Testing MIL-STD-1553 - TEST SYSTEMS

1 Testing MIL-STD- 1553 . After nearly a generation, there are still misconceptions on validating the 1553 databus. An expert draws from his experience to describe the perils and how to avoid then. by Leroy Earhart*. T. he most popular serial data- avionics on the B-52, F-15 and F-16. over the last three years has given us bus in military avionics to- 1553 should continue to broaden its interesting glimpses into the priority day is MIL-STD- 1553 . Since application and probably increase in some companies put on Testing . Some 1973, it has become the choice for areas such as flight control. Next- take the time to train personnel and system integration because of generation aircraft with dedicated acquire test equipment while others flexibility, inexpensiveness and avionics that require higher attempt to ad-lib through the Testing off-the-shelf availability. Today, it's throughput, such as the Advanced process.

2 Most companies fall not only used in military applications, Tactical Fighter, will use 1553 . somewhere in the middle. They gen- but in commercial SYSTEMS as well. Although 1553 has been around for erally have test equipment with partial 1553 is a mature standard for nearly 20 years, there are sill many capability, but lack experience for future applications as well as retro- misconceptions about its test effective Testing . It creates a two-fold problem. First, the full capability of a terminal's design will not be *Mr. Earhart Subcontractors who believe is president of that validation Testing is Test SYSTEMS , too costly and unnecessary The optic version (MIL- Phoenix, Ariz. could find out how STD-1773) and High feed The company expensive it delay can be. Data Bus will supplement supports 1553 1553 , but will not replace it. with test equipment, and validation. Let's consider several seminars and areas that appear to me particularly determined without extensive Testing .

3 Validation troublesome. Secondly, improper operation not The purpose of validation Testing is found prior to a production run or to verify compliance of a terminal's system integration becomes far more fits. Its flexibility enables it to mesh databus interface with MIL-STD- costly to track down in the long run. with other standards and network 1553 . Using published test plans, you Two other factors appear to be architectures such as MIL-STD- can verify, characterize a terminal and responsible for insufficient Testing . An 1760A, which defines aircraft stores. define its margins and limitations. The obvious one is that Testing is The fiber optic version (MIL-STD- information gained is essential to frequently cut back when costs 1773) and High Speed Data Bus will avoid incompatibilities before system increase and time runs short. The supplement 1553 , but will not replace integration. Since validation Testing it.

4 They will not threaten 1553 because does not test the operation or they lack its maturity, and other functional aspects of a subsystem, it benefits. Economics is a controlling can be performed as soon as remote A validated component factor and, with military programs terminal hardware is available. must also being cutback, a drive to modify many Subcontractors who believe that used correctly interfaces with a higher speed bus may validation Testing is too costly and not be feasible. unnecessary could find out how Retrofitting applications should expensive the delay can be. second reason is that there are widely continue to grow, as in the new Our activity in validation Testing held misconceptions about the need for Testing . Let's look at the three Avionics/March 1991. most common fallacies, then at and variations between separate LRUs standard feature. This is a good remote terminal (RT) failures we have all have an effect.

5 Different power example of where accuracy is necessary found in Testing . Finally, we'll supplies can influence board for characterization of an RT's margins. examine several options available for performance. Although these physical Another problem is the r validation Testing . variations mainly affect electrical and sponse time of an RT to a command. The first misconception is: "Vali- noise rejection, it should not be taken We have seen the allowable us dation Testing is not necessary if for granted that protocol tests will exceeded many times. We've had validated components are used to produce identical results with different several RTs respond incorrectly in build the RT. "This is the most widely software or firmware. protocol tests by setting the Busy bit or held false assumption we have come Subsystem Flag bit inappropriately. across. There are chipsets and Using the wrong part (transformer/. components which have partial transceiver) or using the right part validation Testing by the SYSTEMS Improper operation not incorrectly is common.

6 Improper Electronic Analysis Facility found prior to a production initialization of the chip is also common (SEAFAC) at Wright Patterson Air run or system integration because everyone writes software Force Base. Those which SEAFAC becomes far more costly to differently. In fact, some failures are found to be satisfactory were track doom in the long run. actually problems in the chipset design "validated" or "SEAFAC certified." that SEAFAC missed in its Testing ! While using such parts in an RT Each RT, of course, has its own minimizes problems, they do not problem areas. Take, for instance, the eliminate the need for thorough Past Operation output amplitude of one RT that Testing . It is important that the correct The last misconception is: "Vali- validated components be used dation Testing is not necessary together. This seems quite obvious, because the LRU has been operating but we have tested RTs that had, for in the system .

7 " The fact that an RT is instance, the wrong transceivers or already functioning ( , it's flying) Margins are not checked transformers ( , incorrect turns ratio does not mean it satisfies MIL-STD- 1553 . We once evaluated in normal system M. or quiescent state). A validated component must also an RT that had been flying and found operation or be used correctly. One company, a broken address line - among other operational Testing . after its first production run, brought problems. The standard has built-in us a unit with an eight-layer board margins and the Test Plan does test and discovered how costly it was to for them. Using an RT without started transmitting at V but have the wrong taps on transformer required margins in a system reduces decreased to V by the end of the connections. operating margins for the entire 33 word message. While this is not a There can still be problems if the system . Margins are not checked in failure, it indicates a potential problem.

8 Right parts are used correctly. normal system operation One blatant error we've had was an RT. or Consider, for example, physical operational Testing . Normal system that transmitted in response to a receive location. The proximity of command! Another RT, on power up, operation or operational Testing cannot components in the layout can have an verify proper handling of detected began responding, stopped responding - effect. Card placement in the LRU errors or proper noise rejection. then started responding again. Most (Line Replaceable Unit) also affects Consequently, Testing of an integratedproblems are not that exotic; improper performance. Excessive bus cable RT under actual flight conditions willoperation is usually due to length can change electrical never be as thorough as validation misunderstanding the standard or data characteristics and noise rejection. Testing . sheet of the protocol chip. The bottom Another problem is that operation of line is, in three and a half years of We find, in performing RT validation Testing , not one RT passed the RT is affected by user software.

9 Validation Testing , that many RTs have the RT Validation Test Plan on its first problems with the same tests . One is try! Interfacing Verifying that a design meets The second misconception is: "Be Different power supplies MILSTD- 1553 and that options are cause the interface board was validated in performing correctly is an enormous can influence task. To obtain acceptable results it is one LRU, validation Testing isn't necessary on subsequent LRUs. " Remote terminals board performance. necessary to have appropriate test using an interface board that passed equipment and experienced personnel. validation Testing should still be tested Two alternatives are available. One is with at least the electrical and noise the Zero Crossing Distortion (ZCD) to train personnel who can be rejection tests of the RT Validation Test test which fails an RT for detecting committed to validation Testing and Plan. One factor that affects errors for a ZCD of 150 ns.

10 Inad-- acquire dedicated equipment so performance is placement of the card in equate test equipment is the major measurements and results are the LRU. Different bus cable lengths, culprit. Not all test equipment is repeatable. If several MIL-STD- 1553 . variations in proximity to other devices created equal and measurement projects are in the works, it may be resolution to 2 ns or less is not a feasible to set up a test facility. Avionics/March 1991. of the standard. He can assist in EQUIPMENT FOR 1553 VALIDATION Testing solving problems on the spot. As a third party, the specialist may also give more credibility to test results. 1553 Bus Tester Oscilloscope Because of the complexity of Testing , 1553 Noise Generator True RMS Voltmeter the Air Force requires Douglas Aircraft Company, prime contractor for the Connection Panel Impedance Analyzer C-17, to evaluate and approve test Host Processor Function Generator facilities (equipment, test procedures and reports) before validation Testing .