Universal Synaptics TWA-800

Subject:  Inclusion of documents for TWA-800 investigation.

Dear Mr. Swaim,

Your request as I interpret it, consists of three parts;

1.  Permission to include the last document sent to Dr. Smith.  Permission is hereby granted and this document is included.  We have also included a previous written document concerning the direction his study is going in, looking for shorts and opens while excluding intermittents.  We feel the information in this document is just as important.  You have permission to publish it as well.

2.  You requested documents concerning the IFD-2000's capabilities and specifications.  These documents are part of our NFF booklet which dwells heavily on the NFF problem and explains why, where, and how our tester should be used, and what it will cost.  Please use any of these documents as you see fit.

3. In your last email you requested information on the pros and cons of non visual wire testing techniques vs. visual inspections.  The rest of this document will address that issue as does Walter Shawlee's reprint from Avionics magazine, and our recent letter to NASA which can be found inside the NFF booklet.
 

Visual Inspections vs Testing:

Actually, we have never considered testing for latent defects vs. visual inspections to be a pro or con debate.  To be truthful, I was taken back a little at your suggestion that one method might be considered at the EXCLUSION of the other.

From my years of testing and test program development for the Air Force, it was drilled in to every technician and engineer - that a complete and rigorous visual inspection would be accomplished, on each and every item before any testing was ever done.  This included all depot level functional, LRU and SRU tests, up through final installation into the aircraft.

The purpose of these visual inspections was two-fold.  First, if any components were shorted or burnt, it could cause serious damage to our million dollar test stands.  Second, if components are not assembled correctly, they are certain to cause problems later on.  Screws can come loose that hold components together. Frayed wire strands can break open or short to adjacent pins under vibrational stresses, etc.  It's a reliability issue.  No one would even begin to consider that a visual inspection would not be performed several times over the course of testing and repair.

Visual inspections are extremely important!  However, not all problems will be detected.  It is virtually impossible to see everything that needs to be examined.   The very first thing you are taught in any Quality SPC program is that inspections will not catch everything.  This was aptly demonstrated by the instructor having the entire class read a 5 or 6 sentence paragraph while counting the total number of "F"s.  No one gets the same answer... it's human nature to miss simple issues.

The biggest problem with visual inspections ONLY is that interconnection components often look good, but work intermittently.  Fretting wear and corrosion on connectors is impossible to inspect without a microscope.  Looking for the "green fuzz" of heavy corrosion is not even in the same class as looking for flux residue or other contaminants and oxidation films on connector pins.  You cannot see damaged or dirty wiper surfaces on potentiometers and engine controls, or the arcing inside a conduit that passes through fuel tanks or other areas. The results of these common degradation mechanisms MUST BE TESTED!

The terms "inspected" and "tested" are all too often used synonymously.  In many peoples mind's they refer to the same thing.  It would be much better if the term "inspection" was used only in the context of a visual examination.  Testing always means performing some rigorous examination by applying signals and measuring their performance against a set of standards.

The term inspection simply says "look at it".  The Air Force inspection standards should be adopted by the FAA, as the visual inspection standards to visually "test" against.

We have discussed at some length with the Airlines and FAA the need for both inspection and testing.  To do one without the other is simply asking for trouble.

When the risks are high, such as in the case of national defense or in the case of accidents caused by equipment malfunctions, there is plenty of reason to do both visual inspections and rigorous testing.  Neither should be sacrificed for the other.

If you want some comparisons of the relative merits of "inspection" vs. testing, again look at the Air Force's numbers. They perform rigorous visual inspections, yet have a NFF or CND rate as high as the airlines, which runs as high as fifty percent or more on older systems.  They, like the airlines, don't do intermittency testing.

They do have wire testing equipment however.  For ringing out an aircraft they have a huge tester mounted on a semi-trailer that they pull around to the airplanes when necessary such as after extensive electronic or avionics modifications.  They also have top-of-the-line, expensive wire testers and complete adapter sets and programmed continuity tests for ringing out each LRU chassis.

They seldom use the aircraft wire tester because it takes considerable time to ring out the entire aircraft and in the case of the LRU wire tester it's almost never used because they very rarely find any problems at all.  The biggest complaint against each tester is that they are especially incapable of finding intermittent type problems.

As you can see, rigorous testing must include Intermittent Fault Detection (dis-continuity testing)!   This type of testing must not be confused with other wire tests measuring continuity and insulation breakdown parameters.  Continuity testing and dis-continuity testing are two completely different sciences and require diversely different technologies...  This confusion in testing is the KEY to the whole problem!

In your efforts to compare visual inspections to wire testing, beware that there are two different things that need to be tested.  First you need to measure the connectivity of the wires which includes testing the insulation properties and other components, to make sure they meet engineering specifications. Do they attach where they are supposed to be attached and only there?  Is there any unwanted resistance in the lines?  This takes the kind of equipment presently in use.  You need a "continuity tester" for this purpose.

The NFF problem is caused by intermittent connectivity brought about by wear and corrosion in the connectivity elements.  Due to the randomness of this failure it cannot be detected efficiently using standard "wire testers" such as continuity testers or ATE.  You need a DIS-CONTINUITY tester to do this job.

The purpose of our IFD-2000 is to look for intermittent breaks in continuity.  The other test equipment simply cannot see these short duration, low level breaks or "indicators" of failures about to happen.  Their measurement "mechanics" are on a serial oriented, fixed-timing schedule and cannot sync up with these randomly occurring events.

The IFD-2000 monitors all lines, all the time, while continuity testers look at one single line at a time, and even then they are insensitive to short duration or "ohmic" breaks.

We have been saying for some time, that the problem the Airlines, FAA, NASA, and Defense Department are having with all the rampant unexpected system failures, is that they are skipping steps in the testing process.

There are three testing steps!

1.  Functional tests insure that the system works as designed.

2.  Visual inspections, insure that defects that cannot be tested for can be eliminated before they cause failures in the future.

3.  Reliability or intermittency tests insure that the age related defects, that cannot be visually inspected, can be eliminated before they also cause failures in the future.
 

Every organization is performing functional testing.  Some are doing visual inspections.... at least to some, limited degree. None of the aforementioned organizations are doing reliability testing.... It's the missing link!!

In your first email message you said that you were compiling information from four of our competitors in the wire testing business.  In actuality we have no competitors that we know of.

If the other four testing products you mention are claiming to catch intermittencies by taking measurements of ohmic values on one or even several lines at a time, and then compare these values to some preconceived or "on-the-fly" standard, or if they are using a sub-mode measurement scheme that applies digital techniques to "measure" several lines at a time in an effort to find "intermittencies",  then neither type system is in the same "ball-park" or class as the IFD-2000.

Its the difference between "looking" at something with a set of reading glasses versus an electron microscope!

We don't consider these products our competitors.  They are continuity testers that were designed to perform other semi useful functions but are very marginal when it comes to finding intermittencies.

In contrast, the IFD-2000 cannot take a resistance measurement. It has semi useful modes that allow us to estimate the relative continuity of a wire about as good as these continuity testers do intermittency.  We cannot tell you that a certain line has a finite number of ohms associated with that line.

We can however, tell you if any line is intermittent, down to the sub microsecond range and we can see "ohmic" changes as small as 20 ohms.  (Newly developed testing technology now lets the IFD-2000 measure down into the sub 1 ohm range.)  We can monitor all the lines simultaneously and continuously looking for these micro breaks.  A single event on any line, at any time, will trigger the IFD-2000 and it will tell you exactly which line had the intermittent event.

It's takes entirely different underlying technologies to perform these two very different testing jobs that need to be done. You don't need to compare "competitors" products, you really need one of each.

If for economic reasons you don't want to have two separate wire tests in the testing regimen then possibly the continuity test function can be accomplished via the functional test.  If it tests ok, then obviously the wires are connected.  It's what they're doing at the present time.

I don't recommend it necessarily, it's just that there is more to be gained by visual inspection and reliability testing with the IFD-2000 then by continuity testing.  The Air Force has already gone the continuity route time and time again with very little success as far as NFF and intermittencies are concerned.

I hope this clarifies the issues between testing methodology and sheds a little needed light on the problem you and the others are trying to resolve.

More details on the testing problem are featured in our IEEE white paper and briefing slides which can be found in our enclosed NFF booklet and on our website at www.usynaptics.com.  I suggest that if you are not clear on the issues raised that you take a more in-depth look at them.  A phone call to us may help also.

If we can supply any additional information please let us know.

Warmest regards,

Brent Sorensen
President