Selecting a Systems Diagnostic Design Development Tool
By: 

Selecting the Best Technology For Troubleshooting No Fault Found / Intermittent Conditions
By: Brent Sorensen,President, Universal Synaptics

Diagnostic Tools Fit Specific Needs; Success Comes from Best Match-Ups
By: Tim Webb, DiagnoSYS Systems, Inc.

Diagnostic Tools using Automatic Probing
By: Jim Crosson , Huntron, Inc.

When tasked with selecting a Systems Diagnostic Tool, the user must understand and mitigate all business case risk related to this selection.  This can only be accomplished with the understanding of how the tool will meet the needs of the total systems engineering process.  Alternative and complimentary tools need to be identified and understood in order to best leverage the strengths of a tool without compromising the most paramount objectives of the systems’ diagnostic design.

Systems Diagnostic Tools need to be researched, evaluated, and accepted before most of the Subsystems Design Requirements are derived and flowed down to the developers, both internal and subcontracted. It is imperative that such tools serve to bring together, analyze, and optimize all aspects of systems and subsystems integration of the diagnostic design.

Since this Systems Diagnostic tool needs to absolutely be in place prior to the identification of the parts or components, the tool must be able to efficiently use the functionality of the design.  The design functionality must be defined in the Diagnostic Design tool early enough to effectively address System Diagnostic requirements feedback from users such as Production, Operations, and Logistics Support. The total systems design influence enables well-informed decision making to occur during the earliest stages of design. This systems process provides the influence needed for the early higher level (or System) design decisions such as sensor placement, repair item partitioning, redundancy, access, safety, etc. A systems Diagnostic tool must therefore have functional modeling capabilities as a fundamental requirement.

As parts are selected, the Systems Diagnostic Tool needs to support a model and process that captures all of this lower level data and integrates it with the highest levels of the System Model. The tool must also support rapid trade studies during the lower level part selection. The tool then needs to support a fully integrated, system-wide, FMECA to identify the lower level failure modes and the effects of these failures at the higher level.  The criticality of the failures must also be identified.

Since it is not always possible to enter a program at the early design phase, the tool must be able to effectively support modeling and analysis at any stage of development, including work on legacy systems.

The Systems Diagnostic Design Development Tool must be able to fully support diagnostics design analyses at any stage of development or use. This requires the capability to provide, at will, instantaneous vertical and horizontal visibility and tracking throughout the subsystems and systems layers. This capability of the Systems Diagnostic Tool requires it to function equally efficiently from both functional and failure-based information (a characteristic of an advanced hybrid Diagnostic Development tool).

Since many complex system designs require large and multi layered breakdowns, the Systems Diagnostics Design Development Tool must be scalable to efficiently model and analyze these large structures as well as small structures such as a circuit card or encapsulated device.

The selected tool must meet your and your customer’s technical and business case needs within the full Systems Engineering process.

SYSTEMS DIAGNOSTICS DESIGN DEVELOPMENT TOOL CHECKLIST

The Systems Diagnostic Design Development Tool must:

• Drive System Design Requirements
• Fully support very early design influence decision making process based on functionality of the system
• Provide a graphical means to efficiently communicate design functionality that describes diagnostic coverage
• Provide effective design assessment and produce a requirements Gap Analysis at any phase in development or product use
• Capture iterative and evolving data and design knowledge while effectively reporting system or subsystem diagnostic capabilities at any point during development.
• Integrate all program disciplines such as Reliability, Maintainability, Production, Logistics Support, Test Engineering, Cost Analysis, etc.
• Provide true hybrid integration of functional and failure mode analysis
• Fully support both common cause (single fault) and Multiple Failure diagnostics
• Provide scalable and Open architecture to allow the importing, exporting and exchange of data for rapid design and life cycle support of embedded Operational Health Management, and Logistics Support.
• Not be dependent on any industry standard, but can easily adapt to such standard(s)
• Be widely used within Industry and Government Agencies
• Provide accurate and repeatable analysis results, with high confidence, from solid and proven algorithms backed by world wide industry experience
• Supported by a company with a long-standing reputation in industry

 Brent Sorensen,President, Universal Synaptics

There is little argument that one of the most important things for success or even survival is having superior tools and technology.  As such we have seen the unparalleled development and fielding of complex electronic systems that support our defensive, communications, power, transportation, space, industrial, and other infrastructures.  Keeping these now critical systems running reliably, in the face of relentless aging, has made the role of maintaining these costly systems one of critical importance.

Test equipment manufacturers have been working diligently to develop equipment with expanded features and phenomenal accuracy to help test and service these systems, yet maintainers still find themselves increasingly unable to keep them running with any significant degree of sustained reliability or cost efficiency.  This is especially true in military and commercial aviation systems, where it is not uncommon to have 50 percent or more of all reported operational malfunctions going undetected and therefore unrepaired during subsequent static or ground-based testing.

Diagnostic labels such as NFF (No Fault Found) or CND (Can Not Duplicate) and their statistically increasing rates of use, quantify the direction and extent of this testing-void problem.

From a root-cause perspective, electronic malfunctions can be categorized as either hard or intermittent failures.  If it’s a hard failure, the failure repeats every time and there is an estimated $100 billion worth of test equipment in place to accurately test and diagnose these comparatively easy failures.  You might even say that it would be impossible to misdiagnose a constant failure.  However, if it were an intermittent failure some would say that your best tool is a big bag of luck, and luck can be reduced to a notion of increased probabilities.

The probability of traditional test equipment being able to detect a randomly occurring intermittent failure or event is extremely low.  There’s simply too much fixed scanning, sampling and digital averaging involved to capture a brief, one-shot, low-level failure causing event.  There’s hope however.

As a rule, active and passive electronic components either stop working altogether or drift out of their original design parameters over time. In contrast, all the electromechanical connectivity elements (wiring, connectors, crimps, splices, solder joints, relays, circuit breakers, flex circuits, backplanes, etc) or the part of electronics that “glue” all the components together, rarely abruptly fails.  Instead, like machinery, they loosen or degrade over time, due to thermal, vibrational and contamination factors occurring in their operational environment.  With age, their operation becomes compromised and their failure mode is mostly intermittent in nature.

Because of the random nature of the failure mode, direct testing with expensive and highly accurate digital technologies (Digitizers, DMMs, DSOs), or other on-off technologies are simply not going to work.  And to be practical, a lot of analog technologies will not work either when large numbers of circuits or wires need to be tested.  An expensive analog oscilloscope is no better than a simple $3.00 test-light when the 30 millisecond blink rate of the human eye is the limiting factor and you need to be detecting intermittencies at least into the microsecond range.  A Time Domain Reflectometer (TDR)  and Standing Wave Ration (SWR) can tell you a lot about conditions on a transmission line, but if the line does not exhibit any intermittency during their short test period, what are they going to report?  And if you are going to test longer or on multiple lines, you are going to have to put hundreds of these devices to work at the same time, which is not likely to take place.

One technology from Universal Synaptics is based on analog neural sensing technology that does work effectively to find intermittency. . With this method of testing, as exemplified by the IFD-3000, all lines or wires of interest are connected to 256+ individual sensors arranged as a neural network. If a change in current is sensed on any of the lines, the network will report a problem, as well as automatically identify the failing line.   The accompanying computer will then capture and display a trace of the severity and duration of the failure event and it will update an on-screen graphics display and report its physical address in Unit Under Test terminology.  At the end of testing, the time stamp of each intermittency is also available for printing along with a reliability validation report for documentation requirements.

While legacy testing methods for continuity, functional or reliability testing delivers rather poor or non-existent performance when testing specifically for age-related intermittency or reliability, the IFD-3000 delivers increased levels of performance that is literally millions of times better on any individual test line. In addition, because of the IFD’s parallel nature, the increased performance in probability is orders of magnitude better when large systems need to be tested. 

For a more in-depth discussion of comparable testing capabilities and other aging-intermittency/NFF testing issues, see related article “The Achilles Heel of Modern Electronics” available at the link www.evaluationengineering.com/archive/articles/0604/0604modern_electronics.asp

 Tim Webb, DiagnoSYS Systems, Inc.

When diagnosing printed circuit board assemblies, selecting the right tools can be a daunting task. Whether you need to debug a design, catch a production fault or repair a unit, knowing what kind of information you need to obtain can reduce the time and energy to effectively evaluate the tools you will need in your test and diagnostics arsenal.

Problems can come from almost any direction in a variety of forms. Faults on printed circuit boards can emerge as a shorted or open trace, shorted substrate or leakage on a pin. Voltages can be too high or too low. Output pins may not transition at the right time. Waveforms may be the wrong shape, frequency or amplitude. Resistance levels may be out of tolerance or capacitance levels incorrect. The data bus may be wrong; the address bus may be stuck high, firmware may be the wrong revision - preventing the board from booting. The list can be endless.

Knowing how to address each problem and knowing in advance what repair action to take when they occur can save a lot of time. Here is a quick checklist for determining at an early stage the likelihood of successfully diagnosing a problem at the manufacturing and support stages.

1. Is the UUT built with DFT in mind?
2. Do you have access to component pins or test points?
3. Does the UUT utilize JTAG?
4. Do you have a schematic?
5. Do you have a working board to compare to?

If you answer NO to all of these, troubleshooting will be virtually impossible.  With each YES answer, the troubleshooting process should become easier.

The more information you have about the board, the faster you can determine appropriate corrective action. If you don’t have sufficient information, you need to acquire some in order to successfully and cost-effectively troubleshoot the board. If you have a good board for comparison, some tools and test methods can help find variances between the two. This may uncover some basic faults – but should not be relied on for all faults.

ASA (Analog Signature Analysis) is a good way to start. It provides a “power off” method to compare the impedance values of each node between good and bad boards. Some tools store the data to a file, allowing you to generate a program from a known good board for use at any time. This technique will work for analog and digital boards, but it can be time consuming, especially if there are many surface mount components. Also, it requires the availability of and confidence in a “known-good” board. Automated versions are available on a flying probe platform, but they require higher volumes to justify their cost. ASA can find a number of faults, including some opens and shorts, changes in resistance and capacitance as well as leakage and defective PN junctions.  Because the signatures are taken of specific devices, it is possible to diagnose to a small group of components. 

A schematic can definitely make the troubleshooting process easier. In most cases it is essential. Some products on the market can generate schematics for undocumented boards, and you should consider them if you must have a schematic. A good system, such as the DiagnoSYS PinPoint II Functional Tester, will be able to generate a gate level schematic in a few days depending on the complexity of the board. The system should be able to create a file for use with similar undocumented boards coming in for repair.

One of the best methods to troubleshoot IC failures is by using a functional in-circuit test system with component-level adapters. This method uses a back-driving technique to functionally prove the condition of the component.  Functionally testing a component in-circuit, can provide a good level of confidence that the component operates as expected.  A good in-circuit test should be able to determine true tri-state levels on appropriate pins. Tests covered should include open circuit, pin-to-pin shorts, internal substrate shorts, functional, voltage checks, memory cell contents, and more. Knowing that all components are functioning correctly is important, but not all faults materialize in this form, especially complex ICs for which functional tests may not be feasible. You will also want to test the traces between components and from connectors around the board.  This option, available on some testers, can be very useful.

In-circuit testing presumes you can access component pins or test points, but this is not always possible. Ball Grid Arrays (BGAs), for example, provide no direct access to pins, so you should consider using boundary scanned BGA ICs. If some components on the board are JTAG-compatible, refer to the schematic to make sure the scan chain is accessible – usually through a Test Access Point (TAP). Boundary scan can perform a series of tests, including some functional testing, interconnect testing, and in-system programming.

If the defective board is CPU-based, you could use a ROM emulation system that accesses the board through the boot ROM. You may also need to use logic probes and oscilloscopes to further aid the detection of faults.

For analog circuitry, a good selection of instruments is handy, but the required specifications will depend on the task at hand. These would typically include an Arbitrary Waveform Generator with more than one output and the capability to generate various waveforms. An oscilloscope with storage features can be very useful for sampling and storing signals for analog and digital situations.

With so many possibilities to consider, it’s virtually impossible to have a “Swiss army knife” to cover every test and repair need. DiagnoSYS provides a series of products designed for troubleshooting faults to the component and or to the connection. Faults are detected using a combination of the previously described techniques.

Jim Crosson , Huntron, Inc.

The move towards miniaturization means less PCA real estate can be devoted to accessible test points or connections. Surface mounted components make manual testing a serious challenge of one’s dexterity and eyesight. The introduction of ball grid array ( BGA ) components has further reduced accessible test points.

To assist the test engineer to adequately test and verify boards using leading edge technology several test tools are needed such as precision probing equipment to eliminate probing errors, accurate vision systems able to recognize areas of the board and test systems capable of testing BGA circuits.

Visual Inspection is a key point when looking at diagnostics, approximately one third of PCA field returns problems are diagnosed visually. (Broken connectors, lifted pads burn spots…). High density SMT limits the effectiveness of a visual inspection with out the aid of vision tools, (magnifying glasses are not good enough).  Enhancing vision is an effective way to help the diagnostic process products such as video microscopes, and camera based repair viewers are migrating from the production and quality departments and are proving very useful in diagnostic support.

Cost effective diagnostics a function of applying the most appropriate test method to the problems that are best prepared to isolate and identify faults for corrective action. As capable as some of the test equipment available on the market today may be, they are not always the most cost effective when it comes to all of the potential test problems that can exist on a circuit card.

It may also be impractical to "fixture up" a piece of test equipment to be capable of recognizing all of the potential problems that a board may have when many of those problems may occur infrequently. The question then is "What is the balance between..."
1. "...a significant investment in the software and fixture development of a high end tester as the primary test method..."
versus
2. "...a balanced investment plan across the primary and another test method which cost effectively maximizes fault coverage"?

The answer is a function of the cost of the board, volume of production, cost to develop test software/fixture and the precision measurement ability of the primary tester.

A more recent development is the use of robotic platforms to interface to specific points on a PCA. Flying probe technologies from companies such as Huntron, Inc. are now being used to replace manual test methods. Interfaced to PCI/PXI based powered test instrumentation from companies such as National Instruments, Tektronix and GOEPEL electronics, these robotic probers can be used as a guided probe for waveform and voltage measurement and take measurements on test points, via’s or IC pins.

Robotic probers can be used to apply test signals from power-off signature analysis instrumentation such as the Huntron TrackerPXI. This would greatly reduce the time needed to troubleshoot complex printed circuit assemblies and reduce the burden on engineering resources. Custom cables, probes and interface heads can be developed to increase measurable bandwidth which is especially important when testing RF signals.

Using commercial off-the-shelf software packages such as National Instruments LabVIEW, the X, Y and Z movements can be precisely controlled. Useful software additions such as the CAD Toolkits, allow for the integration of CAD-based PCA layout viewing directly into the test application. Robotic probers are very flexible in that changing between different test routines is simply a matter of mounting a different PCA and selecting the appropriate test sequence.

The sophistication of robotic systems makes the initial capital investment higher but the return on investment can be a significantly short amount of time. The burden of manual probing on engineering resources and the risk to PCAs that are tested can have a tangible cost put on them. This cost will be reduced by the significant time savings and repeatable, extremely accurate probing.

Robotic Probing provides at least a 10:1 improvement in probing time.  This saving is due to the fact that the operator does not have to interrupt the program to look up reference points in a diagram or other documentation.  Also, Robotic Probing does not require judgment or manual dexterity on the part of the operator.  It eliminates the need for ancillary documentation to identify physical location of the probe points, and it further eliminates errors in probe placement due to operator error and variations in probe pressure

The selection of diagnostic tools may not be as critical as the determining how to offset the limits of one’s dexterity and eyesight.

• If your friend forwarded this newsletter to you, please register as a member and receive The BestTest Newsletter -- absolutely free!
• If you wish to update your news preferences or cancel the subscription, please unsubscribe.
• If you have any questions, please email experts@BestTest.com

Get the widest selection of test related books and software at the BestTest Online Store.