Issue link: https://iconnect007.uberflip.com/i/1114420
MAY 2019 I DESIGN007 MAGAZINE 33 Well, what good is that? All I'm going to do with the DFM is learn what problems I'm go- ing to have to solve on the manufacturing floor because it's already baked in concrete. I can't go back and say, "Stop the fabrication and start over again. Move that BGA 300 mils to the left so that it's not right on the edge of the board and won't need fixtures to run it down the line." I don't have that option. DFM is really just a preemptive alert for the manufacturing engineers to say, "This is the pile of crap that's heading in your direction, so figure out how to build it, and you can't change anything. Too bad." We have all of these rich CAD tools with lots of information about constraints, land siz- es, via holes, and annular rings, and then we reduce that rich data down to Gerber data, which is about as dumb as a post, or ODB++, which is more intelligent. Then, we give it to a fabricator and an as- sembly shop. In that whole process, I don't be- lieve the humans and business paradigms in- volved are going to evolve in such a way that this can be done collaboratively. Due to com- mercial and technical reasons and how busy everybody is, I can't make a designer under- stand everything a manufacturing engineer knows and vice versa. The only place left to go is the software. We can create a software para- digm where this stuff happens in parallel. We have to. Happy Holden: We've had that for 20 years in integrated circuit design. Arcuri: Exactly! The parallels to that are shock- ingly obvious. I'm in love with whoever said that (laughs). Nolan Johnson: That's a good one, Craig. Hap- py and I have IC backgrounds too. The PCB design tool flow seems to be 10–20 years be- hind where IC was. IC has been driven by huge monetary constraints to getting to silicon, which then forced the discipline that isn't pres- ent in PCB right now. Instead of having three shops and a $60,000 price tag for getting your first silicon, you're looking at a few hundred bucks at any one of 800 shops. Arcuri: That's a very good point. The capital re- quirements for an IC factory are grossly large. There are 400–500 medium to small CMs in North America alone competing for all of this business. It's such a fragmented market, and it's not driven by the turn time. The cost of spinning an IC is astronomical compared to, "If we have to relay out the board and fabricate another board and simplify it more, that's not going to kill us." We're preaching to the choir. Johnson: Let's take that complexity and add an- other layer because there are some trends go- ing on here. We have that sort of complexity going on plus new design constraints that are coming all the time. There are so many special- ized functions. You can't expect one designer to master them all, so how do the tools help them? Arcuri: I've always wanted to know so much about something that I could call myself a thought leader in a subject area without laugh- ing at myself, and I'm still looking for that thing. I am a pilot and am pretty good at fly- ing, but to the extent that I am old now and have a lot of experience. Globally, we are em- bracing more and more collaboration across all walks. With the gig economy and services like Uber where software applications connect a customer with a seller, there's a piece of tech- nology in the middle that facilitates that. Peo- ple use the Uber analogy a lot, but the idea is that the devices we have are much more pow- erful and portable. We are much more comfort- able, although it might be scary to many of us, with our data being someplace in the Cloud. Because when it's in the Cloud, it's a lot easi- er to create collaborative experiences because everybody can access it. Globally, we are embracing more and more collaboration across all walks.