Carmen Parisi is an applications engineer who co-hosts an engineering podcast in his spare time. In this interview, he describes his work, shares some engineering tips, and tells us about a fun prank he played on an unsuspecting designer.
CIRCUIT CELLAR: Where are you located?
CARMEN: Currently, I’m living and working in Raleigh-Durham, NC, around the Research Triangle Park area between the two cities with my wife and new dog Sadie. Kelly and I moved down about three years ago from Buffalo, NY, and really like it here. There’s a lot of tech companies and engineers around, tons of stuff to do, and great food and beer scenes. Plus, as a hearty Northerner, I get to laugh at the “cold” winters we experience. Come summer, though, I melt into a puddle on the pavement. Snow all the way for me, but Kelly disagrees.
CIRCUIT CELLAR: When did you decide to pursue electrical engineering and why?
CARMEN: Ever since I was a kid I had a fascination with tools and how things worked. I would always have a toy sword and various tools stuffed into my belt and would volunteer to help my dad around the house building a deck around the pool or fixing the fence.
Once I got into high school, I took a few basic engineering courses during which time I got bit by the engineering bug. The course that really “doomed” me to a life of electronics was a Robotics course taught by my favorite teacher C, as we called him. He put me through my paces learning how to solder, reading schematics, programming in BASIC, and robbing Fort Knox using a LEGO Mindstorms robot. C’s class solidified my choice to go to college for engineering, and shortly thereafter, I picked electrical over mechanical for my major.
CIRCUIT CELLAR: When was the first time you used a microcontroller in a project?
CARMEN: If we’re counting LEGO Mindstorms, then the Robotics class in tenth grade with C where we had to build a robot to lift a golden brick and run away with it (thus “robbing Fort Knox”). I met all the individual milestones with my group for the project, but we couldn’t get the whole thing working smoothly from beginning to end. I guess that was my first time learning how to successfully fail too which has turned out to be a very useful skill.
My first real microcontroller experience was the summer after sophomore year when I took a college course at a local community college offering a few classes to high school students interested in engineering. During that course I learned more basic circuit theory, got introduced briefly to SMT soldering, and built some robots using the Parallax BOE Bot. Looking back, I’d say this was the time my analog career kicked off as I slowly started to realize that I was more interested in the circuits themselves than the overall robot.
CIRCUIT CELLAR: Tell us about your university-level schooling.
CARMEN: I still consider myself a student in that I’m always looking to learn new things and grow as an engineer, but my formal schooling is over for the foreseeable future. In 2011, I completed a combined BS/MS degree in Electrical Engineering at the Rochester Institute of Technology in Rochester, NY. I initially started off interested in robotics but after working with a great analog designer on my first co-op at GE, I switched into the analog circuit and semiconductor track and never looked back.
CIRCUIT CELLAR: Can you tell us about your work in graduate school?
CARMEN: Sure thing. My graduate work was primarily with the Communications professor who needed a proof of concept built to test out a theory that looked plausible on paper. Prior to my joining the Comms Lab, my advisor and two past grad students had worked out a method of securing wireless channels using the randomness of the channel itself. There was an initial front end of sorts to test the idea out but I don’t think it was ever tested.
I looked over the circuit design, decided to scrap it and start fresh, and immediately realized I had a big job ahead of me. Cue the analog professor becoming my co-advisor. Mixing circuits, active filters, phase detectors, ADCs, and communication theory swam through my head as I slowly cajoled the circuit to life. Two PCB revisions later the circuit worked in that it took the RF input signal and spat out some bits at the other end, but after my advisor applied his algorithm to the data, we weren’t able to generate symmetric keys on different boards. Whether this was from an error in theory or with my board I never found out, as I ended the project there to focus on my full-time job leaving with a grad paper instead of a full thesis.
I still have all my old lab notebooks, schematics, and board layouts on my bookshelf at home. I think the files are sitting on a hard drive somewhere too. Looking at them now, I can spot a lot of little errors I’d like to fix due to my inexperience at the time and some maybe a few not so little errors too.
CIRCUIT CELLAR: What did you do after school?
CARMEN: After I left RIT, I moved down here to Raleigh-Durham to start my career as an Applications Engineer working on switching regulators with Intersil. Back in 2009 I had done a summer stint as an FAE at a small field office in Long Island with the company which got me interested in working in the semiconductor industry.
Life on the road as an FAE didn’t appeal to me after spending my college years constantly moving around for co-ops, so my former boss set me up with an interview here at the RTP design center. On the way down for the interview, I got stuck in Dulles for the night thanks to some bad weather in Rochester causing me to miss my connection. I wound up getting a bare 3 hours of sleep that night on an empty terminal bench. The next morning, groggy and sleep deprived, I suited up in the family restroom and flew out for six wonderful hours of technical interviews. I was absolutely wiped out by the end of the day but managed to survive the ordeal. The rest is history.
CIRCUIT CELLAR: Tell us about the work you are doing as an applications engineer for Intersil.
CARMEN: Well, for starters, being an apps engineer is exactly the rock n’ roll lifestyle I’m sure all your readers expect it to be. I roll into the office every morning and have the roadies warm up my iron for me!
In reality though, I work on buck regulators for computing applications like notebooks, tablets, ultrabooks, with maybe a bit of desktop work from time to time. Most of the parts I work on are for the primary core voltage on Intel processors. Sometimes, should the part integrate multiple regulators, I’ll work on a graphics rail or one of the other many voltage rails present on a motherboard. For each new processor tock (tick? I always confuse the two), Intel releases a laundry list of specs that have to be met in order to provide power to their CPUs and my parts are designed to those specs.
When I work apps on a brand spanking new chip, I’ll first work with the design engineers to run some feasibility studies and help define any new features for the IC. These tests range from tuning a similar part to the new Intel specs to see if the control scheme hits any corners or has stability issues to beating up some power FETs to determine if they can handle the new current requirements we have to meet. Once the chip tapes out, I’ll start work on preliminary documentation—a rough datasheet draft or early reference design based on feasibility testing and simulations—for the field to use when working with customers. During this time, I also design the evaluation board I’ll use to validate the part and send to customers for sampling.
The real meat and potatoes of my job is silicon validation. I’ve got an exhaustive spreadsheet of bench tests to do that functionally verify the IC over a wide range of corners. The first few weeks after silicon comes back I’m working full throttle, round the clock if need be, to make sure there are no show stopping bugs we need to address. I never see my office during validation. Instead I’m spending all my time in the lab hunched over the eval board or squinting at my scope.
Things calm down slightly after the initial validation, but the work is still nowhere near done. Now I’m working with design and test engineers to debug any issues that crept up during validation and implement fixes. Ideally, a board-level change is found because PCB or apps level schematic changes are much easier and cheaper than silicon spins. In conjunction with this work, I’ll also refine my reference designs and documentation as well as work with the field on initial customer designs by answering questions and checking over layouts and schematics to make sure everything’s optimal for their builds.
Up until the part releases, I’ll continue cycling through validation, debug, and customer support as needed, squeezing in documentation when I get a chance too. At any given time, I’m also supporting old parts still in production or, if I’m in a lull with my work, getting pulled onto other chips to help out other apps engineers in a jam.
The last part I released, and my first as the lead apps guy, was the ISL95813, a single phase regulator for Haswell and Broadwell systems. My next part is scheduled for release next year which I can’t talk too much about, but it’s really cool.
CIRCUIT CELLAR: During your time at Intersil, you must have learned some important lessons about professional engineering. Can you share one or two things you took from the experience?
CARMEN: Most importantly, good communication skills are key. A large chunk of my job is talking to other engineers and customers across the country and overseas. Their whole interaction with me is through the emails and reports I send out and I want to make sure they’re top notch. You don’t need to be a poet laureate by any means, but if you come across like a rock head, it will be much harder to get taken seriously and problems will drag out longer than necessary. Proofread your work; make sure you’re getting your point across clearly; and tailor your email, report, PowerPoint, whatever, to your audience’s level of technical expertise. Study up on how to make a slideshow that won’t bore your audience or read a technical writing guide. It can’t hurt.
Secondly, document, document, document—even if it’s only for your own reference. And keep it somewhat organized so you can find what you need again without too much hassle. Yes, it can help CYA, but also I’ve saved myself a ton of time not redoing the same derivations or looking back at a difficult test setup I had documented in my notebooks. It’s especially nice being able to pull up old data from past parts to see why the heck we did what we did years later.
CIRCUIT CELLAR: Tell us about your most recent electrical engineering project. What did you build and why?
CARMEN: Well, I can’t talk too much about work since all my projects at the moment are either customer related or under development, but suffice it to say I’m working on a lot of low power, multi-role chips.
Outside of work though for nearly two years now I’ve been co-hosting a podcast which keeps me plenty busy. The show’s called The Engineering Commons and it gets released every other week by myself and three other engineers scattered across the US. It was originally started by Chris Gammell and Jeff Shelton, but when Chris left the show for other projects back in 2013, I threw my hat into the ring when Jeff put the word out he was looking for new co-hosts. We discuss the engineering discipline as a whole rather than focus on any one field and some of our favorite topics include education, the value of co-ops, life in the workplace, and the stories of other engineers we bring on to interview.
The semiconductor field is pretty niche, and so through the show, I get exposed to all sorts of new ideas and philosophies, whether it’s from researching a topic when coming up with show notes or hearing the stories of engineers and professors from across the globe. Some of my favorite episodes are the ones while interviewing a guest I barely have to say anything and not just because I hate hearing my voice when I re-listen to an episode! Hearing someone get really into a story and talk about their passion I can’t help but get drawn in and become excited myself. All us engineers are alike; no matter the field once you get us going about that tricky bug we finally tracked down, the ridiculous meeting that happened the other day, or those ah-ha moments when a solution just clicks in your head we just can’t help but gush and it makes for great content. I’ve put out nearly 50 episodes with Jeff, Adam, and Brian, and I can’t wait to do the next 50!
CIRCUIT CELLAR: Tell our readers about the prank circuit gag you pulled on the designer you worked with. And can you share an image of the prank circuit?
CARMEN: A good way through the 813 development I found some problems that ended up being non-issues because I misinterpreted a spec, had a test setup issue, or made a silly component choice in my design. The designer started ribbing me a bit by immediately calling everything a board issue from that point on. This kind of back and forth goes on all the time between apps and design and it’s always good natured in tone. I didn’t take it personally and took strides to be more thorough before ringing alarm bells going forward but I couldn’t let him get way Scot-free.
With my boss’ permission I waited until a slow day came along and rigged up a little circuit to the bottom of the eval board that would overdrive the compensation node of our regulator, propagate through the control loop, and cause seemingly random spikes in the output voltage. I took some waveforms and sent them off to the designer explaining how I found an operational corner that affected regulation we needed to address. Since he was a thorough designer and liked to regularly pop into the apps lab I actually spent my morning running the tests he asked me to just to keep up the illusion something was wrong if he showed up.
I kept him digging through the schematics trying to find his mistake until mid-afternoon before I brought him in the lab and slowly flipped the board over while telling him I found the error was caused by a parasitic circuit. At this point a couple other engineers who were in on the gag had found reasons to be in the lab for the reveal and we all had a good laugh. The designer took it pretty well, and I even bought him a beer for being a good sport.