By modern era is meant the realization that for such near-field RE measurements to be useful, the sensor/pickup/antenna must closely model the actual victim protected by the EMI standard in question, in physical appearance, orientation and separation from the victim.
Prior to 1953, various sensors were used somewhat indiscriminately, with small attention paid to repeatability and correlation between equipment-level RE measurements and vehicle-installed electromagnetic compatibility (EMC).
The 104 cm (41 inch) rod antenna is conceptually one of the simplest devices in our arsenal of measurement tools, but it is a sad fact that it is poorly understood today.
Most EMC engineers understand intuitively that the rod antenna sums the electric field components parallel to its length, and the radio frequency (rf) potential available at the base is one-half the potential difference between the rod base and top. It is less well-known that the potential at the rod base is not measured as an absolute, but like any potential measurement, the measurement is a potential difference – in this case, the potential difference is between the rod base and the potential of local ground near the base, meaning the counterpoise. The counterpoise potential is often considered to be a zero potential, but in fact it is not, by virtue of being exposed to the same field as the rod. It is this latter fact, amongst others, which is addressed by the MIL‑STD-461F rod antenna configuration change, and ongoing efforts to develop MIL‑STD-461G.
After a historical retrospective, misinterpretations and errors, up to and including an unfortunate description in this year’s EMC Symposium Record, are reviewed and explained.
The historical discussion is taken largely from a monograph on the same subject available in the “History” section at www.emccompliance.com, and an article by the author entitled, “On the Nature and Use of the 1.04 m Electric Field Probe” [1] – henceforward Javor 2011. Javor 2011 goes into depth on the physics and math of electric field coupling to a 104 cm rod, whereas this article is qualitative and simply references the analytical and test results demonstrated back in 2011. Unless otherwise noted, the test data and test set-up photographs used in this article were borrowed from work done developing Javor 2011.
The Beginning
29 May 1953 was the release date of MIL‑I-6181B, “Interference Limits, Tests and Design Requirements, Aircraft Electrical and Electronic Equipment,” the first to adopt the use of the 104 cm rod antenna. The first use of the 104 cm rod antenna is explained in NADC-EL-5515, dated 10 August 1955, “Final Report, Evaluation Of Radio Interference Pick-Up Devices And Explanation Of The Methods And Limits Of Specification No. MIL‑I-6181B.” This report was essentially a rationale appendix for MIL‑I-6181B.
Before, during, and after WWII, right up to 1953, standard practice for connecting a communication radio to an antenna made use of unshielded wire that was essentially a continuation of the external antenna (Figure 1). The antenna connection was just as sensitive to rf as the antenna itself, and within the vehicle, it was exposed to numerous sources of radio frequency interference (rfi).
Figure 1: Antenna connection wiring run inside of WWII-era bomber (National Air & Space Museum, Washington DC). The bare wire is covered by steatite beads, which provide the required insulation when the antenna is transmitting. In transmit mode, as much as 5 kV potential was on this wire, at medium and high frequencies!
The input impedance of radio mixer tubes of this era was very high, and it was necessary to separate the antenna lead wire from aircraft structure to limit capacitive loading. That is the function of the porcelain standoff highlighted in Figure 1. The high impedance wire was very susceptible to capacitive crosstalk, and pains were taken to keep it separate from other wiring. Notice the separation of the wiring along the top of Figure 2 suspended on the porcelain standoff from other cable assemblies.
Figure 2: Separation between unshielded antenna lead and closest adjacent wire bundle (National Air & Space Museum, Washington DC). Wire separation was the only available control on crosstalk in these days of open wires and minute signals in receive mode and extremely dangerous potentials in transmit mode.
The rfi problem was well understood, as evidenced in the Figure 3 drawing excerpted from a WWII-era War Department Technical Order.
Figure 3: Drawing showing coupling of aircraft internal rfi sources to an internal antenna lead in, and how to minimize crosstalk. It is quite clear that rfi occurs internally to the aircraft, and that coupling to the antenna itself isn’t even an afterthought! – from the 1945 “Handbook of Elimination of Radio Noise in Aircraft” which in turn was an update of a similar 1942 publication (United States War Department and the Air Council of the United Kingdom).
The long-term fix for the resultant rfi was to eliminate the unshielded antenna lead-in from future procurements. MIL‑I-6181B banned such procurements, replacing them with radios compatible with coaxial cable. But there was a very large inventory of the older radios, and the aircraft that had them installed, and so MIL‑I-6181B still had to protect (grandfather) those installations. William Jarva, the author of NADC-EL-5515, picked the 104 cm rod antenna as provided with the Stoddart Aircraft Radio Company AN/PRM-1 meter (new at the time) as a reasonable simulation of the unshielded antenna lead-in at frequencies below 20 MHz. Above 20 MHz, a horizontally polarized tunable dipole was used.
NADC-EL-5515 describes a measurement made by Mr. Jarva to develop a radiated emission limit to protect the BC-348Q radio installation. Figure 4 is a re-enactment of this set-up.
Figure 4: Re-enactment of the NADC-EL-5515 set-up used to create the RE limit in MIL-I-6181B in 1953. With this set-up, there was a nearly one-to-one correlation between failing an EMI requirement and causing an EMC problem in a vehicle. Once the problem of open-wire lead-ins had been solved, the correlation was much lower. In fact, MIL-E-6051D, a system-level EMC standard released in 1967 cautioned thusly: “Unless otherwise specified in the contract, subsystems/equipments shall be designed to meet the requirements of MIL-STD-461 and MIL-STD-462. Since some of the limits in these standards are very severe, the impact of these limits on system effectiveness, cost, and weight shall be considered.”
A full description of the measurement is to be found in the aforementioned articles. For the purpose of this retrospective, it is sufficient to note that the AN/PRM-1 EMI receiver in the foreground was battery-powered and the only connection to the receiver was a short bond strap to the tabletop ground plane. Further, the 104 cm rod emanates directly from the EMI receiver; there is no intervening cable. In this way, the EMI receiver very closely simulated the period-piece BC-348Q radio placed on the ground plane that was the victim protected by the RE limit. Impulsive noise (represented by the impulse generator) coupled equally to the unshielded antenna lead connected to the BC-348Q and the 104 cm rod antenna. The impulsive noise source was placed equidistantly from the BC-348Q antenna lead and the 104 cm rod antenna, and the separation was one foot, as opposed to one meter today. This reflected the separation achievable between culprit and victim wiring in the aircraft of the time. When rfi was detected by listening to the BC-348Q headset, the meter deflection was noted on the AN/PRM-1 meter, and a limit was built in terms of the rf potential measured by the meter (Figure 5), as opposed to the modern day practice of measuring a field intensity. Such a limit is termed “antenna-induced” (dBuV) as opposed to field intensity (dBuV/m). An antenna-induced limit necessarily requires tight specification of the antenna. Modern one-meter military RE limits control field intensity, but still retain control of antenna type, vs. RE measurements specified at three meters or more, which do not. This reflects once again the difference between near and far field measurements.
Figure 5: Limit determination for radiated emissions below 20 MHz in NADC-EL-5515 and MIL-I-6181B (note inherently broadband units – all sources used were broadband, reflective of electrical culprits at the time)
The following excerpt from NADC-EL-5515 (available from the emccompliance.com history page) explains the physics of the situation in Mr. Jarva’s own words:
“ANTENNA SYSTEMS FOR RADIO INTERFERENCE MEASUREMENTS
In the frequency range 0.15 to 20 mc, radiating elements, pick-up antennas and distances, generally used for radiated radio interference measurements, are small compared to wavelength. The amount of energy transferred from field to antenna depends on the nature of the signal source and the type of receiving antenna used. For instance, if the radiating interference source is a single, small closed loop of wire, a great deal of current can flow without developing much voltage across the loop. Consequently, a large magnetic component is developed in the induction field in conjunction with a comparatively small electric component. To extract a large amount of energy from such a field, a similar loop antenna, correctly matched to a receiver, should be used as the pick-up device to provide what may be compared to a good impedance match in ordinary circuit theory. If a short rod antenna, sensitive to the electric component of the field, were used as the pick-up device very little energy transfer would result and a situation comparable to a condition of impedance mismatch would exist. When a short rod antenna is the signal source, a large voltage can be developed on the rod, but with very little current flow. Consequently, the field developed is composed of a large electric component and a small magnetic component. In this case, another rod used as a pick-up device would indicate the presence of an intense field, whereas, a loop antenna would give very little indication. Typical radio interference sources in aircraft include the extreme cases described and all other variations. In general, the ratio of the electric to the magnetic components surrounding an unshielded lead will vary directly as the impedance of the load terminating the lead, and the apparent impedance presented to the various pick-up antennas will vary in the same manner. This statement applies to radial and tangential field components as contrasted with the more usual concept of wave impedance encountered in shielding theory, which applies only to the components tangential to the line of propagation.
Although it would be desirable to require the measurement of both the electric and magnetic components of the interference field, it is felt at the present time that such requirements would make specification testing excessively complex. Experience has indicated that aircraft electronic equipments, which operate in the lower frequency ranges (0.15 to 20 mc), are more sensitive to the electric field because of the unshielded high impedance antenna lead-in, which has been in general use. Present practice is to control the electric field by radio interference measurements. This is done by utilizing a 41-inch rod antenna and treating any difficulties arising from equipments generating strong magnetic fields as special cases which require particular attention when the equipment is installed in the plane. Reference (e) requires that all equipment used with antennas be designed for use with a shielded antenna lead. If and when the unshielded antenna lead is completely eliminated from use in aircraft, a review of present methods and limits in the frequency range 0.15 to 20 mc will be required. Radio interference meters using the 41-inch rod antenna are so constructed and calibrated that they read directly the microvolts which are induced in the antenna by the interference field.”
Note: The reference (e) cited is MIL‑I-6181B.
The above excerpt is remarkably lucid and showcases how well understood the problem and the solution were. The reader can demonstrate the electric field nature of the rod. Set up a wire above ground per MIL‑STD-461E/F, driven at one end by a 50 Ohm signal generator. Load the other end of the wire with 50 Ohms, and place a 100 kHz, 100 dBuV amplitude signal on the wire. Record the measured field intensity (~1 mV/m, 60 dBuV/m). Now remove the 50 Ohm load from the far end of the wire. The wire potential will increase 6 dB due to being unloaded, so decrease the signal generator setting by 6 dB to keep the wire potential constant. The rod antenna measurement will indicate precisely the same field intensity reading as previously – despite that around 80 dB less current flows in the second configuration! A clearer demonstration of electric field sensing and magnetic field rejection cannot be had.
With that background as to how the 104 cm rod antenna came to be used for EMI testing, we progress to its implementation in the EMI test chamber. Figure 6 is a diagram and recreation of a rod antenna set-up from MIL‑I-6181.
Figure 6: MIL-I-6181 rod antenna diagram and set-up recreation. As noted in the text, this is so near field that it isn’t even an attempt to measure a field intensity; instead the potential induced in the rod is measured and is a crosstalk control, or very close to one.
In Figure 6, the rod connects directly to the EMI meter. As bands were selected, the antenna was internally properly matched to the mixer input. The mixer tube presented a high input impedance, so that the 104 cm rod was not loaded as it would be by a modern mixer with an input impedance more nearly approximating fifty Ohms.
Note also the very short bond strap between EMI meter and ground plane. The rod antenna was only 12” from the test sample front face. That reflected achievable wire separation in that era’s aircraft. The purpose of the bond strap was to make the ground plane the reference for the rod antenna’s pickup potential. The EMI meter was battery-powered in this application; the ground plane is the sole ground reference.
As time went by, complaints arose about the difficulty of using the AN/PRM-1 meter in immediate proximity to the test sample. While a remote meter was provided with the AN/PRM-1, the controls still had to be adjusted on the meter face itself. Stoddart Aircraft Radio Company then provided a more modern version of the rod antenna, with its own base, passively tuned. This allowed remote use of the EMI meter itself. Figure 7 shows a set-up using the rod antenna with its own base in a picture of an EMI test from the 1950s or early ‘60s.
Figure 7: Picture of MIL-I-6181 RE test using a rod antenna prior to 1963. Now the engineer could be removed from the set-up. The connecting cable was shielded twin-axial transmission line, so if desired it could be run through a bulkhead feedthrough grounding the shield without introducing a ground loop into the instrumentation!
MIL‑I-6181B specified the rod antenna only to 20 MHz, with dipoles used at higher frequencies. Later revisions pushed the breakpoint to 25 MHz (viz. 25 or 30 MHz for MIL‑STD-461/-462, all versions).
Tri-Service Standardization
The Army, Navy and Air Force had their own Service-unique standards up to 1967. The purpose of MIL‑STD-461/-462 was to provide a single Tri-Service standard, with attendant economies of scale.
MIL‑STD-462 placed the rod antenna at a one-meter distance from the test sample, and floated the counterpoise from the ground plane (Figure 8). This was per the NADC-EL-5515 observation (quoted earlier) that as the open-wire lead-in radios were phased out of use and replaced by 50 Ohm coaxial input radios, the test method using the rod antenna at 12” would have to be revisited. The use of modern coaxial-shielded lead-ins moved the sensitive high impedance unshielded victim to the aircraft exterior where the antenna was mounted. Increasing the antenna-test sample separation was a response to the new radio-to-antenna connection technique. The counterpoise was only grounded through its coaxial connection to the EMI receiver, which was very important because at 14 kHz, the bottom of the RE02 band, a single point ground was necessary for measurement integrity when using a passive (octave-band tuned) rod antenna.
Figure 8: Rod antenna use per MIL-STD-462 basic release (1967). Many of the approved EMI meters at this time could be run on battery power, so that this set-up did not inherently ground the isolated counterpoise.
Another change related to the removal of the sensitive victim wiring from within the aircraft was the consequent attention placed on protecting the antenna from rfi. This resulted in a change from the antenna-induced limit to a modern field intensity limit.
A grey cloud appeared with the silver lining of removing the highly sensitive unshielded wire from within the vehicle. The easily modeled and universal internal interaction devolved into a vehicle specific geometry where mostly internally generated fields had to find their way out of the metal vehicle to interact with an external antenna. Now failing to meet the RE102 limit is not a cause for immediate rejection, but something that needs to be evaluated by installing the device on the applicable vehicle and checking compatibility. This is recognized in MIL‑STD-464 paragraph 5.2.4, which requires quantitative measurement of rfi coupled to vehicle antennas from vehicle electronics. Such measurements have been made even before MIL‑STD-464 basic was released in 1997, but MIL‑STD-464C made it a hard requirement in 2010.
MIL‑STD-462 also replaced tunable dipoles with the 1.37 m tip-to-tip biconical antenna above the rod band. Whereas MIL‑I-6181 and similar specifications required a dipole tuned as low as 28 or 35 MHz (on the order of 5 meters end-to-end), the shorter biconical can be tilted to be used vertically as well as horizontally. Pre-MIL‑STD-461 EMI standards only required control of horizontally polarized coupling or fields, above 20-30 MHz, but MIL‑STD-461/-462 controlled both polarizations above 30 MHz (log-spiral antennas were used above the biconical band, which captured both horizontal and vertically polarized fields simultaneously).
And finally, it was about this time that active rod antennas became commonly available. This was a technology development, not a specification or standard requirement. Instead of tuning a rod antenna through octave bands that tracked those of the remote receiver, the rod antenna drove a FET gate that acted as a near open-circuit load. This means that the 104 cm rod antenna’s inherent open-circuit effective height of 0.5 meters (or 6 dB/m antenna factor) was achievable. Compared to tuning out the rod’s 10 pF source impedance with one inductor per octave, the improvement in antenna factor was on the order of 50 dB at 10 kHz. This development facilitated the use of spectrum analyzers for EMI testing when they became available. The analyzer’s sensitivity wasn’t as good as that of the EMI receiver, but it didn’t need to be, using an active rod antenna. The downside was that the active circuitry placed a limit on dynamic range for high-level signals, both for the rod and spectrum analyzer electronics. Response to a broadband signal could be quite limited, and if there was a strong out-of-band signal, that could diminish the ability to receive a low-level signal. The latter issue wasn’t as important inside a shielded test chamber.
A Problem Creeps In
In 1970 and 1971, Notices 2 (Air Force) & 3 (Army) were released. One area of commonality between Notices 2 and 3 was a change to the rod antenna configuration: whereas previously the counterpoise was floated from the ground plane, now it was bonded to it. This change found its way unimpeded into MIL‑STD-462D (1993) and the consolidated MIL‑STD-461E (consisting of both requirements and procedures) (1999). Notice 2 wording is as follows: “4. Paragraph 4.2.3.2 Add this sentence: When a counterpoise is used with a rod antenna, it shall be bonded to the ground plane with a strap at least 30 cm wide.” Note that the 1967 set-up is similar to that of the 2007 MIL‑STD-461F RE102 rod antenna with the exception of the lack of the lossy ferrite bead increasing the impedance of the bond path. The 1970/1971 changes were a mistake, but it took three decades to realize it.
Recall that before MIL‑STD-462, there was a change in antenna type and polarization at somewhere between 20-30 MHz, depending upon the particular specification and vintage. The efficiency of the vertical rod and horizontal dipole at 30 MHz was quite different, so that the Figure 9a antenna-induced limit on rf potential was discontinuous at the breakpoint, and the signatures were as well. But with MIL‑STD-461 going to a field intensity limit, and MIL‑STD-462 requiring both horizontal and vertical polarization of the biconical antenna, it is reasonable to expect some degree of continuity at the antenna change breakpoint for vertical biconical polarization. In fact, the RE02 and RE102 limits of all versions of MIL‑STD-461 are continuous at the breakpoint (Figure 9b). The slope may be changing, but the limit amplitude is continuous. But after Notices 2 and 3 were released, it was not always the case that the signatures were continuous at the antenna breakpoints even for vertical biconical polarization. This is even more obvious if an overlap of data is taken between 20-30 MHz, where both antennas are calibrated for operation. Another and related issue that comes to light is that a surprising number of totally different test items all seem to have a broad peak between 20-30 MHz.
Figures 9a & b: MIL-I-6181 antenna-induced emission limit at left, showing discontinuity between vertical rod and horizontal dipole rf output at 25 MHz vs. superseding MIL-STD-461 RE02 limit (1967). Note how corresponding portions of both limits have similar slopes, reflecting the effective heights of the vertical rod and tunable dipole. One other change that should have occurred in 1967, but didn’t until 1993 (MIL-STD-461D) is that with the external antenna as the focus of radiated emission control, there should have been separate limits for equipment installations depending on whether they were in or outside of a metallic vehicle.
Problem Recognition
Messrs. Steve Jensen and Luke Turnbull separately identified shortcomings of rod antenna measurements in the last octave of use in 2000, and 2007, respectively. These were critiques against MIL‑STD-461E RE102, below 30 MHz and similar automotive test standards. The issue was a great discrepancy between measured fields at the 30 MHz breakpoint between the rod antenna and the biconical antenna, vertically polarized. While one would not expect precise agreement, due to significantly different physical apertures, the 20 dB difference in the data below is problematic. Mr. Jensen showed by overlapping the biconical antenna and rod antenna from 20-30 MHz that the biconical always returned much lower levels.
Mr. Jensen flagged this issue in a critique of the draft MIL‑STD-461F dated 23 March 2007. Air Force EMI personnel separately along with others had also noted enhanced levels of any noise present in the 20 to 30 MHz frequency band. In conjunction with the Tri-Service Working Group (TSWG), the Wright-Patterson Air Force Base (WPAFB) EMI laboratory undertook a detailed study of the rod antenna setup. John Zentner and Steve Coffman with participation by the author performed the work. John Zentner had been deeply involved in the development of the “D” revisions of MIL‑STD-461/462 and was the chairperson of the MIL‑STD-461E TSWG. Steve Coffman was the Air Force EMC engineer for special operations aircraft with 30 years of experience in the EMC field. The result of the effort became the basis of the setup changes for the rod antenna introduced in the final version of MIL‑STD-461F.
Figure 10: Data from Jensen, Steve. “Measurement Anomalies Associated with the 41 Inch Rod antenna when used in Shielded Enclosures,” dated 17 July 2000. Once you know to look for this, you will see it over and over again – unless you are working to MIL-STD-461F or later. While modern versions of RTCA/DO-160 no longer use the rod antenna, DO-160C rod use was no different than that in MIL-STD-462. Modern DO-160 replaces RE measurements with a common mode current control below 100 MHz. Recall the rod antenna’s electric field response and magnetic field rejection. Does DO-160F/G really control electric fields below 30 MHz from a high impedance cable?
Problem Resolution
Various configurations with different size counterpoises and grounding/bonding techniques were explored at Wright Patterson AFB. It was found that a resonance condition with the MIL‑STD-461E/462 configuration caused the potential of the counterpoise to rise to a level that swamped the potential induced in the rod antenna itself. Measurements from various configurations were compared. The configuration that produced the best results was a traditional size counterpoise that was not bonded to the bench top ground plane, closer to the floor than the tabletop ground plane, and with a short coaxial cable electrically grounded to the shielded room floor. Due to a remaining resonance between the counterpoise capacitance to the floor and the coaxial cable inductance, a lossy ferrite sleeve was applied to the coaxial cable to dampen the resonance. The required impedance of the ferrite sleeve was defined in the main body of MIL‑STD-461F and a statement was included in the Appendix of the standard that a ferrite sleeve “lossy with minimum inductance” should be used. The results of the study were presented in public forums at the 2007 IEEE EMC Symposium and at the 2008 Department of Defense (DoD) Electromagnetic Environmental Effects (E3) Program Review.
The pre-461F parallel LC trap formed by the counterpoise over the floor and the coax cable connected at one end to the rod antenna base and at the other to the chamber wall caused the impedance between counterpoise and chamber to increase greatly; hence the same field that coupled to the 104 cm rod was able to increase the potential of the counterpoise as well. The impedance between counterpoise and chamber in the absence of any detuning at WPAFB is shown in Figure 11.
Figure 11: Classical parallel L – C trap impedance (yellow is amplitude, blue is phase). Every so often theory and practice coincide so perfectly that the term “textbook” just begs to be used.
Javor 2011 showed the result of detuning this resonance (Figure 12). That effort used a one-meter long wire suspended 5 cm above ground driven and loaded by 50 Ohms as the electric field source. The voltage on this wire was constant vs. frequency, and the article provides the quasi-static physics and math to demonstrate that the resultant electric field as measured by the rod antenna should also be flat vs. frequency, so that any departure from flatness is a measurement error.
Figure 12: Pre-MIL-STD-461F resonance on left; MIL-STD-461F lower-Q resonance on the bottom (both plots have same reference level, are 10 dB/division and cover 2 – 32 MHz). Measurement technique fully described in Javor 2011. The low frequency peak on the top is due to an extremely long coaxial cable being used – a variety of lengths were tried to see the effect before your author figured out what was happening.
Looking Ahead
It is clear that the MIL‑STD-461F technique went a long way towards eliminating a large error source. It is also clear that it is not perfect. Javor 2011 showed the result of completely floating the counterpoise (eliminating, not detuning, the resonance), and that gave a nearly flat result, and more importantly, almost exactly the same result as when both test sample and rod antenna are both referenced to the chamber floor as in Figure 13. A US$50 Mini-Circuits isolation transformer was used in lieu of the ferrite sleeve.
Figure 13: Both the radiating element (wire above ground on the left) and the measurement antenna are referenced to the shield room floor, ensuring a common potential for the measurement and the absence of any sort of resonant condition. Above ground rod antenna measurements such as MIL-STD-461, RTCA/DO-160 (obsolete versions), CISPR 25 et al. should all produce results commensurate with a common ground plane measurement. It is the gold standard.
Figure 14 shows measurement results when the source and the rod antenna are both referenced to the shield room floor as in Figure 13. The flatness of the measurement is very close to perfection. Figure 15 compares the -461F result (top trace) to using an isolation transformer (lower trace) in lieu of a ferrite sleeve in the -461F set-up. The Figure 15 isolation transformer lower trace response is very close to that of the Figure 14 floor-based measurement.
Figure 14: Field intensity measurement results when radiating wire and rod antenna are both referenced to test chamber floor per Figure 13
Figure 15: Upper trace is -461F result; lower trace is result using isolation transformer in otherwise -461F set-up
It should be clear that the isolation technique of eliminating the resonance is potentially superior to the detuning technique, but a few hurdles remain. These include transmission line transformer vs. a true isolation transformer, and the efficiency (loss) associated with the transformer. With a 24 dBuV/m limit above 2 MHz, not much loss is acceptable. Members of the MIL‑STD-461 Tri-Service Working Group have been working this problem since 2011. There is a separate practical motivation for using isolation instead of detuning. The -461F technique requires grounding the coax shield directly below the rod antenna. Many test facilities don’t have a readily accessible floor grounding point available everywhere, due to various coverings that are sometimes used over the metal floor, such as tile or concrete. An isolation technique eliminates the need for grounding. That advantage has many people interested in this approach entirely separate from a desire for better test data.
A Concern with Misinformation
Unaware of the body of work resulting in MIL‑STD-461F and the continuing work by the TSWG based on Javor 2011, Mr. Harry Gaul of General Dynamics published and presented an article on the same subject in/at the 2013 EMC Symposium in Denver (Gaul, Harry. Electromagnetic Modeling and Measurements of the 104cm Rod and Biconical Antenna for Radiated Emissions Testing Below 30 MHz. 2013 IEEE EMC Symposium Record. Denver, CO). Gaul 2013 comes to the same conclusion as Javor 2011; namely that counterpoise isolation is superior to detuning a resonant circuit. The Gaul 2013 approach was entirely different than employed in Javor 2011, and it is reassuring that the two entirely different techniques ended up with the same conclusion. Mr. Gaul used a method of moments code called FEKO, whereas Javor 2011 performed a closed form analysis calculating the coupling from the electric field from a wire above a ground plane to a 104 cm rod antenna, based on first principles – quasi-static electric field formulation based on Gauss’ Law. Both efforts compared the predictions with measured test data, but that is where the similarity ends.
Both Javor 2011 and Gaul 2013 evaluated several different antenna configurations between 20 – 30 MHz. These are complete isolation of the counterpoise, -461F, -461E, -462, and floor mounting of the rod antenna, as well as use of a vertical biconical.
Your author, a member of the Tri-Service Working Group on MIL‑STD-461F, was aware of the precise damping ferrite described in the standard, and the test results in Figure 12 verify how well it works. MIL‑STD-461 cannot identify any commercial product or service by trade name, and instead has to specify a device by its salient characteristics. Mr. Gaul unwittingly identified a (non-technical) flaw in MIL‑STD-461F: part of the ferrite description is not in the main body of the standard, but in the appendix, which is not contractually obligatory. In the main body, the ferrite is identified as having an impedance of 20 – 30 Ohms at 20 MHz, and in the appendix the following statement is made: “Floating the counterpoise with the coaxial cable electrically bonded at the floor with a weak ferrite sleeve (lossy with minimum inductance) on the cable produced the best overall results.” One ferrite that meets this requirement is the Leadertech (used to be Ferrishield) CS28B1642 using 28 material. Its measured impedance (courtesy of Mr. John Zentner at WPAFB) is listed in Table 1.
Table 1: Leadertech CS28B1642 ferrite bead impedance. The highlighted row shows the critical parameter at the frequency specified by MIL-STD-461F: the resistive impedance is higher than the inductive reactance. Note that this model has the two impedance components in series, which is standard in the industry, not in parallel as in Gaul 2013.
Note that at 20 MHz the impedance is higher than the MIL‑STD-461F requirement; the requirement was based on the manufacturer’s data, not measured data. But the resistance (real Ohms) is higher than the inductive reactance (imaginary Ohms), as required, and this provides the necessary damping performance.
In contrast, Gaul 2013 uses the following analytical model for the ferrite sleeve, which provides no damping whatsoever:
“The ferrite bead (when used for the MIL‑STD-461F configuration) is modeled as a parallel circuit of 480 ohms, 0.255pF, and 250nH to match the characteristics of the actual bead used.”
At 20 MHz, 250 nH provides about 30 Ohms inductive reactance, but as noted, the parallel 480 Ohms provides no damping at all. The effective impedance is still 30 Ohms inductive. Unfortunately, based on his misunderstanding of the MIL‑STD-461F approach, Gaul 2013 states that:
“The MIL‑STD-461F test setup had the worst agreement with about 18dB difference.”
That pejorative conclusion is unsurprising given an assumption of purely inductive bead impedance: it doesn’t take a sophisticated computer program to determine that adding inductance to the inductive leg of an LC trap simply reduces the tank frequency, without reducing the circuit “Q.”
Misunderstanding of the purpose of the ferrite sleeve is distressing. The TSWG thought it would be clear that the purpose of the sleeve is for damping, based on the overall wording in MIL‑STD-461F. If damping is retained over isolation, future versions of MIL‑STD-461 will certainly clarify this issue in detail.
Conclusion
In today’s world, there is an expectation that technology improves over time. With respect to rod antenna measurements over the last threescore years, we might instead be inclined to quote Jerry Garcia and say, “What a long, strange trip it’s been.”
The radiated emission problem was well understood six decades ago, and they had both a test and design solution in hand. But when the design solution took hold, the test solution changed and became problematic. We have spent the last several decades in a sort of mini-Dark Ages, where we were not doing as good a job as when we began, and weren’t even aware of it. But thanks to Steve Jensen, Luke Turnbull, and the DoD TSWG, we have turned the corner on a mini-Renaissance and now (in MIL‑STD-461F) have a much better test, and for the last several years have been pursuing an even better approach for -461G (counterpoise isolation).
Aside from the 60th anniversary aspect, the motivation to write this article is due to multiple interactions with people reacting to the MIL‑STD-461F change without adequate understanding of the background. One objection was that the -461F configuration lowers measured emissions from previous configurations. That is true (especially in the resonance band) but that is also the point – more accurate measurements reveal that the true field is lower than previously measured. Careful measurements at WPAFB revealed that when the resonance is removed via -461F, a previously masked dropout is visible, but that is because of a destructive interference node due to room dimensions and inadequate absorber performance. The MIL‑STD-461F Table I absorber requirement is a considered compromise between rf performance and economic impact. If the dropout is a concern, the solution is better absorber (hybrid based on a ferrite tile foundation), not maintaining a counterpoise resonance to offset a chamber-induced destructive interference. And of course the frequency coincidence of these two resonances only works for specific room dimensions, not in general.
The author wishes to thank the following EMC engineers for above-the-call-of-duty reviewing of this article. John Zentner and Steve Jensen, both mentioned and identified within the article, also took the time to review this effort. Mark Nave of Mark Nave Consultants, Inc., Vince Sutter of Raytheon, and Tim Travis of ASRI all contributed to making this article a more user-friendly reading experience. To the extent that it isn’t, the fault lies entirely with the author.
Reference
- Ken Javor, “On the Nature and Use of the 1.04 m Electric Field Probe,” Interference Technology Engineering Master (ITEM), 2011.
 |
Ken Javor has worked in the EMC industry over thirty years. He is a consultant to government and industry, runs a pre-compliance EMI test facility, and curates the Museum of EMC Antiquities, a collection of radios and instruments that were important in the development of the discipline, as well as a library of important documentation. Mr. Javor is an industry representative to the Tri-Service Working Groups that write MIL-STD-464 and MIL-STD-461 (the “G” effort presently underway). He has published numerous papers and is the author of a handbook on EMI requirements and test methods. Mr. Javor can be contacted at ken.javor@emccompliance.com. |
