Low-frequency signals appearing on the power lines—so what? The level is low relative to the primary power for an AC-powered device. And DC-powered systems are designed to function with wide variance in the input voltage, so why bother?
Yet the MIL-STD-461F: CS101 Conducted Susceptibility Testing requirement is applicable to all services and applications. Problems are not infrequent, and solutions may be elusive.
The purpose of CS101 testing is to assess the capability of the EUT to maintain the designated level of performance during the presence of interference on the power leads at low frequencies. The power distribution of most facilities and platforms is rife with power frequency harmonic current.
Many products contain nonlinear switched-mode power supplies that contribute to the contamination of the power lines, and variable frequency drive systems are prevalent, providing another source of low-frequency current. A typical PC may produce a very high harmonic distortion as shown in Figure 1.
The testing is performed from 30 Hz (for AC-powered systems, the start frequency is 2x the power frequency) to 10 kHz. The specification is set as a voltage; the effect is to simulate ripple in the power source.
Over large portions of the frequency band, the test circuit may present a very low impedance, so a maximum power is specified if the voltage is not produced by applying a precalibrated drive level. Accordingly, this drive level is defined as a secondary compliance limit.
The Limit
Primary Limit—Test Voltage
Two test curves are presented: one for 28 V or less and the other for greater than 28 V. The nominal operating voltage is used to make the determination. If a unit can operate from 12 to 40 VDC but normally is used in a 28-V system, then the nominal 28-V test limit would be selected.
Secondary Limit—Power Limit
A secondary limit is defined by adjusting the power source to the level that dissipates the power limit curve into a 0.5-? load without producing the test voltage. The power limit curve is in MIL-STD-461F, but at the lower frequencies, the power limit is 80 W and decreases for frequencies greater than 5 kHz. Figure 2 shows the power limit converted to voltage terms (normally used to measure the pretest calibration load voltage).
The Precalibration Test Process
1. Calculate the voltage to be measured across the 0.5-? load that equates to the power level of the power-limit curve. This calculation should be predetermined so the target voltages are known prior to starting the maximum drive calibration.
2. Configure the test equipment as shown in Figure 3 with the 0.5-? resistor across the secondary terminals of the coupling transformer.
3. Adjust the signal generator to the start frequency and the amplitude to the voltage associated with the power limit. The power is 80 W so there is significant heat dissipated. Do not touch the load resistor or place it on a material that could easily burn.
4. Record the power amplifier output voltage. Scan the signal generator over the test frequency range maintaining the drive voltage necessary to produce the power dissipation voltage. This scan will allow the voltage to vary so a decision is needed prior to the scan on acceptable tolerance. The decision will need to allow the load voltage to stay above the minimum calculated dissipation voltage and below the maximum level you will allow.
Tolerance is not clearly defined in MIL-STD-461F for this type of measurement. If the 3-dB measurement system amplitude is used, the voltage would be allowed to vary by more than 41%, but the dissipated power would double. Selection of the 5% resistor tolerance would keep the min-max voltage band very tight with up to 10% variance in the dissipated wattage. This allowance will need to be decided to support the process with either a manual or automated process.
As the scan is accomplished, the waveform must be observed to verify that the signal remains sinusoidal throughout the test frequency range. Examine the oscilloscope for indications of waveform distortion and clipping that would indicate saturation of the amplifier. If an automated system is used, this observation should be an integral part of the precalibration process each time the hardware is configured. Allowing an automated system to set the drive levels without observing the signal is not appropriate.
The Test Process
1. Configure the EUT and test equipment as shown in Figure 4. Before going to the testing, let’s take a moment to review the test configuration. The test requires the application of the interfering signal at the specified voltage across the EUT power input circuit. With the coupling transformer secondary in series with the power phase or positive lead, the current loop includes the EUT, the LISNs, and the power source.
At low frequencies, the voltage drop typically is dominated by the EUT. But as the test frequency is increased, the inductive reactance of the LISNs increases, and the voltage drops in the LISNs. This causes the EUT exposure to be low but not from a low-impedance power input circuit of the EUT.
To compensate, the 10-µF capacitor across the power lines tends to short-circuit the LISNs and provide for the voltage drop to be at the EUT. In other words, the absence of the capacitor represents a significant undertest.
The oscilloscope is powered through an isolation transformer to prevent grounding of the EUT. This places the oscilloscope chassis at the voltage of the reference power input. If the neutral is at an elevated voltage, such as a phase lead of a 3-phase input, the oscilloscope becomes a live circuit, and contact will result in a shocking test experience.
2. Power-on and start operation of the EUT. This may be easier said than done because the transformer is in line with the power circuit and represents a reactance limiting the in-rush power. Also, the primary may be an open circuit and the reflected impedance exaggerating the impedance to the in-rush current.
A low-value resistor of approximately 5 ? across the primary may allow the start-up current for the EUT. However, the situation may not be complete if the EUT current is relatively high. The power frequency fed back across the transformer, secondary to primary, is a function of the current flowing in the circuit, I1 = I2N where N is the turns ratio and may produce damage to the power amplifier or at least a shutdown of the power amplifier if it is protected.
3. Now that the EUT is operating and the power amplifier able to output a signal, the testing shall begin. While monitoring the EUT for indications of susceptibility, set the signal generator to the start frequency and then adjust the output amplitude until the lesser of the test voltage or the precalibrated drive is reached. Assuming that the EUT is not susceptible, scan the test frequency range at the scan rate specified in MIL-STD-461F, maintaining the test voltage or precalibrated drive.
4. Document the test results presenting the applied voltage and frequencies at which the test was conducted. Provide any susceptibility indications and threshold measurements.
5. Repeat the test for each phase lead identified for test. Testing with the coupling transformer in the neutral is not applicable because the testing places the test voltage across the EUT power leads.
It can be difficult to measure the low-level signal in the presence of the power frequency at high levels, especially at the lower test frequencies. Figure 5 shows a power filter that works with a 10x scope probe and 1-M? oscilloscope input. It is designed to attenuate the power line frequency. This circuit must be calibrated or characterized to determine the insertion loss across the frequency range of the test.
Threshold Measurements
If the EUT is affected by the susceptibility test signal, it is necessary to determine the threshold of susceptibility. MIL-STD-461F provides guidance on threshold measurements:
•?At the frequency of interference, lower the signal amplitude until the indication of susceptibility is not present.
•?Reduce the signal an additional 6 dB.
•?Gradually increase the amplitude until the indication of susceptibility reoccurs. This is the threshold of susceptibility. Record the threshold level and the indication of susceptibility.
This process generally works, but if the failure indication is a curl of smoke, analysis would be needed to determine the failure mode.
As often happens, EUT may be susceptible over a wide frequency range. The question is posed: How many frequencies should be measured? This guidance is not part of the standard, so the laboratory needs to establish an approach to making a sufficient number of measurements that envelopes the frequency range where susceptibility occurs and identifies the lowest threshold level.
Consider an approach of measuring thresholds at the frequency where susceptibility is first noted, highest frequency, lowest threshold, and at least three frequency points per octave over the range of susceptibility. In addition, other points of inflection on the threshold curve may be needed. If multiple indications of susceptibility are present, then each indication needs to be addressed.
Countering Feedback
High AC current flowing through the coupling transformer will induce a voltage in the transformer primary. This often faults or damages the injection source.
The solution is to negate the feedback with another transformer. The feedback is relatively simple, using a second coupling transformer with the secondary connected to a load simulator drawing current that is near the EUT current (Figure 6). The difference is that the transformer is out-of-phase with the EUT coupling transformer. This out-of-phase provides a net cancellation of the power frequency fed back into the power amplifier.
However, two issues may present difficulty in implementing this control:
•?The high-current EUT plus the additional offsetting current may be too much current for the circuit.
•?If the EUT presents significant harmonic distortion, the load simulator will not produce the cancellation desired. If the feedback control method will not be effective, tailoring of the test may be necessary, including consideration of another coupling method that meets the intent of the test.
Conclusion
MIL-STD-461F CS101 testing lends itself to automating the test process. But it doesn’t imply that a recorded precalibration should be played back without regard to the lesser of the two limits.
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]
January 2009