Below are the questions asked during the live event, along with their respective answers.
Q: The diagram you showed with the two frequency dropouts – is that realistic?
A: The diagram with the two spot frequency dropouts absolutely happened. My website blog page has the complete article.
Q: How much deviation should be allowed in a verification check before you redo the calibration?
A: When the verification deviates by 3dB the MIL-STD requires resolution. I expect to see the uncertainty of a dB or so depending on the type of verification (very controlled configurations normally vary by less than 1 dB). Reviewing the data for trends (.5 dB low followed by 1 dB low followed by 1.6 dB low) indicates a problem and should be addressed even though it is still within the acceptable range.
Q: Is it better to test the cables on site or before installation?
A: Cables take a beating in the laboratory with many connect/disconnect cycles and moving between workstations. Running a network analyzer sweep prior to use is fairly quick and is thorough – so do this often and record the data. Let the frequent full checks compensate for reduced calibration intervals. The verification spot checks are still needed as part of the total configuration.
Q: This verification should be performed by laboratory equipment?
A: The verification checks should be accomplished with common lab equipment.
Q: While the background noise (or the receiver noise level) that is 6 dB below the limit is allowed in the standard, how can we perform the verification in emission tests at “Limit minus 6dB” under the presence of this background noise? They are at the same level!
A: This situation does occur but most standard limits allow enough room for the noise to be more than 6dB below the limit. If you are unable to reduce the noise, the need for greater sensitivity, pre-filters, or the like may be identified. Some special limits may require lower noise than is attainable with the standard equipment. When this issue is recognized, reducing the bandwidth of the measurement receiver could help but that could affect the measurement of broadband signals. Sometimes low noise pre-amps or better antennas can also offset the issue. Examination of a specific case may yield information on the details contributing to this issue – I could review a specific case if desired – contact me to discuss.
Q: Can we use the pre-calibration method in the RS103 test instead of the closed-loop method? Is it justifiable and allowed in any part of the standard?
A: RS103 pre-calibration was supported in earlier revisions (prior to revision G) where the field was calibrated prior to placing the EUT. Revision “G” removed that approach. Not many laboratories used the pre-cal method because the sweeps had to be done twice (1- pre-cal without EUT and 2-test the EUT). Note that the support equipment was present for the pre-cal. To use the pre-cal method, the test procedure would need to document that approach and obtain test plan approval. IEC 61000-4-3 uses a leveled field calibration where the EUT plane is calibrated, which MIL-STD-461 does not specifically allow, but it is a good approach that could be used if the test plan approved that method.
Q: What is the best way to verify the modulation, used in CS114 and RS103, is present on the drive signal
A: A simple approach is to examine the output in the tie domain to verify the 40 dB on/off ratio.
Q: CS114 Cal: I had heard that the emission probe is also placed into the calibration setup since high-gain probes can add impedance to the induced current loop, and might affect the resulting injection levels. Adding the probe provides the same impedance in the loop as would be found during actual testing, especially since that impedance is between the injection probe and the EUT. Is that correct?
A: I agree that the probe addition may affect the circuit impedance, but I’m not sure about that being the reason. When moved to the test configuration the probe influence becomes less pronounced because of the variability in the cable impedance.
Q: In regards to the cable pinch problem, how many test frequencies should be checked to verify the cable is okay? If the pinch point has a very narrow frequency response, it seems like a network analyzer may be needed.
A: I agree that a network analyzer would be a better way to examine the cables.
Q: Why no present about RS 103 and RE 102?
A: RE102 was included in the presentation. RS103 commonly used a closed-loop field sensor measurement method that does not require verification – active measurements are used during test. A reverberation chamber calibration was too complex to discuss in the allotted time.