Product Focus: Evolution of DFDAUs

Avionics evolve. And in true Darwinian fashion, they often can become extinct–at least as hardware. That probably will be the case with flight data acquisition units (FDAUs). FAA’s notice of proposed rulemaking (NPRM) for flight data recorders (FDRs), into which FDAUs feed data pertinent to accident investigations, is furthering the technology’s evolution. However, data acquisition incorporated in the integrated avionics of new aircraft designs, such as the B787 and A380, will accelerate the process, perhaps making the traditional FDAU a dinosaur.

The NPRM, which also covers cockpit voice recorders, requires that the sample rate for up to 10 key parameters–such as flaps, aileron and rudder inputs and corresponding surface movements–be increased to 16 samples per second (s/s). That’s no problem for today’s digital FDAUs (DFDAUs) gathering the samples; they can receive data inputs via interfaces to as many as 48 ARINC 429 data buses at much faster rates. However, the problem arises from the inability of sensors and other data providers to deliver information to the DFDAU at the 16-s/s rate.

A system such as the flight control computer, which delivers pilot inputs to DFDAU in some aircraft, may have to be modified in design. "And that’s no trivial task," says Dan Martinec, with ARINC Industry Activities. Comments to the NPRM (still being submitted at press time) no doubt will negate the need for, or will suggest alternatives to the 16-s/s rate for all of the designated parameters. Honeywell advocates a "performance-based" sample rate instead of a uniform 16-s/s data acquisition rate. Some parameters may not require sampling at 16 s/s, while others should maybe have an even faster rate, says Duncan Schofield, Honeywell’s principal engineer. He cites as an example the rudder movement at 39 degrees per second (according to a National Transportation Safety Board report) on the American Airlines A300 that crashed off Long Island in November 2001. "If that rate really is true, then 16 samples per second may not be fast enough," he comments. "That’s why we think performance-based may be better."

Nevertheless, today’s DFDAUs have the processing power to accommodate the digital inputs; the challenge may be the DFDAU’s data output rate to the FDR. As yet that rate has not been fully established, but some DFDAU suppliers believe it would have to be at least 512 s/s and perhaps as high as 1,024 s/s. This means that older DFDAUs in spares inventories–many with no more than a 256-s/s output rate–won’t work in new aircraft, for which the NPRM applies. (The NPRM doesn’t require the retrofit of FDRs or DFDAUs.)

ARINC 767

Because units with a 1,024-s/s output rate have the same architecture as those with slower rates, they still will meet the ARINC 717 specs for DFDAUs, according to Armen Nahapetian, Teledyne Controls’ vice president of engineering. (The Airline Electronic Engineering Committee [AEEC] recently completed version 11 of ARINC 717.)

Now AEEC is trying to define the DFDAU not as hardware but as software embedded in an integrated avionics suite. "Instead of defining form, fit and function, we’re just defining function as the standard," says Martinec.

"The traditional FDAU that reads analog inputs and makes conversions [to digital form]–that’s gone," exclaims Stylian Cocalides, vice president of Avionica. "Now the data [to the FDR] is digital, and we’ve pushed the analog conversion to the subassemblies computers, such as the flaps management computer." These DFDAUFs (F for function) already reside in integrated avionics cabinets installed in Bombardier CRJs and the Embraer ARJ 135 and 145. The Boeing 777 has the DFDAUF in its aircraft information management system (AIMS).

The benefits of these "virtual DFDAUs," as they are often called, is that they save weight because they require less hardware and they improve dispatchability and the aircraft’s reliability, according to Nahapetian.

AEEC is working on several new standards that will impact DFDAUs. One is ARINC 767, which applies to Ethernet-based communications between the DFDAU and FDR. With Ethernet switching, says Nahapetian, multiple sources of data can provide information to a single receiver, and this could eliminate the DFDAU’s function of multiplexing the different parameters into a single data stream.

With ARINC 767, data acquisition and processing would not be concentrated in a FDAU but rather distributed via a backbone bus throughout an integrated avionics system. And data acquisition would no longer be time-dependent. Which means the data acquisition rates could be performance-based, as advocated by Honeywell.

ARINC 767’s specs also would allow for the inclusion of a recording map–information on how to translate the gathered data–within the recorder, instead of requiring documentation from the DFDAU manufacturer for interpretation. ARINC 767 is expected to be completed by the end of this year.

Another new standard, ARINC 647, would provide a common language for information delivered by the DFDAU and stored in the FDR. "This will make data from different aircraft compatible," Cocalides explains. "The recorded data can be interpreted, typed out and read in plain English. It will be in an ASCII text form, which means anyone with an Excel program can read it."

Companies

AIM www.aim-online.com

Avionica Inc. www.avionica.com

Curtiss-Wright Controls Inc. www.cwcontrols.com

Fluke Corp. www.fluke.com

Geotest-Marvin Test Systems Inc. www.geotestinc.com

Honeywell www.honeywell.com

L-3 Communications www.l-3com.com

Laversab Inc. www.laversab.com

Merlin Engineering www.merlineng.com

Moritz Aerospace Inc. www.moritzaero.com

North Atlantic Industries Inc. www.naii.com