Joe Fjelstad, the nicest guy in the electronics industry, is used to hacking people off.
That’s because wherever he goes, Joe is a disrupter.
Not of people, mind you. Of technologies. And thinking.
Joe was talking flex circuits before flex circuits were cool. Now they are ubiquitous.
He conceived remarkable developments for chip-scale packaging, lead and ball attachment and, later, high-speed interconnection. A quick search of the US Patent Technology Office database reveals 118 inventions with his name on them. Even more ideas are in the pipeline.
As often as not, however, Joe moved on before he had the chance to reap the benefits (financial or peer acclamation) of his work. For instance, he was the first corporate fellow at Tessera Technologies, the inventor of the µBGA, but left before the company went public.
His latest rethink is the so-called Occam Process, which uses packaged ICs to build electronic assemblies in reverse order, thus eliminating the need for a separate PCB. The process calls for putting known-good IC packages and various discretes directly onto a carrier in predetermined locations and then encapsulating the components so they become a monolithic assembly. Then the assembly can be metallized with copper, and circuit patterns created to make the required interconnections between leads of all the various components. For multilayer designs, the process is infinitely repeatable, until all the connections are made.
You’ll notice nowhere in that description is solder mentioned. That’s because what Occam also does is eliminate the world’s oldest joining material, a move guaranteed to do the technical equivalent to kicking the proverbial hornets’ nests.
Occam debuted in the wake of RoHS and was marketed as an answer to lead-free solders. Yet I’ve often thought it was miscast in that role. Electronics aren’t about the connecting material; they are about size and reliability and cost and function.
A decade on, although still aghast at the estimated $90 billion that’s been spent converting the industry to lead-free, Joe seems to recognize this. He’s back to thinking structures first, instead of the other way around.
“At Tessera, we asked people to make new materials and processes in pursuit of a microvia CSP. This time we can use the existing infrastructure. It already exists. I’m hoping to create the will to pursue it.”
That’s key because, as the Semiconductor Industry Association noted in July with the release of its final roadmap, the shrinking of conventional silicon transistors is about out of steam. (See Around the World, pg. 11.) “Smaller” just can’t get much smaller.
In a paper published last spring, Joe revealed results of a study in which a 140 x 100mm 12-layer board featuring a 442-pin FPGA with a 0.8mm ball pitch was reduced, using Occam, to a 30 x 40mm six-layer board using a 0.5mm lead pitch LGA. The design rules didn’t change – 50µm lines and spaces – yet the new version is ~70% smaller in terms of total area and occupies a footprint of less than 20% of the original design, with minimal increase in assembly height.
If fewer pins could be achieved, simpler routing (and probably faster CAD) are next in line. Libraries could be simplified as well. Can Occam spur a return to building structures with the fewest number of transistors, not the most?
Joe thinks so. To light a fire, he is sponsoring a competition for designers, aptly named the Occam Prize. “Redesign anything you’ve done before, pretend all the components exist and set yourself to the task,” he says.
Joe, who will formally announce the Occam Prize this month at PCB West, relates the rethinking of electronics design to the pioneers of aviation. “Those early days were fraught with peril. But today, aviation is the safest way of transportation. If you want to fly, you have to leave the ground.”
Amid all these ever-shrinking products, Joe is always one to see the big picture.
“I’m not going to make a lot of friends in the solder (or component) industries, but the motivation is for the four billion people who make $2 a day and need products to last and be bulletproof. Those who adopt – and adapt – will be able to make product that is superior and less expensive than their competition.”
Designers, it’s time to fly.
P.S. See you at PCB West Sept. 13-15 in Santa Clara!