The Project 8 related HFSS models range from microstrip patch antenna arrays to open-ended waveguides to slots in a waveguide and more. Although the patch models have not been used lately they still serve as a simple and low profile solution for detection of electromagnetic radiation. Also, the current slot antenna technology lacks a conviction as being the FSCD (Free Space CRES Detector) solution and thus we are still open to horns, patches, cavities, slots, etc (as of June 2020). Therefore, we have set up a few HFSS .aedt “project” files and/or tutorials to assist the new Project 8 Member in using HFSS for Phase III/FSCD design. The more important examples have been selected and are shown/described below. You can also find related models in this folder. Please feel free to email Tim Wendler for any questions: timoth500@yahoo.com. It’s always smart to begin with a general non-p8 HFSS tutorial. But if you are confident with Microwave FEM software fundamentals and basic CAD principles feel free to move to the following P8 specific examples:

1. Locust Source in HFSS: Importing and Examining a CRES Source near-field

This tutorial takes you through importing near-field data made by Locust. This is theoretically the field emitted by a stationary event (pitch equals 90). A part 2 is available in the same folder. Various CRES spheres with different size and time-avg can be found here. For example, there exists an instantaneous CRES source for HFSS (just a flashing 6dB relativistic lobe pointed in one direction), which has not proven to be helpful yet so we have stuck to the time-avg sources.

2. Generating a Transfer Function: For Locust’s Finite-Impulse Response Filter generator we made a generalized Transfer Function text file. These files are generated by HFSS, here is an example: TF_for FIR.aedt. This example shows a horn antenna in both Transmit and Receive mode. For the Rx mode, it shows the complex voltage function at the antenna output over a 50 Ohm termination. The TF is 401 frequencies over a 4GHz bandwidth, as used in Locust for the Filter generator. To calculate specific field values in HFSS one typically creates a function using the Fields Calculator Cookbook. Here are a few resources for this more exotic antenna topic, namely antennas in the time-domain.

3. FSCD Related Examples File: FSCD_tutorial.aedt

Contains a range of relevant structures to the current FSCD system such as wr34 waveguide, horn antenna, slotted waveguide, and an actual CRES ball imported from Locust into HFSS as a near-field source. Since this is a more dense model I have decided to briefly explain each model within the single project. Note: requires sphere_N2048_1cmrad.and, .nfd files.

1. WR34 Lumped RLC impedance derivation, includes a key parameter called the propagation constant (attenuation constant + i*phase constant). The WR34 waveguide generally has a wave impedance of around 500 Ohms, but when using the Lumped RLC boundary with Z_pi definition one should use closer to around 270 Ohms for 25.9GHz.

2. Half-wave dipole gain derivation: Note that this is a resonant structure with the S11 being better around the frequency with wavelength of roughly twice the length of the antenna. Also, note the pattern is a dipole donut with a theoretical gain around 2dBi. In reality the model should produce: between 1 and 2 linear

3. WR34 to E and H flare horn antenna (should get a gain between 10 and 20dBi)

4. Single traveling wave waveguide slot unit cell (compare equivalent circuit admittance to Zhang and Elliot et al)

5. Slot array optimized for 5 slots (10 slots has a gain of 17dB, therefore half the elements should be 3dB less). Note that the slot offset from the center line of the broadwall is 1.5mm, while the 10slot array is more like only 1mm offset.

6. Locust CRES source incident onto a 10 slot array (Requires specific .NFD and .AND files)

   1. The general format of the CRES file naming involves the time-avg in terms of Fourier bins and spherical radius for Huygen’s sphere. So the ones above are spheres with a time-avg determined by 2048 points in the transform and a radius of 1cm for the Huygen’s sphere. Various CRES spheres can be found here.. The NFD fields should be peak phasor quantities representing the time-harmonic electric and magnetic fields.

4. Transmission-line Basics: Simple transmission line concepts here in a short microstrip example. In this example you can find some basic t-line analysis of microstrips with different substrate materials. Plot the real and imaginary parts of the key parameter called the propagation constant (attenuation constant + i*phase constant). This parameter will show any cutoff frequencies for various modes with the imaginary part, and attenuation vs. frequency with the real part.

5. Patch Antenna: patch_antenna.aedt

A K-band microstrip patch antenna fed on the radiative edge (with inset for matching). This example shows how the feed length is an important parameter to consider when matching the antenna to 50 Ohms. It also shows the effects of a wider patch and the unwanted moding that results.

6. 10 Slots in a WR34 Waveguide Array Center-Feed: Coax Center Fed, this model made by Maurio Grando is a waveguide antenna for 26GHz. Note the slot offset is around 1mm, which is specific to the number of slots 10, while a system of less slots like 5 or less is closer to 1.5mm all the way to the edge of the broadwall for a single-slot (theoretically).

7. 10 Slots in a WR34 with a Waveport End-Feed: An end fed slotted wr34 waveguide initially designed for Phase III. Note the near-field interference issues for sources closer than 10cm. This can be seen in pure Tx mode by plotting the ComplexEmag over a sheet in the near-field and seeing that more and more nodes/antinodes appear as you move close to the array.

8. Finite Element Boundary Integral Method: FEBI example file

This example is crucial for the electromagnetically larger models that slow the machine down. HFSS has developed a way to separate solution spaces when one does not need to analyze space between largely separated structures. This example shows how to do that effectively thanks to M. Jones for putting this together.

9. CRES “like” source: Here is an example of a synthetic CRES antenna. This model was given to Mark Jones during the early development of the Synthetica CRES antenna. Note the 4 signals are all pi/2 out of phase as you go around the z-axis, we are referring to this as spatial quadrature. This design is based on a crossed Hertzian dipole.

10. EM model for Phase II insert: Full short-LNA model and proposed mode-absorber HFSS projects here.

1. Eigenmodes: A few of these models show the eigenmodes from the insert as a whole. They generally agreed with Elise and Ali’s statistical analysis. However, the system was so complex we could not gather any confident tractions. One difficult aspect was the anisotropic effective dielectric constant of the Birefringence crystal that converts cross polarized modes to the TE01 mode of the exiting waveguide connected to the LNA

2. Wave absorber: In Phase II we ran into an issue where the electron would couple to a Transverse Magnetic mode of the circular cavity when offset from the center. It forces the event to show bursts of power in the form of curves in the spectrogram lines. To mitigate this we developed a TM mode absorber that had longitudinal strips of metal.

11. Hamish’s Diffraction Grating and Alec’s reflector system: For more exotic detector ideas see Hamish and Alec’s grating ideas here.

12. Full Rx chain in AnsysEDT: For an example on how to include the Circuit simulator in Ansys and include the full Rx chain see this to start. Note that you can include the whole down-conversion as well as sample/quantization/digitization steps as they are all indeed separate processes on a fundamental level.

13. EE Friis Calculations: A key folder with Python Scripts written by M. Jones and M. Grando is here. This is key to use because far-field couplings should stay grounded with sanity checks using a classic link budget scheme, or Friis. A script to calculate beam-forming steering vectors and produce a text file for HFSS is here.

14. WSU 13GHz Array Model (10 patch array with amplitude tapering). WSU EE department worked on a scaled version of a Phase III array. The idea was that the properties of an antenna at 13GHz can simply be scaled up in frequency and space so one does not need to work at high frequencies when considering a manufactured realization of the HFSS model. A few of these actual antennas are in the Penn State antenna lab.

15. Radome and Floquet Theory: Radome and Unitcell for Slots are two example files of the use of periodic boundary conditions in HFSS. Ultimately the antennas must be isolated from the 85K tritium gas as well as shielded from the gas’s IR emission. The antennas are meant to stay in the single digits of Kelvin. These models have various metasurface structures that act as low pass (i.e. <50GHz to reflect IR) and bandpass filters (half-lambda thick SiO2 radome) to make the necessary isolating structure transparent to 25.9GHz.

16. Absorber around Slots: slotAbsorber_examples When considering the slotted waveguides in a ring in the full detector their exposed metal surface became an issue, we began to see “ringing” that would reduce the ability to localize any single event within the volume. A series of absorber models were made with PNNL’s guidance that consider surrounding the slots with absorber. The drawback is that we cannot get too close to the slots because the current surrounding the slots are a key to the radiation mechanism. We were however able to find some ideal distances that did not affect the gain of the arrays themselves. The absorber properties for this were difficult since the absorber had to be so thin. You can play around with the effective permeability and magnetic properties (very mild to beefy) to see how it affects the array as a whole, but thickening the absorber quickly ate into its gain.

17. Examples of Full Detector Simulations:

1. Full 10cm radius ring of 60x6slot linear arrays. This is an electromagnetically large model so many desktop computers do not have a chance at solving this. The FEBI method is seen to exploited but even then the memory requirements are enormous.

2. Full 5cm radius ring of 30x short and long linear patch arrays: Example 1 Example 2 Example 3

18. Superlist of Misc. Models: An archive of most of my p8 models can be downloaded from the folder here. All tutorial material in a zip can be found here.

As of June 23, 2020, the antenna design with HFSS has been transferred to Yale. Arina B. and James N. have taken the HFSS side of things while the new antenna WG chair is Pranava S. In this transition we have completed a set of real-time tutorials (12 hours total zoom time over 1 month) which have covered the following topics:

Completed Tutorials with Yale (see Arina B. or James N. for details)

1. Halfwave Dipole Locust TF

2. wr34 section with slot traveling wave getting Zhang/Elliot numbers

3. slot array standing wave optimization of offset

4. Importing Locust NFD source and normalizing to 1.08fW and rE pattern analysis (requires .nfd and .and files from Locust)

5. Locust NFD onto Rx structure (requires .nfd and .and files from Locust)

6. Make an official Rx TF for a horn (401 Complex voltage values from 24GHz to 28GHz with an approximate output impedance)

7. Test for energy conservation and correct phase response with well-defined output Z_load (use built-in horn with plane wave pure Rx mode)

8. Comparing horn field and power pattern to cos^p(theta)

9. Calculating Wave-Impedance from near-field to far-field

10. Changing orthographic to perspective view for papers/posters

11. When in doubt use power with phase (always calibrate/normalize voltage and Z with power in the mode, sometime requires a search for the right eLine and/or using Effective Height)

12. When to need to use "terminal" solution type (single-ended vs. balanced, common-mode vs. differential mode)

13. Importing spirals/helix antennas and other misc preset antenna structures as 3D Components (also making 3D components)

14. Brief intro into using the Transient solver (step and impulse)

15. Optimizing coax port for 50 Ohm input impedance - zero reactivity aka pure real resistance (dismissal of ELI the ICE man is a must for final engineering design requirement)

16. Two levels of 4-slot array TF and calculation of Wave-Impedance of orthogonal directions

17. Showing Babinet's principle of a slot in a metal sheet and a planar dipole.

18. Import coils from Maxwell and assign IE regions to the coils with FE-BI for the antennas and source

19. Reproduce near-field CRES or dipole sweep map as seen in TF channel

20. Plot Jsurf and Q around the slots, then plot poynting Vector over a invisible surface 1mm above the slots (edited)