In 2006, APELC received a request from Lockheed Martin for a MIL-STD-461, RS-105 EMP test system. After analysis of the MIL-STD, we decided this was something well within our wheelhouse and decided to give it a go. Many lessons were learned along the way, and we continue to learn and apply new lessons in the design of our high-altitude nuclear EMP (HEMP) simulators for RS105 and MIL-STD-2169. Below is an abbreviated version of two papers we wrote and presented at the IEEE Pulsed Power Conference in both 2009 and 2019. Hopefully, this gives insight into challenges involved with this type of test, how we overcame it, and some basic principles of HEMP simulators and testing.
We based our initial design by following the definition in the MIL-STD demonstrating a parallel-plate TEM line. We later learned this was not a good idea, and is likely something that needs updating in the standard. A brief explanation as to why a parallel-plate system is not a good idea is presented below.
APELC 1st Generation RS-105 Test System
A pulse is launched onto a parallel plate, TEM structure, which is resistively loaded to avoid unwanted reflections. The standard defines the characteristic waveform, shown in Figure 1, which is a double exponential pulse, characterized by a 1.8 – 2.8 ns rise time, and a Full Width Half Maximum (FWHM) of 23 ns ± 5ns. A peak electric field strength of 50 kV/m is required [1].
Figure 1. The required waveform necessary to perform the MIL-STD 461G (RS105 test method).
Figure 2. An illustration of unwanted waves generated by the suggested geometry.
The geometry described in the standard presents problems posed by reflections, caused by all the bends in the structure. As illustrated in Figure 2, major bends throughout the structure create diffracted waves off the corners located near the input, and reflected waves off the corners and back wall near the load. The unwanted waves distort the primary signal, often pulling the system out of specification.
This is illustrated in the resultant waveform from the original system shown in Figure 3.
Figure 3. Sample waveform from the original parallel-plate RS105 system
APELC 2nd Generation RS-105 Test System
While the waveform above does meet spec, this is only due to the placement of absorber material in the test volume to correct the worst of the reflections. As a more permanent solution, we eventually developed a system with a semi-transparent load taper using rods instead of a metal plate. We also, developed a means of directly feeding the pulse onto the TEM structure without the use of a coaxial to parallel plate converter. The resulting system and waveform are shown in Figures 4 and 5 below.
Figure 4. An illustration of the 2nd generation RS105 test structure.
Figure 5. A sample waveform measuring the electric field inside the test volume of the 2nd generation RS105 test system.
As can be seen in Figure 5, the resulting waveform was free of many of the distortions caused by the original parallel plate geometry. Moreover, the system shown in Figure 4 could be easily set-up or disassembled, and stored in a portable, reusable crate.
APELC 3rd Generation RS105 Test System
The systems we now offer utilized the work of Giri and Baum [2], who did considerable work in modeling and simulating TEM structures to come up with the most ideal design for radiated HEMP test systems. These are accepted as the standard for all such simulators.
The benefit of this geometry is that the single corner is moved well away from the test volume. There are no diffracted waves caused ahead of the test volume, and the reflected waves come very late time in the pulse, and are reduced by the 1/r field propagation.
The structure size is also defined in the MIL-STD, and is designed to minimize interactions between the structure and the test article. Generally, the test volume, or Enclosure Under Test (EUT) volume can be described as LEUT x WEUT x HEUT, and the surrounding structure must have the minimum dimensions of:
Lstructure = 2 x LEUT
Wstructure = 2 x WEUT,
Hstructure = 3 x HEUT
In reality, the structure must be substantially longer, in order to maintain the pulse into the EUT volume and create the desired impedance. Antenna angles between 14 – 17 degrees are common. Table 1 provides the antenna/structure design parameters, with the antenna dimensions derived directly from the standard. The final structure dimensions include the mechanical supports and guy wires.
The antenna is simply a wire structure forming a linear half-TEM horn antenna, often called a conical horn antenna. For this design, the top plate is created by twenty 1/8 inch stainless steel cables; and the ground plane is created by a wire mesh, cut slightly larger than the top plate to minimize fringing effects.
Figure 6. Conical antenna and wave propagation
The design of this system was robust, allowing for a wind loading of 80 MPH, and a maximum wire sag of 6 inches with ice loading. Furthermore, all elements of the support structure were designed to be non-conductive. The vertical supports were made of 12 inch diameter fiberglass poles. The horizontal beams are glue-laminated beams. Phillystrand guy wires maintain position and lateral stability.
Prior to construction, the structure was modeled and simulated using Remcom (FDTD). The primary focus was to ensure that the rise edge of the pulse propagates the structure. It was also critical to ensure that the peak electric field strength does not fall more than 6 dB throughout the EUT volume. Finally, the model was used to finalize the impedance of the structure, and thereby the voltage required from the pulsed power to create the peak electric field.
Figure 7 provides sample waveforms from the simulation. The graphic shows two pulses, one entering the EUT volume, and the second leaving the EUT volume. In this case, the rise time of pulse is maintained, and the amplitude drops approximately 10%- well within the required 6 dB required by the MIL-STD.
Figure 7. Sample results from the Remcom model, describing the pulse propagation through the EUT volume.
Table 1. Structure parameters
The pulsed power requirements were driven directly by the Remcom model. With a 6 m structure height at the EUT, a minimum 300 kV voltage pulse is required to meet the standard (50 kV/m). With a propagation loss of approximately 20% to the EUT volume, a voltage of at least 360 kV is needed.
The pulsed power system is described in Table 2 and shown in Figure 8. A coaxial Marx generator was built to deliver a maximum erected voltage of 455 kV (13 stages, with a maximum 35 kV charge voltage). Since the generator is coaxial, and the antenna is planar, the peaking circuit was designed as planar to match the antenna. This geometry negates the need for a zipper section, and substantially increases the voltage efficiency of the overall system.
Figure 8. The pulsed power source, including a compact Marx generator and a planar peaking circuit.
Table 2. Pulsed power parameters
The system was tested for its performance in delivering the required pulse throughout the EUT volume. Figure 9 provides a sample waveform of the pulse entering the structure, using a ground plane D-dot probe (AD-110) near the launch point. The perturbations in the waveform are due to the vertical supports located at 70 ns intervals. The load is approximately 200 ns from the point of measurement. Figure 8 provides a waveform, measured at 24 m from the source, and inside the EUT volume. This measurement is made with a free field probe (AD-55). At 75 ns, there is an abrupt change in the waveform due to a mismatch of the load. However, with this occurring so late in time, the pulse characteristics fall well within the temporal requirements of the standard.
Figure 9. A sample waveform of the pulse launched onto the antenna.
Figure 10. A sample radiated waveform compared to the ideal waveform
Figure 11. A comparison of the spectral content between a measured pulse and the theoretical waveform.
Ultimately, the waveforms and respective frequency responses must compare with those described in the MIL-STD. Figure 10 also provides a sample measured waveform compared against the theoretical waveform. The early and mid-pulse follows the theoretical shape very well. However, there appears to be some additional, unwanted late-time, low frequency energy in the tail of the pulse due to the mismatch at the load. As shown in the spectral view of Figure 11, the high frequency tracks the desired spectrum closely; however, moving to the left of the curve, there is some deviation from the theoretical trace.
Conclusion
Through a process of both research and trial and error, APELC developed a radiated HEMP test system for both MIL-STD-461 (RS105) and MIL-STD-2169 testing that is robust, portable and reproduces the standard waveform with extremely high-fidelity. This design is scalable from small box-level tests all the way up to full system-level (e.g. vehicles) tests.
REFERENCES
[1] MIL-STD-461G, “Department of Defense Interface Standard: requirements for the Control of Electromagnetic Interference Characteristics of Subsystems and Equipment.” US Department of Defense: March (2015) [2] D. V. Giri, “Design Guidelines for Flat-Plate Conical Guided-Wave EMP Simulators With Distributed Terminators.” (1996)