In 2007, MIL-STD-461 turned 40 and Version F was released on Dec. 10. MIL-STD-461F includes changes as a part of normal evolution to update requirements and methods to more closely evaluate products for the changing electromagnetic environment.
The new method restores a test that was deleted by the release of MIL-STD-461D in 1993. Changes in the standard may be significant when qualifying products to the now current revision. In conjunction with the release of MIL-STD-461F, three Data Item Description (DID) documents were released.
Here is a clause-by-clause review of the standard where differences between the E and the F versions are notable.
Interchangeable Modular Equipment (4.2.7)
A new item in MIL-STD-461F requires qualification of assemblies when new line replaceable modules (LRMs) are incorporated into devices. The qualification can be by test or similarity assessment but requires approval of the procuring agency.
Construction and Arrangement of EUT Cables (4.3.8.6)
“Input power leads, returns, and wire grounds shall not be shielded.” This requires breaking out the power leads that are part of a cable bundle shielded for the test (4.3.8.6.2).
In 1993, MIL-STD-461D definitively established the requirement to test with cables of the type used in the installation. This was a major move to force responsibility for EMI/EMC control into the hands of the integrator/supplier or whoever had providence over the interconnecting and power cables.
This revision removes the capability to shield power cables as an EMI control measure because, in most installations, shielded power cables are not the norm. An exception is discussed (A4.3.8.6) where filtered power is provided from another device: Shielding may be supported.
In addition, shielded cables for Navy surface ship applications may allow an unshielded section for radiated tests but not for conducted testing. The test configuration needs to be described in the test procedure and approved by the procuring agency. But basically for most applications, the power cables should be unshielded.
Computer Controlled Instrumentation (4.3.10.2)
Verification of software needs to be described in the test procedure. For commercial software, identification of the manufacturer, model, and revision must be provided. For locally developed software, the control and methodology must be indicated.
This presents a challenge for test-procedure developers who aren't part of the designated test laboratory. The laboratories should prepare the necessary documentation and make that available to procedure writers to support designation of the test laboratory.
Bandwidths (4.3.10.3.1) ??Alternate Scanning Technique
“Multiple faster sweeps with the use of a maximum hold function may be used if the total scanning time is equal to or greater than the Minimum Measurement Time defined…,” referring to the measurement time found in Table II. This may give an impression that the test duration can be reduced, but clearly that is not the intent.
The discussion in the appendix indicates that the faster sweep allows for capture of low duty-cycle or intermittent signals. However, in section 4.3.10.3.3, the scanning rate must be adjusted to capture infrequent emissions.
To assure capture, the overall scan needs to be at every one-half resolution bandwidth over the EUT cycle. Realizing that signals such as frequency-hopping modulations would require an exceptionally long sweep for all of the energy to be quantified, don't hesitate to use the faster sweep with multiple max-hold capture to show the envelope of this type of signal.
Frequency Scanning (4.3.10.4.1)
The susceptibility sweep rate or step size has been increased for frequencies above 1 GHz, allowing a faster susceptibility test. This change of sweep rate or step size reduces the test time even with a long EUT cycle time since the cycle time affects the dwell period and not the step.
Thresholds of Susceptibility (4.3.10.4.3)
MIL-STD-461F adds a statement, “Susceptibilities and anomalies that are not in conformance with contractual requirements are not acceptable. However, all susceptibilities and anomalies observed during conduct of the test shall be documented.” What is the implication?
It is not unusual to test hot—at levels above that required. And the possibility exists to observe an anomaly at the elevated test levels only to find that the anomaly is not present when the correct test level is applied to the EUT. This new statement requires documentation of all observed, albeit compliant, anomalies.
Emission and Susceptibility Requirements, Limits, and Test Procedures (5.3)
A new CS106, formerly identified as CS06 in MIL-STD-461C and dealing with voltage spikes on power input lines, is listed in Table IV. The applicability list of Table V includes a few changes in applications associated with some of the test methods such as CE101, CS109, CS115, CS116, and RS101.
CE101 (5.4)
Applicability has been added to surface ships. In addition, the appendix provides some tailoring guidance for high current loads or certain wiring considerations. Suggestions are made for the use of a 5-µH line impedance stabilization network (LISN) and limit adjustments with frequency range changes.
MIL-STD-461 has long supported tailoring for several revisions, but this represents one of the few specific suggestions for tailoring. The tailoring falls on the procuring party, but a test plan could recommend the tailoring for approval by the procuring party. Calibration verification with all test equipment including cables, probes, attenuators, amplifiers, and receivers is accomplished prior to test. If multiple limits for different power inputs are specified, use the most restrictive limit to demonstrate.
CE102 (5.5)
No changes in the requirements except the tailoring regarding Section 5.4 affects CE102.
CE106 (5.6)
Testing of both the receive and the standby mode is unchanged. The CE106 transmitter limit for the 2nd and 3rd harmonics was redefined to a level of -20 dBm (87 dBµV) or 80 dB below the fundamental, whichever requires the least suppression. All other harmonic and spurious emissions must be suppressed by 80 dB.
Assuming a 100-W (50 dBm, 157 dBµV) transmitter, the suppression would be 70 dB to achieve the -20 dBm level for the 2nd and 3rd harmonics and 80 dB for all other frequencies. Making this measurement requires a dynamic range sufficient to show compliance to the limit.
Other issues include determining the frequency span associated with the harmonic emissions. How is the transmitter output power verified since in-band testing is not required? How is the fundamental suppressed without sacrificing sensitivity at out-of-band frequencies? Is the power in the harmonics sufficient to cause nonlinearity in the detection system? How do you handle connection to a transmit port with a type N connector during testing up to 40 GHz?
These are some of the questions that need to be addressed prior to testing—typically during test-procedure development—so both the right equipment and test approach are in place to support the test. Be prepared for several hardware configurations and the associated calibration verification.
CS101 (5.7)
While there are no changes in the requirements, don't forget the capacitors. The higher-frequency losses in the LISN without the capacitors are fairly dramatic and result in a significant undertest. Make all personnel aware of the potential for shock hazards from the isolated oscilloscope configuration.
CS103, CS104, CS105 (5.8, 5.9, 5.10)
There are no changes in the requirements because the testing has been a tailored requirement in the contract for a long time. This test normally calls for a lot of preparation to achieve the test method and limits. Often, the procedure is developed as part of the contract preparation, and if not, the procedure and limits are developed and then added to the contract through test-procedure approval.
CS106 (5.11)
This is a new requirement that brings power-line voltage transient testing back into the requirements for some applications. It restores CS06 testing from MIL-STD-461C, superseded in 1993, but with only one pulse duration.
The details are spelled out in the standard, but basically the 5-µs pulse at 400 V is precalibrated into a noninductive 5-Ω resistor. That generator setting is used as the maximum applied if the 400-V pulse is not generated during the test with a lesser generator setting.
The waveform characteristics are very well defined, and some of the older spike generators are not adequate for the specification. Once the generator level is calibrated, the positive and negative transients are applied to all ungrounded power inputs between phases or between the phase/positive and the neutral/return.
The application between the chassis and phase/positive is not applicable. The test duration is 5 minutes with a 5- to 10-pulse/s repetition rate for each polarity. Unlike the old CS06, phase synchronization is not applicable. Again, watch for the shock hazard with the ungrounded oscilloscope.
CS114 (5.13)
The testing is basically the same with an additional common-mode test for power leads in the 4-kHz to 1-MHz frequency range for some applications. Here is a refresher on the process because often it is accomplished incorrectly:
• Precalibrate the applicable calibration curve levels to establish a maximum forward power level for the test-frequency range. The standard also reminds us to use the same hardware as used for the calibration.
• Select a cable for test and apply the lesser of the test current (note that test current is the calibration current plus 6 dB) or the maximum forward power. Why are they different? During calibration, the actual loop impedance is 100 Ω so the current would be reduced by 6 dB compared to a 50-Ω circuit.
• The test-signal modulation is specified to be pulse modulation (PM) with a 1-kHz square wave. Most signal generators in the test-frequency range do not support PM so AM is frequently used, which would add 6 dB if 100% AM was used to attain the on/off ratio specified.
Two issues with using AM: Does the amplifier have the necessary drive in its linear region, and can the test article tolerate the overtest? Assuming that both issues are satisfactory, test with AM. Remember that noted susceptibility may be because of the excessive test level from AM, so measure the threshold considering this factor.
Measuring the applied current also can be tricky, if not impossible, with the modulation applied. The drive level is determined with the test-signal unmodulated CW.
Previously, there were some issues with testing the power leads or phase lead in shielded power cables so often the test engineer would cut the shield to access the leads. This was NOT the intent. The phase or power group should not be tested separately if the cable is shielded. MIL-STD-461F resolves the problem by declaring that shielded power cables are not allowed. Phase leads are tested as a group (common-mode), not individually.
CS115 (5.14)
There are no changes in the requirements or procedure except that we must use the same hardware as used for the calibration. CS115 testing is accomplished with the generator set at the precalibration level, not the lesser of the current or the precalibration level. Also, as with CS114, test the phase leads as a group, not individually.
CS116 (5.15)
The test requirements have two changes that are easily overlooked.
• The requirement to test with the power off has been removed.
• Paragraph 5.15.3.4.c(3) instructs us to apply the calibrated test signal with a note to reduce the level if necessary. We also are cautioned that, for shielded cables or low-impedance circuits, it may be preferable to gradually increase the drive to the lesser of the precalibrated level or the test current. In this case, the test current is defined as the current not plus 6 dB even though the calibration loop impedance is similar to CS114. Since we seldom know the characteristic impedance of the cable loop, I think that all tests should be performed as if all circuits are low impedance.
One notable difference in CS116 vs. CS114 and CS115 is that the phase leads are tested individually (differential-mode) instead of as a group.
Finally, note that the limit is peak current. We have witnessed tests where the rms current was used as the calibration level and the pulse subsequently was applied. Setting the drive level at rms and allowing the pulse to rise to peak usually result in a fairly severe overtest. Normally, rms measurements are used throughout the standard, but the measurement should be in the same terms as the limit for comparison.
RE101 (5.16)
The change in this test method involves dealing with over-limit emissions. If over-limit emissions are detected at the 7-cm antenna location, MIL-STD-461F calls for determining the distance from the EUT where the emissions meet the limit. This data is used to help determine if the emissions need to be suppressed or if some reasonable setback can be prescribed.
RE102 (5.17)
There is not much change in the requirements or procedure; however, the applicability and frequency ranges were slightly altered. For example, the exemption for testing at the transmitter fundamental frequency added the following phrase, “and the necessary occupied bandwidth of the signal,” to the procedure.
The upper test frequency based on ten times the highest intentionally generated frequency of the EUT still applies and may offer some test-time relief instead of automatically measuring out to 18 GHz. The upper frequency level is predicated on the EUT frequencies, which should be determined in advance of the test. Also, specific antennae are called out in the standard.
RE103 (5.18)
RE103 testing presents a lot of challenges. This test is an alternative to CE106 and should be used only when CE106 is not a viable option. The changes from the previous revision are not significant, but because of the issues, a discussion here is merited. The test process:
• Verifies calibration of measurement system including the transmit frequency rejection network.
• Establishes the far-field test location and positions equipment.
• Measures effective radiated power (ERP) assuming the power monitor is not available. Verifies that the ERP compares with the expected ERP based on the operating parameters of the EUT.
• Establishes a limit based on measured ERP.
• Scans the measurement receiver over the test-frequency range to locate harmonics and spurious emissions. Measures and compares to the limit. Note that the measurement system antenna may need to be positioned to maximize detected emissions.
Verify calibration including the rejection network. For most applications, a rejection network is needed to prevent overloading of the detection system. Even if the transmit power is low enough not to damage the receiver input, the high-level signal may cause spurious signals in the receiver system that could mistakenly be observed as valid signals.
The rejection network normally is a notch filter but could be a variety of high-pass and low-pass filters preventing the fundamental frequency from overloading the receiver. Since a tuned transmitter is tested at various frequencies, provisions are required for the rejection network to be tuned to achieve the proper rejection.
A simple attenuator will cause sensitivity loss throughout the test frequency range. In short, consider the RF conditions of sensitivity, receiver spurious response, and dynamic range to get an accurate and valid measurement.
One issue is radiated testing in the far field, which for low-frequency transmitters can be a very large distance. As an example, a 300-MHz transmitter may be installed in a vehicle and connected to a 1.5-meter whip antenna. In this case, the far field would be 4.5 meters using the RE103 calculation method.
Also, the limit may be less than the detection-system sensitivity. Even 1-W transmit power can present issues depending on the system antenna and operating frequency. Lots of planning is needed for this test.
RS101 (5.19)
Changes are minimal with one significant procedural difference: The scan rate of “three times the standard” has been slowed to the “standard” scan rate. Because of the number of test locations for the 30-cm x 30-cm placement, this seemingly minor change results in a significant test time increase.
The use of an alternate Helmholtz coil method may need a closer examination. For most labs, though, building a Helmholtz coil large enough to accommodate the variety of equipment sizes is one big drawback.
One suggestion is to develop rationale and a procedure to be selective about the test locations based on likely sensitivity of the EUT components. The test plan could limit the test locations based on this analysis. Will the contract technical representative accept that approach?
RS103 (5.20)
RS103 saw some changes that may be considered significant or minor depending on the perspective of the responsible party for the EUT or the test facility.
Sensor placement clarification was added to position the sensor at the EUT location and to position the sensor vertically at the point of the EUT illumination. In addition, a test was added to verify that the sensor is responding to the fundamental frequency as opposed to the harmonics of the test amplifier.
Positioning of the radiating antenna at 1 meter or more from the EUT was incorporated, eliminating the capability of moving the antenna closer to achieve some of the specified field strengths at all test frequencies. This will cause many labs to lower the test levels they can support or create a demand for higher-power amplifiers. Either way, the cost of testing will increase.
Changes in the receiver limit may have the largest impact in those cases where receivers are included in the test article, which are ever-increasing with wireless technology being incorporated into more and more products. The standard states that there is no requirement at the tuned frequency of antenna-connected receivers except for surface ships and submarines. Does this mean that the frequency range is exempt from test?
Review of the appendix for guidance indicates that there is no relaxation for platforms other than surface ships and submarines. What is the limit for surface ships and submarines? It isn't provided in the standard.
The prior revision provided a limit of RE102 plus 20 dB. But with the removal of a reduced limit, are receivers with embedded antennae supposed to tolerate and operate in the presence of a 200-V/m interfering signal? Can we expect that a receiver with a sensitivity of -120 dBm to function with a signal approximately 180 dB above that sensitivity? Is the receiver front end subject to damage?
The standard appears to have some gaps in this area that will need to be addressed by procedure development or a change notice. Looks like some test-procedure tailoring will be warranted.
DID EMICP (DI-EMCS-80199C)
This EMI Control Procedure DID documenting the contractor's design procedures and techniques is unchanged from the previous revision. Often this document is prepared to satisfy the contract requirements and goes unused during the development, which is unfortunate because a well-prepared control procedure will direct the design for compliance. It also is an ideal document to analyze the requirements, identify tailoring for the planning effort, and support contract modifications to approve tailored requirements. Organizations that truly use the EMICP are far more successful in attaining compliance with little or no changes to the product as a result of the testing.
DID EMITP (DI-EMCS-80201C)
This EMI Test Procedure DID documents the test procedures to be used to evaluate the product for compliance to the standard. Changes place more emphasis on documenting the software for automated testing, incorporating correction factors, and presenting the results.
Like the EMICP, this document is underused. The details provide an understanding of the exact hardware including support equipment and the test configuration and really help define the pass/fail criteria. This is the location where tailoring of the testing is brought to maturity and, upon approval, reduces the battle over issues that manifest themselves during the test.
DID EMITR (DI-EMCS-80200C)
This EMI Test Report DID documents the test results. Changes involve presenting and documenting susceptibilities. The test standard has called for presentation of charts that support a minimum frequency resolution of 1% or twice the measurement bandwidth and amplitude resolution of 1 dB. This has largely been ignored, especially when the results show compliance with significant margin where knowing the exact values is relatively unimportant.
This DID states that the resolution will be required, and more specifically, a single chart cannot be used to present the emissions data. Rather, multiple charts may be necessary to prove that the necessary resolution/accuracy is presented. Presenting a sample in the test procedure for approval prior to testing is recommended.
In addition, the DID change specifies that susceptibility shall be noted with threshold measurements. In reviewing past reports, susceptibility often is noted, but threshold measurements often are omitted. This poses a question: How many frequency points should the threshold be measured over the frequency range where susceptibility is noted? This should be documented in the test procedure.
Conclusion
Remember that the goal is to evaluate the product, make a sensible test, and get valid results. Blind application of the standard without considering unique aspects associated with the EUT results in a situation that may fall short of the goal.
The MIL-STD-461F states in paragraph 6.4, “When analyses reveal that a requirement in this standard is not appropriate or adequate for that procurement, the requirement should be tailored and incorporated into the appropriate documentation, prior to contract award or through contractual modification early in the developmental phase.” So although there are specifics that must be applied, the standard allows judicious tailoring to prepare a viable test procedure and properly perform the test.
About the Author
Steven G. Ferguson is vice president of Operations at Washington Laboratories. He has been working in the compliance test arena for more than 35 years at test laboratories and manufacturing companies designing products, developing procedures, and performing tests. Mr. Ferguson also presents a hands-on course in testing to MIL-STD-461 for multiple government and industrial clients. Washington Laboratories, 7560 Lindbergh Dr., Gaithersburg, MD 20879, 301-216-1500, e-mail: [email protected]