Featured Engineer

Interview with Joel Avrunin

Joel Avrunin

Joel Avrunin - Field Applications Engineer for Tektronix

How did you get into electronics/ engineering and when did you start?

Growing up as the son of an electrical engineer, I absorbed my initial interest from my father. We always knew the time in our house thanks to his Heath Kit clock. Our basement had a Heath Kit TV. He taught me the fundamentals of electricity and electronics, and helped me with many of my first solder-kits. I think I completely destroyed my Radio Shack 200-in-1 electronics kit by going off-manual, discovering how fun it was to blow up capacitors and make things smoke. I even taught an electronics design class at a computer camp in Boston. On the final day of camp, we used two forks soldered to a lamp cord and plugged into the wall to have a hot-dog cookout and light up a giant deli pickle. My experience at camp led me to realize that I enjoy teaching as much as I enjoy tinkering. The human side of engineering was always important to me.

What are your favorite hardware tools that you use?

It will likely be no surprise that I say oscilloscope, but honestly, that’s the reason I decided to work for Tektronix. I love technology that is personal, where people see beyond utility and make it part of their identity. Non-technical people can feel this attachment with their smart phones. A person is rarely attached to a particular brand of microwave oven, but he will swear his iPhone is better than an Android (and vice-versa). Engineers feel this association with their oscilloscopes, more so than with a frequency counter or power supply. There is no piece of equipment that makes engineers more comfortable. For any given measurement, there is often a piece of specialized equipment that does it best, but doesn’t let you see the waveform. Suppose you buy a precision frequency counter and the readings make no sense. I guarantee you grab an oscilloscope to figure out what’s wrong. People just believe results they can see.

I love hearing engineers tell the story of their first oscilloscope and how they came to possess such a treasure. It may involve a dumpster behind a local school, a surplus auction, or an extra special holiday present, but there is often a story well worth hearing.

You might be surprised to hear from me that there are features from every major oscilloscope vendor that I enjoy. Seeing the same fundamental information measured and presented in different ways is the essence of innovation in test and measurement. While Tektronix owns the majority of the oscilloscope market, products from other market vendors like Agilent, LeCroy, and Rohde often each have some cool little feature that distinguish themselves. It is what keeps us on our toes as the competition drives everybody to create a better product. My favorite oscilloscope right now is the Tektronix MDO4104-6, a 1GHz oscilloscope that includes 16 digital channels and a 6GHz spectrum analyzer all in one box. If I could have just 1 scope, that would be the one.

What are your favorite software tools that you use?

Ever since college I have been a fan of MATLAB. I am naturally hardware oriented, and can struggle with some of the abstraction required for more advanced programming. Without having to know about objects, classes, and inheritance, I can make the computer automate a task quickly using MATLAB. It is also easy to setup automation with Matlab Instrument Control Toolbox, and I can use everything from pre-written drivers to SCPI commands right from the command screen. I also make use of Visual Studio C# to write small applications to automate common oscilloscope tasks, especially because I can easily compile these programs and give them to anyone.

Do you miss leaving the design bench for life on the road as a Field Applications Engineer?

I’m having more fun now than I’ve ever had before. Most engineers have to plead with their managers to buy decent equipment, and I get the hottest technology as soon as it is available. I am also exposed to far more technologies than a typical bench engineer. When you are working on design, you have just a few technologies you must know. Supporting oscilloscope applications, I have to maintain a huge knowledge base to support my customers. As a result, I am exposed to technologies as varied as radar, satellite communications, consumer electronics, high speed serial standards like USB3.0 and PCI-Express, high voltage power systems, and memory technologies like DDR. I get to visit research scientists pushing the envelope of knowledge in fields like microwave spectroscopy and biomedical sensors. There’s likely no other job where I could be exposed to so many different types of technology. Plus, it’s fun to be on the road. One day I’ll be at NASA assisting on measurements for a future satellite or probe, and the next I’ll be wearing a hard hat and trudging through a coal mine to demonstrate a spectrum analyzer. I’m also the perpetual tourist, and find cool spots wherever I visit. Technology museums are my favorite, including visiting just about every air museum I pass.

I also get to teach and share my passion for test with other people. Many engineers are uncomfortable in front of a room of people. I thrive on it, and greatly enjoy the interaction and ability to share new knowledge. The worst thing that can happen when I speak is that everybody listens. I love when I get tough questions and the class becomes interactive.

What is the trickiest bug you have fixed?

I love hearing engineers tell their tales of tricky bugs. The entire company ground to a halt while an engineer fixed the problem. We tend to be fans of the fictional heroic engineers, such as Star Trek’s Montgomery Scott, Atlas Shrugged’s John Galt, or Iron Man’s Tony Stark. Sure, Captain Kirk always got the girl, but how long would it take to get to the planet Risa if the warp core was still offline?

I used to work at a company that made precision laser measurement equipment. There was a power supply connected to a large rotating laser assembly via a cable. One interesting challenge I had was that while the assembly rotated, something was emitting an ear-rattling grinding sound. First, I determined that the grinding was not coming from the rotating laser, but rather from the power supply on the floor. After disassembling the supply, I discovered the power line filter (inside a metal can mounted to the chassis) buzzed during operation, and the attachment to the aluminum chassis amplified the sound. It only happened during high current draw, hence why the motor turning caused the sound. I compared a good filter to a noisy one and discovered that while the cases looked the same, the country of origin stamp was different. I proceeded to use a dremel to saw open the case and grind away the potting from both devices. I discovered the construction had changed and the coil in the noisy one was rotated 90 degrees so it poked through the potting into the metal can casing. We got in touch with factory engineers who discovered that when they moved the assembly line, there was a manufacturing change that was causing our problem. I then spent the next day calling distributors to find one who had stock of the old part, and we bought enough to cover our needs until the supplier could fix it.

Do you have any tricks up your sleeve?

I always advise people when they layout circuit boards to think about probing access for later debug. Basic techniques for maintaining signal integrity are becoming well known, such as keeping paths short, reducing stubs, and preventing crosstalk, but very few engineers consider how they can check their boards later. A 12GHz probe with a 6” long ground lead is likely no better than a passive probe. Even though most high-speed boards do not include designed-in test points, that does not mean you cannot plan ahead. For instance, punch through extra ground vias near high-speed signals. Without a nearby ground, it is impossible to accurately probe. Make ground vias obvious for debug later (I worked at one company that made them square). On differential traces, clear back some solder-mask where you might want to probe. Also, components have height, so make sure that if you do intend to probe a small trace, you do not put a huge tantalum or connector right next to it. Also, look for parts that you may want to adjust and hand solder later, making sure there is clearance for you to work. If your company’s CAD system supports it, try to have a 3-D block model of your board generated before you go to fab, giving you a quick reality check of what it will look like. Also, people underestimate the way silkscreen can help you in debug. I love to write all sorts of notes to myself on the board. If I can read my own silkscreen instead of referring to the schematic, it saves me on debug time.

When you make an early run of a board, make a break-out board with test patterns. Many engineers do this today with impedance coupons, but if time allows, I like to copy and paste any complex microstrip filters, only I add an SMA to each end. I have even made impedance coupons with active circuitry on them so I could verify tricky circuits.

You may work at a company that has all of its boards sent out for fabrication, but learn to surface mount solder yourself. When your board doesn’t work, you’ll be the hero who can dead-bug install a missing IC with bridges of 0402 resistors and capacitors formed across the leads. If the component is RF, you may need to dremel down to the ground layer, so get a board layout print and a stereo microscope.

I was particularly proud of a trick I used when I designed a board with an infrared laser diode mounted on it. Powering up the system, we were discovering that there were lots of problems that could prevent the IR laser from engaging (digital problems, power supply droop, bad components, etc), but the only way to tell was with a current meter in series or a well-placed photodetector because IR light is invisible. Many times I wished I could just glance at the board and tell if the IR laser was active. So when I laid out the board, I placed a normal red LED in series with the infrared laser diode. I found it reassuring that with a quick glance I could see that current was flowing through that leg. It was removed in the final design, but it certainly helped in debug.

Any other tricks?

If you’re teaching people over lunch, you have about 40 minutes after they take their first bite of pizza before they will begin to nod off. If the room is too warm because you are running a 100GS oscilloscope and the lights are off because you are using a projector, they will likely be asleep in 20 minutes no matter how interesting you make the presentation

What is on your bookshelf?

My Kindle and my iPad! Seriously, I read so much (both books and application notes) that I have gone almost entirely electronic for both. Generally my reading is classic fiction or modern books on physics, cosmology, history, and economics. I am also studying for my MBA, so I spend a good amount of time with my head in textbooks.

What challenges do you foresee in our industry?

Time to market is critical today, and engineers are expected to do more with less. Part of my job as an applications engineer is to train customers on the new technologies and how to test them, because they just don’t have the time to become experts. A classic example is Superspeed USB (otherwise known as USB3.0). USB2.0 is ubiquitous and the test equipment requirements to ensure compliance are fairly straightforward. USB3.0 not only requires a more complex piece of test equipment than the USB2.0 solution, but also requires intimate knowledge of high-frequency design techniques. Departments outside of engineering may think it’s easy to change a requirements document to indicate USB3.0 instead of USB2.0, but that small change can easily drive tremendous complexity into the design cycle.

As a result, engineers are pushed to incorporate these new technologies faster today than ever before. I rarely visit a company and find underworked engineers. To some extent, this has created tremendous industry opportunity for niche designers. Years ago, an electrical engineer would design his own switching supply, and had to be an expert on it. Today he just buys a brick and solders it down. Even the board level designer becomes somewhat of a system integrator as it is impossible to be an expert in everything. While this may be the natural progression of more complex designs, it means that when things don’t work, often the lead design engineer doesn’t know where to start looking. Test and measurement applications engineers such as me often fill in the gap, becoming the expert in debug of multiple technologies that a typical engineer may not have had time yet to understand.

Previous Spotlights

x
Like free stuff? Enter Here!
EEWeb Weekly Giveaway Sponsored by Mouser This Week: 4D Systems uLCD-32PTU 3.2" Intelligent LCD Module w/Touch
Enter Here
Login and enter if you're already a member.
Click Here