PCB Design Workflows As We Know It Are Doomed Unless Our Data Can Save Us
I’m starting to notice a problem with how we work. Granted, most problems we solve. They’re design related ones, and that’s what we do every day, we’re problem solvers. But this issue that I’m starting to notice? This one I don’t think we can fix on our own. This problem has been around since the beginning, but I believe that we’ve been skillfully ignoring the elephant in the corner, hoping that eventually,
someone won’t point it out. But there’s no ignoring it now, the elephant is taking up the entire room, and there’s barely any room to move. So let’s just admit it guys, the way we design is broken, and we’re all doomed. That is unless our data can save us.
It Starts With the Way We Work
It’s hard to tell when this problem started exactly. What problem? ECAD/MCAD collaboration of course. You’ve heard the drill before, and you’ve likely experienced it yourself. Exchanging files through email, tracking changes through spreadsheets, and pushing and pulling back and forth with your mechanical designer just looking for some way to get the job done.
But this coming together of the specialists never truly happens. We’re raised as engineering specialists, taught to care only about the electrical, or only the mechanical. And so it’s evident that even our tools and workflows have followed suit. Your PCB design tool? It does 2D layout really well, spot on. And our colleague’s mechanical tool? He’s over there in the 3D world making shapes out of digital clay. All the while, there’s a disconnect in between.
This, my fellow engineers, is the Industrial Revolution of design still in action. Yes, we’re still living in it, and no, we haven’t moved on. Why? Because we don’t have the workflow to do so because we don’t have the tools to do so, and most importantly, because we don’t have the right data to do so, at least not yet.
Look around at most engineering departments around the world, and you’ll likely see the same thing. Fragmented parts of a design just getting passed down the line. From one black box to another, to another, and another, until finally, hopefully, it arrives as a finished product.
This is the world and way of living that every engineer knows to one degree or another, and I’m finally starting to realize that it is broken. Have you ever wondered:
- What are we supposed to do when every engineer is designing an IoT product (multi-discipline environment)?
- What are we supposed to do when every electrical engineer is designing 3D circuitry in a molded-interconnect design (MID)?
- And what are we supposed to do when the products we design blend electrical, mechanical, and software so closely that we can’t tell the difference?
- What are we supposed to do when the data from these product becomes integral to the uses of these devices?
- What are we supposed to do when the software is not just a tool but an actual design partner, providing us innovative engineering solutions?
I’ll tell you what we need to do; we need to focus on solving the root issues first. And one of those happens to be our data.
Data Here, Data There
Data, what exactly does that word even mean in the context of engineering? We hear about ‘The Cloud’, we hear about ‘Big Data,’ but for an electrical engineer, what exactly is it? One of the most primary data points we use it every day, let’s just call this PCB data a library. That little nugget of gold where every and every component is centrally stored, revisioned, sourced, and includes a 3D model, right?
This is one of the biggest problems with our data today. We have these libraries, and we depend on them so dearly for the success of our design, but so often they’re incomplete or missing. Take for example 3D models. How many of your parts have a linked 3D model? How do you fill the gap? Maybe you go scouring a manufacturer’s website for some free downloads. And if that doesn’t take care of the job, then your mechanical guy is going to be spending some serious hours making his own parts.
How are you even going to communicate with your mechanical designer about what models are for your board? Sure, that 0805 package makes sense to you by name alone, it’s a resistor, a capacitor. But for a mechanical designer, you might as well be speaking a foreign language. While you speak in package types, they speak in shapes and materials. It’s like apples to oranges.
What a Mess!
This process gets even more complicated. Now you might have your 3D model, and you know its origin point in your ECAD software. The centroid or Pin1 is usually the center of your component. But then you get that model over to the MCAD software, and now you’ve got a third axis (z), and things never quite line up between the two. Footprints need to be lined up with 3D models; origin points need to be aligned, it’s all a tedious process, all because the data didn’t line up right in the first place.
Are you starting to see the problem here? We’ve got all our data, and we’re just hacking it together. We’re supplementing it with manual processes, manually aligning origins, manually sourcing 3D models, manually communicating. It’s like you’re keeping that 3D model held together entirely with duct tape.
What You Really Want
As EEs, I think what we care about, or should care about, is not whether our designs are done and can be passed down the line, but whether they actually fit, and whether they actually work.
Data Done Right
Here’s the question, why even keep libraries separate between electrical and mechanical domains? What we need isn’t to separate, but to centralize. Imagine the foundation to a house that you can build anything in your imagination upon. So long as you have that common foundation, everything else lines up. You wouldn’t build a separate foundation for the walls and then another foundation for the roof. Your data (libraries) needs to be this singular foundation.
Because when you’re designing products these days, it’s not just about the electrical considerations, like whether it’s the right connection or the right value resistor. It’s also about product considerations like:
- Do you have accurate and realistic 3D models for every component?
- Based on the accurate models, will your board fit its enclosure?
- And most importantly, can this entire thing be assembled and powered on?
Isn’t that the end goal for both electrical and mechanical teams? At the end of the day, you want to know that you can clamp that enclosure together, bolt it tight, and power it on for years to come. You’re both after the same objective, so why does your data need to be separate?
What we need as engineers to solve this problem is a singular data model that is communicative and intelligent.
Think about that for a second, what would this mean for your design process?
- It would mean you aren’t relying on an email or phone call to know when changes have happened to your design.
- It would mean having changes occur in real-time, as you work because your electrical and mechanical workflows are connected.
- And most importantly, it would mean products can be designed in tandem, right rather passed down the line. True Co-Design!
When you have a stable foundation of data, then every engineer can connect around it. In the EEs perspective, that data might be a symbol, footprint, board layout, etc… And in the MEs perspective, that same data set might now be a 3D model, a set of parametrics, and modeling for simulation. But the data itself? That never changes, you just see your perspective in your editor or engineering medium.
So once we’ve got our data problem solved, then what’s next? When every engineer on a project is working from the same source of truth, what exactly will this mean for how we engineer? Here’s a few predictions:
- Farewell industrial revolution. We’ll be able to finally extract ourselves from our siloed and segmented engineering workflows. By unifying our data, we can finally stop designing circuit boards and enclosures, and finally, start to design connected products for our IoT future.
- Farewell file formats. As our engineering workflows come closer together, this might also might be the end of exchange file formats like STEP and IDF. And in line with that, you’ll likely be replacing emails and spreadsheets with real-time engineering collaboration tools.
- Hello, new problems to solve. Once we solve our data problems, then we can start to focus on the higher level issues. Like connecting engineering and business through our supply chains, quality assurance, and even our customers.
We’re Doomed, Data Save Us!
The way we design is completely doomed, and good riddance about that. We’re not just creating circuit boards anymore. We’re not just designing enclosures anymore. We’re not just designing software anymore. We’re designing products. Look around you; the facts are in the air. In 2020 there’s going to be more than 34 billion devices floating around that are connected to the internet, and you’re going to be designing them. Doesn’t that number terrify you right now, knowing how you currently design?
When that IoT train takes off, there’s no turning back. When we start to embed circuitry within enclosures, fabrics, plastics, etc. and your PCBs live in a three-dimensional world, there’s no turning back.
This is the New Age of Engineering, where the way we work is just as connected as our products. And to get to this reality, it all starts with our data. So please, ECAD industry, don’t give us another file format. We’ve got bigger problems to solve, and only our data can save us.
Want to be a part of the solution? Download EAGLE for free and start designing the future of electronics!