Abstract

Location of intermittent faults in wiring systems is a severe maintenance burden.  “No fault found” conditions when the faults appear in flight but cannot be replicated on the ground are frustrating and expensive, and may result in the long term grounding of an aircraft or complete replacement of major electrical systems.  This briefing describes a method for locating intermittent electrical faults while the circuit is live.  The spread spectrum time domain reflectometry (SSTDR) method has been demonstrated on 400 Hz and simulated MilStd 1553 data bus signals with realistic aircraft loads including a landing gear light/motor, strobe light, and deck of fluorescent lights.  Development of a temporary in line sensor that can precisely locate faults while a circuit is powered (but the plane is on the ground) is underway.  Further development includes miniaturization to integrate the sensor into an arc fault circuit interrupter (AFCI) and other locations imbedded within the wiring system, and automated interpretation of results through integration with the wiring data base and maintenance manuals for improved maintainer support.

The Faults

“No fault found” conditions are commonly caused either by an avionics fault, an intermittent short circuit typically caused by vibrational contact or moisture ingress, or an intermittent high impedance (open) discontinuity caused by vibrational or temperature stress, corrosion or similar high impedance contact, poor or broken solder/crimp joints or splices, mismated, loose, or damaged connectors, loose or poor grounding, etc.  In order to create a detectable electrical fault, the condition must have asserted itself as an electrical discontinuity, causing something to go wrong during flight and prompt a maintenance request.  By the time the plane is safely on the ground, and the maintainer is on board, the electrical problem may no longer exist.  The wet arc that tripped a circuit breaker, for instance, is dry now.  The loose splice that opens up while the plane is banking left is connected enough for now.  The maintainer must replicate the problem, often several times, before being able to isolate and locate it.  Most often, expensive avionics are replaced multiple times before the problem is traced to a wiring fault.  

The technology described in this briefing is different than other available technologies in one major aspect.  It can be used on live circuits without interfering with the aircraft signals, and without the aircraft signals or associated noise interfering with the test signals.  The fact that this system can run live means that it can be there, testing, when the fault occurs, rather than trying to locate it afterwards when the electrical fault no longer exists.    

Spread Spectrum Time Domain Reflectometry (SSTDR) Technology

Time domain reflectometry (TDR_ has been used for decades to locate faults on aircraft and other wiring systems.  Although effective, this method is notoriously difficult to interpret, and it can only be used on unpowered wiring systems.  Time domain reflectometry cannot be used well on live wires, because it sends only a single voltage step onto the wire, and looks for the echo from this step function to determine where the end of the wire is.  Any noise on the wire (and there is a lot) would interfere with the TDR, corrupting the readings.  If the TDR signal was large enough to prevent this corruption, it would large enough to corrupt the aircraft signal.  

Spread spectrum time domain reflectometry uses a pseudo noise (PN) code, which is a long sequence of voltage steps (perhaps a thousand or more) in a defined pattern.  This PN code also creates echos on the wire, which can be read at the source end using automatic correlation hardware.  The PN code can be very small compared to the signal, and can be well below the noise that already exists on the line.  Thus, it does not interfere with the aircraft signal.  The noise from the aircraft does not interfere with the test signal, either, as the correlation on the long pattern of the PN code averages out the effect of the noise.  

A prototype test system has been built and demonstrated on live aircraft signals including 400 Hz power and simulated Milstd 1553 data bus signals. Aircraft loads added noise to the wire.  A strobe light, deck of fluorescent lights, and landing gear light with moving motor were all tested.  The test system was accurate to within a few inches over several hundred feet of cable. Intermittent faults were simulated using a wet arc drip test and a dry guillotine test with an inline AFCI.  The circuit located intermittent wet and dry arcs lasting less than 20 ms.  

For more in-depth technical information, please contact us: