Peter BigelowThe HKPCA Show revealed just how far robotics for the PCB shop has come.

Industry exhibitions never disappoint, and during the last couple months of 2017 two of the biggest ones took place. Seeing the multitude of options that other parts of the world have in the way of equipment, materials and supplies is always staggering. I am always amazed how many suppliers of drill bits and drilling machines exist – matched only by the number of suppliers of via fill chemistries and paste.

Equally impressive is seeing what’s new. These days the really groundbreaking concepts, equipment and materials seem to be first launched anywhere but in North America, more often than not in Asia. While walking the aisles of the HKPCA Show in December and drooling over the multitude of opportunities to invest the capital dollars I wish I had, on equipment not available back home, I observed some interesting and definitely new equipment that was being described in a decidedly old way as robotics.

Now robotics in itself is not really a new category. As marketing people are known to sometimes do, however, they describe some of the robotics in adjectives that capture the imagination. One pitch I especially liked was for equipment “on the cusp of artificial intelligence.” Not to take anything away from the marketing folks, but I could describe much of my old equipment, and even older employees the same way: “on the cusp” of artificial, or some other type, of intelligence! The evolution of robotics has progressed to a point where there are some interesting and very useful capabilities that now seem, for the first time ever, in reach.

Historically, robots utilized in manufacturing have been relatively basic. Equipment such as loaders and unloaders – where the robotic task was to coordinate and time the loading and unloading of panels to a machine’s optimal feed and process rate – is an example of the kind of robotics most circuit board companies have successfully utilized over the years. Today’s robots, however, take those tasks to a new level, as they can grip and flip over panels quickly and locate even relatively tight openings flawlessly, enabling automation to be added to equipment that previously lacked it. Add to that the flexibility of being able to move the current generation of robotics from machine to machine – or task to task – quickly and easily, combined with far-easier-to-program “user-friendly” software that drives it, and the value proposition certainly has improved over previous generations of robotic shop floor automation equipment.

What about that artificial intelligence aspect of robotics some companies were hyping? Based on what I observed, their AI was simply an alluring way of saying that with the aid of cameras, the machine could locate and maneuver in tighter areas. Regrettably, today’s AI in robotics is strikingly similar to cars that can park themselves. It’s robotics, not intelligence, and while impressive, not really a game-changer.

But all is not lost. While robotics may not have reached the often-hyped utopian level of artificial intelligence, it has advanced dramatically – and at breathtaking speed – in a number of ways that offers promise and provides a glimpse of the shop floor of the near future.

The agility of most new robotic systems is a vast improvement over static machines of the past. Seeing machines that could turn with the same flexible movement of a human’s arm and hand but with precision – repetitive precision – opened the imagination to what might be possible. Current generation equipment is by no means a one-trick pony destined to be tied to one, and only one, piece of equipment. Instead, today’s robotics are truly flexible platforms that can be moved, reapplied and, like an employee, redeployed based on the task at hand.

While robotic loaders and unloaders of the past were controlled by static PLC-based controllers, today’s robots utilize software that is far easier to program and does not require an engineering background to manipulate. Some systems have even transitioned to app-based programming, making change-on-the-fly programming possible. The advances in programming that today’s robots utilize again makes them similar to retraining a worker.

With the advent of relatively low-cost and highly accurate camera-driven vision systems, robotics today can see what in the past only a sharp-eyed employee could. The ability to get an x-y-z view while also being able to clearly magnify an image to identify even the most minor imperfections is a radical improvement over robotic systems of the past, and is the most visible value differentiator today driving sales of robotics for the shop floor.

Now if the sum of all those parts could only be relied on to do some of the subjective tasks performed by humans that often create bottlenecks in production, cause human fatigue or demand delicate handling of fragile product, then we would truly have artificial intelligence!

Virtually every manufacturing facility – especially those where processes require knowledge and understanding, and the end-product can be intricate, requiring exacting capabilities – has tasks that could benefit if smart robots were available with human-like tactile sensitivity. These tasks are the tedious ones that take time, cause human fatigue, and often lead to the highest levels of rework and scrap in any facility. The cost of such failure is real; its significant reduction, if not elimination, would make robotics seem a bargain.

While such advanced sensing and artificially intelligent robots could be well-employed anywhere, they would seem most valued in North America, with its high labor costs and high-reliability product mix. Yet so often here the new arrives late. Walking the show floor, seeing the significant number of companies offering robotics, and the advances in flexibility and programmability, sent a strong message: Cost-effective and truly usable robots are not a matter of “if” but “when.”

Peter Bigelow is president and CEO of IMI (; His column appears monthly.

Submit to FacebookSubmit to Google PlusSubmit to TwitterSubmit to LinkedInPrint Article
Don't have an account yet? Register Now!

Sign in to your account