announced a new document, IPC-SM-782, that would attempt to provide a land geometry to coincide with every standard SMT part. The granddaddy of workmanship standards for assemblies, IPC-610, was slated for a major updating. Other issues of the day included:

* SMT is difficult. SMT components have great complexity on sizes, shapes and terminal configuration. Terminal styles are a source of great confusion, including the j-bend, the gull-wing, gull-U, L-bend, V-bend, flatpack, butt and many others. But SMT is spreading fast--by 1990 we can expect that 30 to 40% of all components will be manufactured in a configuration suitable for surface mounting. However, some miniaturization will be lost to permit package standardization.

* In an effort to achieve zero-defect soldering, computer-controlled pick-and-place machines and IR or vapor phase reflow soldering have now replaced the tweezer and hot plate.

* Our company advocates maintaining a constant solder creme (paste) thickness of 0.10 to 0.14" across the board, regardless of what type or style component will be placed.

* The "isles of automation" concept was heavily promoted. Each machine center would be fed PCBs by magazines and unload into magazines. These magazines would then be transported to the next machine center. This is an alternative to attempting to integrate a pass-through conveyorized line.

* One of the benefits of SMT is reduced board complexity. The number of layers in multilayer boards can be reduced.

* The pick-and-place machine has to be flexible enough to handle a range of component types. Components can range from 0805s to about 1.25".

* Enough of a variety of placement machines, in terms of flexibility and speed, exists that the user can consider acquisition based upon volume requirements. Levels range from less than 2,000 placements/hour to greater than 50,000 placements/hour.

* The most common methods of applying adhesive with automatic equipment are: screen printing (generally used only in low-volume production and prototype fabrication); air pressure operated syringe (the most common method, and also capable of being controlled by microcomputer and dispensing many different drop sizes to suit the various devices being placed in one cycle); template system (most suitable for high production rates of unchanging units; rates of up to 50,000 cph are possible.).

* All good reasons for using surface-mount devices can be negated by a poor design. For example, if the design leads to a high rate of manual touchup, any gains seen through assembly automation are undone.

* With regard to solder creams, variations in furnace construction and PCB heating rates can increase or reduce the number of solder balls.

* In one session at least half the attendees had never seen a surface-mount assembly.

* Recent studies show that possibly up to 70% of electronics companies are, or will be, using wave-soldering in the production of surface-mounted assemblies.

* The value added cost to a soldered surface-mount assembly due to visual inspection can approach 50 to 75% of the assembly's labor cost. The visual inspection of surface-mount solder joints is time consuming, fatiguing, inefficient and at times, just impossible.

* Slightly more sophisticated COB (chip-on-board) packaging will incorporate bubble memories.

* The use of flip chips with controlled-collapse bonding has potential for very high assembly rates.

Some of these perspectives were prophetic. Others suffered a bit of erosion by the winds of change. As you reminisce, recall how much simpler times were. The smallest component was the 0805 and most ICs had 0.050" lead spacing (a few "fine pitch"--0.025" pitch--parts were starting to appear). Most of us used RMA fluxes and we had tri-chlor and other Freon-based solvents to clean with; the Montreal Protocol was yet to come. Only a handful of contract assembly houses were capable of doing SMT. And not a single presentation was interrupted by a cellphone.

The good old days. Remember, just as it was back then, we're all in this together.

Phil Zarrow is president and SMT process consultant with ITM Consulting (ITMConsulting.org); [e-mail].