Frying the Flight Data Recorder: How Not to Impress


In this month’s In Compliance Magazine the focus is on military and aerospace topics. Along those lines, we worked on a project for an aviation company some years back. In mining the archives, so to speak, we head to the lower peninsula of Michigan, to the city of Grand Rapids.

Grand Rapids is the birthplace of Gerald Ford and the home of the Gerald Ford Presidential Library & Museum, where the late President was laid to rest. Ford served a partial term on the bully pulpit after a long run in Congress (1949-1973). He succeeded Spiro Agnew as Veep under Nixon after Agnew resigned, in disgrace, in December 1973.

Those were crazy days in politics and Nixon’s presidency crumbled nine months later under the weight of the Watergate scandal. It was a turbulent and challenging time, indeed, what with the imminent collapse of Saigon, the oil embargo, the threat of thermonuclear war, pastel leisure suit sightings and the woeful growth of disco music.

With all of that on his plate, I wonder if Mr. Ford, sometimes, wished he had stayed in Grand Rapids to practice something more peaceable than politics, like divorce law. Now, Michigan has a “no-fault divorce” provision on the books which, I suppose, makes it a bit easier to get un-hitched. One day, a few years back, I wished there was protection for “no-fault consulting.”

More on that later, but first a bit on Grand Rapids. Native Americans have inhabited the area around the Grand River for millennia. Westerners, first French fur trappers and later woodsmen, started arriving in the late 1700s. Foundations of the city’s industrial base were developed one hundred years later as craftsmen and entrepreneurs built “The Furniture City.” To this day, she hosts several high-end furniture manufacturers. Grand Rapids is also home to a wide variety of technology industries, including machinery, automotive and aerospace.

Where once rapids roared and riffled on the Grand River, dams, power canals and tailraces have subdued her natural flow, now largely serene as it meanders through the town. In the spirit of post-environmental awareness, there are movements in the city to bring back the natural rapids to Grand Rapids.

Motu Viget means ‘strength in activity,’ which is her motto and is a paean that harks even back two thousand years to Native American Mound Builders, who stayed active constructing impressive edifices of dirt and clay to bury their dead.

As the town grew in the mid 1800s and accepted immigrants from Holland, Germany, Ireland, Sweden, Italians and Poland, the city’s location in the industrial heartland of America lends access to major markets and a diverse salt-of-the-Earth populace that boasts a strong work ethic, typical of the solid mid-western towns in the US. Later, refugees from war and foreign strife came to settle in the town. After the pogroms of the 1930s and World War II, European Jews found safety and peace in Grand Rapids. Hungarians settled after 1956 and Vietnamese after the fall of Saigon.

A place of activity—and strength—indeed.

The economy of Grand Rapids is diverse, too, with not as much vertical reliance on a single industry as, say, Motor City, a hundred and sixty miles to the East, with the avionics industry supplying sophisticated systems, to where this story arcs. One project that we were called to investigate was a Flight Data Recorder (FDR). These so-called “black boxes” (actually painted orange or yellow for better visibility at a crash site) have been standard in aircraft beginning in the 1950s. The earliest versions were analog and relied on a metal foil in a crash-survivable case.

The initial development of electronic FDRs for the Air Force (for the F-16) began in 1982, starting a continuing evolution of solid-state based digital devices. With no moving parts, they were much better suited to survive the tremendous impact and related injuries associated with a high-speed nose into terra firma.

The designs of FDRs have tracked heavily integrated with flight systems; very rapid advances have been occurring since the 1980s. This is especially true as more aircraft are developed with fly-by-wire controls, which allow the monitoring and recordation of dozens of sensor signals. The FDR systems consist of not just a single box, but of multiple, linked modules including the Signal Acquisition Unit, Cockpit Control Unit, Data Acquisition and Recording Unit and the Voice and Data Recorder. By the late 1990s, these systems were growing in capacity and capability. Memory units with up to several hundred megabytes meant that more parameters could be stored, for longer periods of time, giving incident investigation loads of data for analysis after crashes. This means the integration and density of electronics has increased tremendously, along with rise of clock and data rates, and EMI.

Versions of these data recorders are now being deployed in railroad applications. Following a fatal crash in Chatsworth, CA in September 2008, the National Transportation Safety Board (NTSB) made the following recommendation to the Federal Railroad Administration to “Require the installation of crash- and fire-protected inward- and outward-facing audio and image recorders capable of providing recordings to verify that train crew actions are in accordance with rules and procedures that are essential to safety as well as train operating conditions.”

Passenger vehicle Event Data Recorders, monitoring speed, brake condition, seatbelt status and cruise control settings are now common in motor vehicles and have been used to prove assertions of reckless driving in court.

Our client in Grand Rapids was developing the Signal Acquisition Unit, a critical part of the FDR’s distributed component of the avionics system connected via a serial 1553 bus. The unit we worked on, a gray squat rugged aluminum box, had four connectors on its face. These were dense MIL-style circular multi-pin affairs that connected to harnesses made of dozens of individual wires. The device was failing RE102 emissions in the 30-60 MHz band. A quick look at the signals showed a mix of narrowband clock-like emissions and broadband data-generated noise: ugly stuff.

We were set up in a small shielded enclosure. A biconical antenna was set up and connected to a spectrum analyzer that was sitting on a cart next to the bench. The wiring connected to the EUT was splayed out on the bench. The analyzer display bloomed with green spikes, picking up energy radiating off the harness wiring.

Because of weight issues, shielding the wires was out of the question, not to mention the complexity of making a complete shield with routing for branches and to connect to various locations on the airframe.

So, what to do? Well, these kinds of situations are really the nasty ones: few options, a complex box, tons of wiring and any number of sources, any one of them could put the unit out of compliance.

We tried clip-on ferrites inside, squeezing them over the internal wiring. No luck. The thing about ferrites is: either they work, or they don’t. Besides, the mechanical guys weren’t too thrilled about installing an unsecured hunk of sintered iron inside the device that had to survive multiple-G environments.

After a morning of fiddling around, checking schematics and looking at pinouts, it was time to break for lunch. Reality engineering sometimes means knowing when to put down the scope probe, but I didn’t. The question kept nagging at me: which of these pins had the most energy? Could we knock these signals down one-at-a-time? One of the tricks of the trade, so to speak, is to take a dental pick and carefully connect it to suspect pins, all the while looking at the spectrum display. If a particular node was “hot” the radiated energy off of the pick would peak frequencies on the analyzer.

One of the project technicians, a calm, older guy named Al who, as I recall, was a whiz at surface mount surgery, stood by as I poked the pinouts on C1, the largest of the four circulars.

“Be careful.” Al said.

The eye-hand coordination was better back then, but even so after a half-dozen pokes, my pick slipped and I crossed two pins. Something inside the unit popped and sizzled.

“Aw crap,” Al said and sighed. “There goes my afternoon.” I looked at him and he shrugged his shoulders. “Might as well go to lunch now.”

I put the dental pick down and followed him out of the room down the hall a few paces behind, head lowered and just a little shaken.

Al turned into the room where the engineers’ cubicles were arranged, tapped the lead engineer on the shoulder and told him, calmly. “The consultant fried the unit.”

A nanosecond later, I walked into the room to hear the lead, jumping up from his desk yelp: “What?! Kill Him!”

Lunch that day at the company cafeteria was quiet, I sat a few seats away from the project manager, trying to stay cool. He was picking at a plate of lukewarm lasagna, muttering something about the schedule getting ‘shot in the ass.’ Oops.

Back at the ‘shop’ I debriefed John, my supervisor, somewhat sheepishly. John was calm. “Those things happen. I blew out the front-end of a forty grand receiver when I was younger. It’s reality.”

They ultimately got the unit working again (it turns out that I blew out a simple line driver; thankfully back then DIP packages were pretty easy to de-solder and replace). Al did a fine job, but it did take the rest of the afternoon.

Ultimately, we found some filter-pins that were retrofitted to the circular connectors as a quick-fix and sold them an EMC design course. favicon



author_violette-mike Mike Violette
is President of Washington Labs and Director of American Certification Body. For reasons that have nothing to do with frying that flight data recorder, he has not been back to Grand Rapids. He can be reached at