The Cable Test Problem
Acquiring the Connection Map
Test Data and Match Data
Checking for Intermittent Connections
PC-Based vs. Standalone Testers
Unique Benefits Offered by a PC-Based System
Back to "Finding the Right Tester"
Introduction Next Topic | TOC
Interconnections between computers and peripheral devices often require complex multiconductor cables. The design and assembly of such cables rely on various combinations of manual and semi-automatic procedures that are extremely vulnerable to human error. Complicating matters further, many cables do not contain parallel point-to-point wiring, but involve crossovers and internal jumpers not evident through visual inspection. Two cables that appear to be the same and have identical connectors at each end may have totally different internal wiring. If we are particularly unlucky, the equipment we are connecting may sustain permanent electrical damage by the use of an incorrect cable that obligingly fits all connectors.
Detecting faulty or inappropriate cables is achieved through testing. Once tested, labeling helps to insure against errant use. A variety of test equipment have been developed to address these needs, most of which were intended for production testing at the point of manufacture. CAMI Research produces a general purpose cable test instrument driven by PC-software that can be used effectively for many aspects of cable test, design, and manufacture, including production test. By combining wide-ranging capabilities in a single low-cost instrument, benefits accrue to anyone who deals with cables, not just those at the manufacturing site.
The Cable Test Problem Next Topic | Previous Topic | TOC
The bulk copper cable and connectors that are used to make cable assemblies are manufactured using a high degree of automation and rarely contain defects. Completed cable assemblies that are defective owe their problems primarily to errors in attaching the connector to the bulk cable. The most likely possibilities are
1 - a missing connection (an "open"),
2 - an extra connection (a "short"), and
3 - a miswired connection (wrong pin).
Consequently, the most important single test applied to a completed cable assembly is an exhaustive continuity check to ascertain the electrical relationship between each pin of every connector in the cable. Without proper continuity, other electrical tests, such as electrical resistance, insulation breakdown, and impedance, are irrelevant. For many cables, if the continuity is correct, the probability is very high that the cable will function properly, and other electrical tests, if necessary at all, may be performed on a sampled basis.
Defects not revealed by a continuity test include improper clearance between connector pins, insulation thinning, foreign particle contamination, and inadequate contact between wire and pin (bad crimp or cold solder joint). In these cases, either a resistance check or a "hipot" (high potential) measurement may reveal the problem.
When a moderate volume of cables must be tested, firms generally purchase a commercial tester from any of about a dozen major suppliers in the industry. A serious test instrument provides continuity tests for opens, shorts, and miswires as the primary test, and, at greater cost, will also test for resistance, insulation breakdown, or other electrical parameters. Serious professional instruments vary in price, ranging from around $1,000 for basic continuity measurements to well over $5,000 when tests for other electrical characteristics are included.
Testing Continuity Next Topic | Previous Topic | TOC
Checking a cable for opens, shorts, and miswires is really accomplished by the same test: measure the continuity of every pin on a connector with respect to every other pin on every connector attached to the cable. In doing this, every mathematically possible connection is tested, and a precise connection map developed. A cable under test is deemed "correct" when its connection map matches that of a properly functioning cable.
Acquiring the Connection Map Next Topic | Previous Topic | TOC
A two-ended multiconductor cable may be considered topologically to be just a single, large group of pins whose total number is just the sum of the number of pins at each end of the cable. For example, a cable with a DB-9 connector at one end (9 pins plus shield) and a miniDIN-8 connector at the other (8 pins plus shield) will compose a group of 19 test points. Note that for a simple serial interface cable, many of these 19 test points may be null (not connected) and are present only to conform to the physical construction of the connector.
To determine a cable's internal connections, the continuity of every test point must be measured with respect to that of every other test point. For 'n' test points, raw cable connection data is represented by a matrix of 'n' rows by 'n' columns. It is developed by sequentially applying stimulus signals to the cable in which a test voltage is applied to a single test point, and a response acquired that shows which test points receive the applied voltage. Each response constitutes one row of what we refer to as the continuity matrix. Following 'n' successive stimulus-response cycles, acquisition of the continuity matrix is complete. For a mathematically comprehensive test, data acquisition time is proportional to the square of the number of test points. A 256-point scan, therefore, requires four times the acquisition time (and four times the memory) of a 128-point scan.
Figure 1 shows the continuity matrices that would be produced for some simple cable configurations. Note that the diagonal of the continuity matrix will always contain '1's; this represents the unity connection that exists from every test point to itself. Also note that every wire in a cable is tested twice, once for each direction of current flow. As long as there are wire connections only, the continuity matrix will be symmetrical with respect to the diagonal. If diodes are present, however, the continuity matrix will be asymmetrical, showing that for a given conductor, current may flow in only one direction.
Test Data and Match Data
Next Topic | Previous
Topic | TOC
A continuity matrix in combination with the cable's connector types fully specify the point-to-point function of a cable. One data acquisition cycle obtains this information from the cable under test. The resulting test data constitutes the Test Data Buffer. If a model cable is not available, some systems allow manual entry of test data via a keyboard without a physical cable being present.
The most important function of any tester allows direct comparison of the wiring and connectors of two cables. To achieve this, information on a match cable, sometimes called a "golden" cable, must be available. This information may come either from a stored database of valid cable data (sometimes in the form of an EEPROM) or from a model cable whose test data has been recently acquired. In any case, the continuity matrix and connector types of the match cable are referred to as match data and stored in the Match Data Buffer. See Figure 2.
Both test data and match data are considered to be raw data
and require further processing to provide a meaningful display
of wiring or a netlist. Valid test data or match data must be
present for most functions of the tester to operate. In many cases,
once test data has been acquired from a test cable (1 or 2 seconds),
the cable may be removed from the fixture.
Checking for Intermittent Connections Next Topic | Previous Topic | TOC
Intermittent connections (both opens and shorts) result when the continuity map changes in response to flexure of the cable or physical pressure exerted on the connectors. This can be a serious problem because it is so easy to miss during routine testing. Checking a cable for intermittent connections involves the continuous reacquisition of new connection data as the Operator moves the cable around or applies pressure to various locations. Changes in continuity or resistance while under physical stress indicate the presence of an intermittent connection. Under certain circumstances, hipot and resistance testing may find an intermittent connection without flexure, however, many problems are found only when the cable is physically stressed.
PC-Based vs. Stand-Alone Test Systems Next Topic | Previous Topic | TOC
The involvement of PCs in electronic test equipment spans a wide range. In the most loosely connected applications, data collected by portable test equipment is uploaded into the PC for analysis or report generation, however, neither the test instrument or the PC software are dependent on each other to function. At the other extreme, the test instrument is controlled totally by the PC and cannot function independently. In some cases, advertising suggests to equipment buyers that if their product utilizes an Intel 80x86 Because the phrase "PC-Based" has been casually applied to many situations,
PC-Based Equipment: the equipment requires a personal computer for operation and cannot function without it. To justify its expense and bench footprint, PC-based equipment must make effective use of full-screen high-resolution color graphics, high-speed computation, large data storage capacity, and the full-size keyboard. Other input devices such as a mouse, trackball, and voice control, may also be employed that would otherwise be uneconomical or impractical for a non-PC-based equipment, or would defeat the portability requirement of such equipment. Note that numerous test instruments permit the upload and download of information to PC software for off-line data analysis and storage, but operate on a stand-alone basis. As such, these instruments do not qualify as "PC-based".
Microprocessor-Controlled Equipment: an embedded, on-board microprocessor effects control of the system and executes whatever function that system is intended to perform. Such systems are usually stand-alone and totally self-contained, with the presence of microprocessor control being invisible to the User. Dedicated switches, keyboards, indicators, and numeric display units comprise a non-standardized User interface. "PC-controlled" stand-alone equipment using an embedded single-board PC (Intel 80x86 microprocessor with Microsoft operating system kernel) functions largely as an embedded microprocessor-based system, and offers advantage primarily to the manufacturer of the equipment as follows: (a) software development may employ high-level programming languages and debugging tools for which many skilled programmers are available, and (b) hardware implementation uses high-performance, off-the-shelf computer modules minimizing any custom design. The human interface, however, still relies on dedicated electrical components, although more sophisticated components, such as a touch-screen display, may be employed.
Over the last ten years, a number of relatively inexpensive microprocessor-based testers have been introduced that quickly acquire the connection map of a cable and store it in on-board memory. By saving in RAM the connection map of a model cable known to be good, it is a simple matter to acquire data from a cable under test and compare it to the model. A matching connection map indicates that no opens, shorts, or miswires are present. In accomplishing this, the cable under test becomes certified, and the benchtop tester will have performed its job.
In evaluating the manufacturing process, CAMI Research has found that other steps necessary in the production of cables are currently handled separately from the test process and often do not involve computerized equipment. These steps include cable design, cable assembly, fault location, labeling, hard-copy documentation, and cataloging in a database for future reference. We believe that integrating these functions in a single general-purpose instrument greatly facilitate cable production by saving expensive engineering time and production labor, reducing waste, and minimizing error. Achieving the desired multi-purpose functionality, however, is beyond the ability of the 8-bit microprocessors and rudimentary display devices used in most benchtop units. By moving all computational, storage, and display functions to the PC and leaving behind only data acquisition hardware, CAMI Research has produced a cost-effective PC-based cable test system with new capabilities that address multi-faceted needs.
Unique Benefits Offered by a PC-Based Tester Next Topic | Previous Topic | TOC
Wiring Display - Any tester must convert raw connection data from the continuity matrix to a wire list. For the purpose of wiring analysis or fault location, wire paths may also be displayed as a graphic wiring diagram when a PC is employed, as shown in Figure 3. The direction of view, indicated by a small connector icon, may be changed to show wiring looking into the pins or looking into the termination. Individual wire paths may be highlighted using the cursor control keys.
Descriptive Notes and Labels - When new cable data is acquired, the test technician may annotate it with descriptive notes, part numbers, vendor data, color codes, signal assignments, or other text and store all information in a searchable on-line database. This data may be recalled by matching against wiring, file name, descriptive text, or by manual selection. Database size is limited only by available disk space. Label text may also be entered and stored with each cable.
Comparison Against a Model - In addition to checking test data against a golden cable, the on-line database may be searched automatically for matching wiring. This ability permits cable identification and the automatic association of descriptive text and labels to the measured cable. For a manual comparison, the operator may quickly alternate between the schematic display of test and match wiring. Because the connectors and wiring are geometrically located in exactly the same position on the screen, rapid alternations will instantly reveal subtle differences in the wiring, particularly if the wire paths of interest are highlighted.
Printing - The complete cable schematic, along with the wire list, descriptive text and label text, may be printed on a laserprinter or inexpensive ink jet printer. This high-quality documentation is essential for maintaining accurate wiring and construction records on cables and is based on measured data, not data redrawn by a draftsman from a rough sketch. Thus, engineering time is saved, and another opportunity for error is eliminated. Data produced in this manner may be used directly in equipment manuals or other printed documentation, or it may be enclosed with each cable that a contract assembly house ships to its customers.
Intermittent Connections - When intermittent connections are found, the wiring display is again useful in showing which wires were found to be intermittent. In combination with the audible tone sounded during flexure, the location of the problem may be readily identified.
Automatic Test Sequences - In a production environment where unskilled operators may do the testing, it is unnecessary for the test operator to enter commands at the computer keyboard, or read results on a computer screen. For pass/fail testing, a test engineer can predefine a desired test sequence and store it as a script on disk. CAMI's CableEye tester is equipped with a single TEST pushbutton, and three LED indicators labeled READY, MATCH, and ERROR. Pressing the TEST pushbutton triggers a predefined Macro, with results indicated on the lamps. The operator's only task, then, is to mount the test cable, push the button, apply a label (optional), and place the tested cable in an appropriate bin.
Cable Design - To design and test a cable, the test engineer simply enters the desired connector types and wire list using a built-in editor. From this, a cable schematic and ordered wire list are generated by software and may be printed or stored in the database with descriptive notes. If cable design and production occur in different facilities, the engineer can modem the design data, or send a diskette, to the cable assembly plant. Technicians at that location may then use their CableEye system to load the test data and reproduce the printed wiring design, as well as have test specifications preloaded in the tester when completed cables are ready for testing. Again, time is saved and the chance for human error greatly reduced.
Summary TOC | Previous Topic
PC-based cable-test systems offer a suite of benefits that are neither available nor economically feasible on traditional stand-alone testers. A PC-based system not only tests cables but also provides an integrated software package for cable design, labeling, documentation, cataloging, data logging, on-line assembly checking, and test scripting. The added cost of a PC is easily justified by (a) the reduced cost and increased reliability of the test fixture itself; (b) the elimination of errors in transcription, drawing, and rekeying of wire lists as cables pass through design, test, and documentation phases; and (c) the ongoing savings in engineering time, fault-location time, and documentation. Finally, continuing advances in test software become available to customers at low cost without requiring any change in the test fixture, data-acquisition electronics, or connector cards.
Christopher E. Strangio
Director of Marketing, CAMI Research Inc.