Universal Synaptics  Doing the Math on NFF

This presentation will describe No Fault Found's (NFF) causes and effects,  examine the human and political factors involved, offer some metrics, and suggest a solution.

Click on Slide for a Larger View

	No Fault Found or NFF is not new. Strangely, no consensus exists as to it's causes or cures. Before ANY problem can be resolved it needs to be fully understood. To understand it, we need to define some terms so we can communicate.
	NFF has many aliases. One person's NFF is another's CND. Every organization has it's pet name for it and associated database to track it. Some break it down into sub-categories, confusing the issues even more. We need to get back to basics and call it what it really is: INTERMITTENCY . I2LM.
	The Military, Space, Airlines and others with their "systems in motion" seem to be getting hit the hardest by NFF. The equipment is more integrated and complex. The environmental abuses are worse. The margin for error is less. The NFF monetary and political impact is greater.
	In any well designed system, maintenance of that system is engineered into it, with some goal in mind. In aircraft systems the maintenance target is quick fixes. Diagnostics are geared to replacing suspect boxes quickly with boxes believed to be good.
	By the "Law of Probabilities", engineers saw that pulling the suspect box off the aircraft would usually fix the reported defects, since more system components are in the box than are left in the aircraft. However, they never calculated how aging factors would affect those probabilities.
	In the past few years, great strides have been made in electronic component reliability. Fewer components are being replaced. More connectivity type repairs such as soldering, cleaning or inspecting of wires, connectors, switches, sensors, etc. The failure mode of these connectivity elements is random intermittency that gets worse over time.
	It's this "Random" factor that causes diagnostic chaos. You "can't fix it, iff'n it ain't broken". Is there really a problem or not? Did the technician pull the right box? Was the problem really in the aircraft wiring? Did the Pilot really know how to use the system? "Lets just call it NFF, we can't spend forever testing it.
	After the spares pool is polluted with intermittent boxes, NFF rates rise rapidly. Operational Readiness suffers greatly. Avionics systems are replaced or upgraded. All new training and testing needed for these new systems. New aircraft replaces old. Cycle repeats.
	Engineers have tried every diagnostic ploy imaginable except one, Directly testing for the problem. Some don't believe that intermittencies are even a problem. Never considered the "randomness" factor. Mistakenly "hoped" that ATE and other testers could detect all intermittent events. They didn't do the math.
	In comparing depot repairs: In one-stop, the final test is the unit functioning properly. In multi-level, ATE determines if the unit is good. One-stop sees a lot of intermittence and little NFF. Multi-level sees little intermittence and a lot of NFF.
	In "doing the math" the LRU's with the highest connection counts, usually also have the highest NFF rates. Common sense tells us that if we have never properly tested the connectivity elements, then, over time, we may have created a niche for these latent failures to accumulate.
	We've looked at the many complex issues associated with NFF, now lets look for a way to simplify the solution.
	Intermittents can be classified into one of three types. The key to solving NFF problems is to know which type of intermittent is responsible. After you define the problem correctly, it's much easier to fix. Observe the clues:
	Aging factors: When a system is new, you get more Engineering and Test-void problems, until all the "bugs" get worked out. Environmental stress factors, over time, cause connectivity problems to become much more prevalent..
	Where and how? Engineering type intermittents are happening to all the units and usually during operation. Test voids are happening with a few units and always fail a test at the next higher level of testing due to failed components. Connectivity failures occur in a few units, during use, and can not be found at test.
	Virtually all of our test equipment is geared towards finding hard failures only. Technical, engineering and support processes are devoted to hard failures only. Technicians who actually see and understand the problem are given no support or equipment to fix the problem.
	Engineers design tests for new systems. Not familiar with connector science or aging problems. Mistakenly believe that ATE sees all intermittent problems. Don't understand the problem randomness plays in testing for intermittents.
	ATE can't "see" random intermittent problems due to: Fixed test windows.. Maybe the intermittent won't show up this week. Scanning: Even if the intermittency is active ATE will probably be measuring the wrong pin at the right time. Filtering: ATE tries to eliminate the noise. No stress is applied.
	Sample and Hold: Typical equipment for "Shake and Bake" Can have intermittencies up to tolerance level without reporting any. Mistakenly apply statistical science, sampling methods in belief something will show up if enough samples are taken.
	Think about this: To measure volts we use a VOLT-METER. To measure current we use an AMP-METER. To measure continuity we use a CONTINUITY tester. To measure random DIS-CONTINUITY we use everything but a DIS-CONTINUITY tester.
	Some continuity testers have a fast scan mode that sends digital signals down groups of wires to increase throughput. After doing the math, you may find that the detection capability is still too low. Attenuation problems also arise when TTL signals are pushed down long lines. TTL not effective for ohmic intermittencies.
	This chart shows graphically why the IFD-2000's neural network technology is superior overall in detecting intermittent failures. Note: Probability bar on IFD-2000 is actually 10 times larger than shown.
	IFD-2000 is easy to setup and use. The only programming involved is in telling it what to "SAY" when an event occurs. Inexpensive, common, D-50 connectors are used on the IFD-2000 interface. Simply add your connector to the other end.
	A versatile instrument: The IFD-2000 can be used by technicians as a diagnostic tool, on aircraft or LRUs. Periodic inspections for prognostics and trending. After trending, repairs or mods can be done without further testing. Produces print-outs and records to verify quality.
	Diagnostic Workstation: Enhance ATE testing. Numerous 3rd party hardware and software add-ons integrate easily. Generic Service numerous systems with the same equipment.
	Where best to apply the IFD-2000? Worst systems first. For any given failure, both the Aircraft and the LRU each contain nearly equal numbers of possible NFF failure sites. While SRUs may have just as many sites or more, they are thought to be more reliable due to soldering
	IFD-2000 Effectivity: 26 units were tested. 14 units had one or more intermittents. 12 units required recalibration only. ATE testing found intermittent problems in only 2 units. ATE NFF rate = 46% Simulator, after hours of testing found 8 units to have latent failures. SIM. NFF rate = 15%
	When "doing the math", to cost-justify intermittency testing, you have to look way beyond the repair shop budget to find the real savings. You need to also consider what the "worst case" risks are from not testing.
	The "father" of quality control outlined the testing philosophy that has guided the United States Navy and the world for decades.
	Considering the huge ATE costs, and that it is only finding 50% of the problems, then: If it was necessary and cost effective to invest in ATE, in the beginning, then investing now in intermittency testing is a bargain.
	You've looked at the problem. You've looked at the solution. Now, look at the costs, Consider the risks, And then,
	Go Figure! Transportable or
	Benchtop.