Engineer Spotlight: Open Source Signal Integrity Engineering with Davi Correia and David Banas
The future of electrical engineering, the freedom of giving code away for free, and the importance of professional curiosity: an interview with Signal Integrity Engineers David Banas and Davi Correia.
Signal Integrity Engineers David Banas and Davi Correia share their thoughts on the future of electrical engineering, the freedom of giving code away for free, and the importance of professional curiosity.
The electrical engineering community is a tightknit group full of cooperation and support, and two great examples of that are David Banas and Davi Corriea. They've achieved a high level of success in their careers, but they also give back to their fellow engineers. Banas, of Haskware and eASIC, is the developer of PyBERT, an open-source serial communication link bit error rate tester, and Correia, of Carlisle Interconnect Technologies, developed many tools to automate parts of the product design process, increasing efficiency for his company and the EE community at large.
The two teamed up with Jason Ellison from Amphenol to deliver a bootcamp at DesignCon 2019 on The Art of Signal Integrity Analysis, where they led participants through the use of their free source code. For those who couldn't attend, you can watch a video of the session. They have also freely distributed their code on DropBox.
AAC's Mark Hughes had a chance to talk with Banas and Correia after their DesignCon session to learn more about their careers and perspectives on the EE profession and its community.
Davi Correia (l) and David Banas (r) at DesignCon 2019. Image courtesy Mark Hughes.
Mark Hughes (AAC): How did you guys become electrical engineers?
David Banas (DB): I got hooked on electronics when I was about 11. I wandered into the local Radio Shack store in my hometown and I saw these gizmos and books about how to build things with these gizmos, and I was hooked. I started buying the little kits they sold, taking them home and trying to build them and make them work. And they never did, but it didn't matter. I was a junkie. I would go back and buy the next one and build it, and finally about a year in they started working, and boy was that magic.
Davi Correia (DC): I started off school as a math major. I wanted to do math. Physics, a little, yes, but math. Two years into college I said, "Eh, not exactly," but I didn't know exactly what I would do. But then one of the classes that I took before was electromagnetics, and that was the only class that I liked that had a lot of advanced math. And then I decided to do something that would be just electromagnetics. So actually I came for electrical engineering from electromagnetics.
Electromatic wave diagram. Image courtesy Davi Correia.
AAC: So what excites you or what do you draw enjoyment from in your careers at this stage?
DC: There are two things that I really like. One is the design, is the challenge, is the days you're pushing and the thing hasn't been solved yet or there is no technical solutions. These designs that don't exist and you have to come up with.
DB: Yeah. I love when elegance finds its way into something like this and then proves to be the correct technical solution as well. That is so beautiful when that happens.
DC: The second one that I really like, is taking something that is perceived as advanced or, "Oh man, this is almost impossible to understand." And then you realize, "Oh, it's not. It's actually accessible." If people spend some time explaining, it's not a big deal. One thing is the design, the research, the new design. The other one is teaching. "Hey guys, it is that simple." Those are the two things that'll say, "Made my day!"
DB: I love pulling the curtain aside on these self-proclaimed wizards that are trying so hard to keep everybody else from learning what they know and to be a part of a free tool flow like this. It's exactly what that effort is focused at and I just love being a part of that, because I want people to learn and understand that these things they're doing are not black magic. There's math underneath and there's science underneath. And that's all! The math and the science work, if you just are able to learn it.
AAC: What advice to you have for new electrical engineers, or those graduating soon with their B.S., as they enter the profession?
DC: I have one that is a recipe for disaster. It's when you stop being curious, when you start just turning knobs. Five years later, if you learn just to turn the knobs, you don't know what you're doing. Because somebody tells you, "Run this simulation. Send me the data." And if you have no idea what you're doing, that is very dangerous because five years down the road, you are very replaceable. If you don't open the box and look inside, your position in the marketplace goes down.
Yeah, there will be pressure. Yes, sometimes you have to click and deliver on time. But don't make that your everyday work. Spend some time learning and going deep. That is what's going to keep you afloat or else you're gonna be out of the game.
DB: I would say, if it's not too late, considers photonics instead of electronics. If they're popping out of college right now, seriously, give photonics consideration.
AAC: Why is that?
DB: Because if we can do it with light instead of electrons, we're going to. The power, and the cost, and the size are all in favor of light. So, if it works—and it looks like it may, at least in certain applications—it's gonna win. And if you're just popping out of college now, you wanna be signed on to the winning ship.
AAC: What are your thoughts on the educational approach to electrical engineering these days?
DB: I just don't think you should be awarded a B.S.E.E. unless you have made, designed, fabricated, and soldered parts on a circuit board and there are no cold joints. It's so cheap and easy to run a prototype board. There's so many houses out there available and there's free tools for doing the designs.
DC: I'm just thinking out loud, and I don't know if it's the case. But couldn't it be that they go through a path that does not involve hardware design? Because I see a lot of [focus on] software.
AAC: So they're choosing the path?
DC: I think now that what happens is, the software side is being overpopulated because people always [say], "Oh, today, the software engineers are making the big bucks. So, why don't I jump in?" Well, that was ten years ago, fifteen years ago. If you jump in now, I think fifteen years from now, you're gonna have a lot of those guys. And the market will probably stabilize by that time. It might be, just a guess, that in the future there are so few engineers going into hardware that those will be the next big bucks that we see.
AAC: What projects do you reflect on and say, "I did a heck of a good job here"?
Screenshot of PyBERT's GUI. Image courtesy David Banas.
DB: I'm extremely proud of PyBERT. I started it as a simple, experimental labor of love, because I wanted to play with serial link design and I couldn't because I didn't have access to one of these expensive tools. I knew from talking to a few gurus that it really wasn't anything more than applied linear systems theory. So, I started reading what I needed to read to write PyBERT and then I just started growing it and last year it was awarded a place in EDN's Top 100 Products of 2018. And I felt really good about that because all of the other placers were commercial products and PyBERT is free.
DC: This is the first time I'm releasing the source code for free. As you can tell, this was almost planned but it was not very well thought through, so to say. The code is not as clean as I would want it. But the same code is also available within the company. And I know the feedback that I got from all the engineers that were working with the codes was like, "Man, this saves me hours of work."
And part of it was identifying the bottleneck. This should take five seconds. It was taking an hour. Because they were loading with a program that was not meant for that. This should be more efficient, which resulted in these tools that we are making available.
But I don't have exactly one product that I would say, "This is the best."
AAC: So you two have joined with Jason [Ellison, from Amphenol] to create some pretty sophisticated software projects. You've decided just to give these away for free instead of attempting to monetize them. That has to be driven by some pretty strong feelings towards the electrical engineering community. Can you tell me a little bit more about how you came to create these, and just give them away?
DB: I've been a part of this community for almost 30 years. It's tribal almost, at this point. My folks raised me with a fairly strong sense of community service and I guess part of it is that. Part of it is totally selfish. Something like this is a wonderful personal marketing vehicle. It's a great way to get your name out in the community, make people aware of you.
And then, part of it is the joy of making a tool. I think the act of making the tool itself is very very enjoyable for me and it always has been. And when I get the feedback from other users and I find that the tool is working well, man, I can't tell you the degree of validation I feel when I hear that. When I get an email or a phone call from a PyBERT user and they're really happy with what it's doing for them, it's huge.
DC: For me, I was little bit lost on what to do with the tools. The reason why I developed them was first, for me to be more efficient. Then I started developing for all the engineers in my company and releasing them. I said, '"You can be more efficient as well. Here. Take it, take it, take it." So, it spreads out. But then, I moved to another company and developed more and more and then I was sitting on those tools. I was using them. I did consider, should I sell them? Maybe, but why would I take the risk? I'm not a software developer. I'm not a professional developer.
I want to have the joy to develop the tool. But if you're in the business of developing the tool, that joy goes down very very fast. I like developing, I like to check, and I like people to use it. But the fact that it's not a business, I don't have any pressure. Well, yeah, yeah there is a bug, I understand. And Dave can say, "Yeah, I thought of adding that. It's not there yet. And if you're not happy, sorry." This freedom when you have when you're not selling is great.
AAC: What's the plan to get this out to more people?
DB: Well, more of these joint endeavors I think. There's power in numbers. The impact of three of us teaming up together today is huge, probably, compared to anything either of us has done alone trying to make people aware of what's available.
DC: I must say, I didn't plan that far off. My plan was, "Let's together and do this. Do you have any idea what's gonna happen next?" Nope, nope. My code is not as clean as it should be. There are a lot of things that will probably change, but I don't want to wait for it to be perfect. It will probably help people now. So, put it out there now and then figure it out. That is my thought process. Dave's is way ahead. The code that he developed is way ahead in terms process and quality than the one I did. But so I said, "Well, at least we have something that is more developed than ours, but ours is good enough to get people started with some basic understanding." So for next year, I don't know, let's see.
AAC: Especially for these younger guys that are coming up, what do we need more of or less of in the electrical engineering community?
DB: Physical design skills are gonna take off. It's not gonna work to have E&M to be solely the realm of wizards anymore. Everybody's gonna have to have awareness of pretty solid E&M.
DC: That's a good one. My background was electromagnetics. I never thought of designing circuits or anything. The reason why I was first brought into the community was because I had a good understand of electromagnetics. It's happening everywhere. You're getting to the point that electromagnetics and more advanced type of things are the last class you take only because it's required as a senior, and sometimes you don't even take it. And you barely pass it.
Screenshot of video showing an electromagnetic field propagation from an RF connector to a microstrip board. Image courtesy Davi Correia.
This will not be sustainable for the devices that we are dealing with. You've got to understand that. And then what we see is, we are doing the training in-house. The reason I was brought to the company is because a lot of people needed that understanding because they took electromagnetics in college but they didn't know.
And here's where it ties up with the idea that if [you] use HFSS or CST, yes, you can click the button and you have something out. But if you don't have that deep understanding of the physics, electromagnetics, you're out of a job very easily. Because clicking a button and running a simulation, anyone can do.
I think it will come a time that it's gonna continue on splitting. Like engineering first split into civil and military. Then civil split into mechanical, civil and electrical. Electrical seems to be going in the direction of splitting into robotics, power, computer and telecommunications. It might come a point when telecommunications is its own major, because you have to go deeper into some of those things.
AAC: Is there anything else you'd like me to know or like to say to the engineering community?
DC: Come to DesignCon next year!
Read More
DesignCon 2019 Begins, Highlights Signal Integrity and Test and Measurement Needed for 5G
Great!