Presentation on theme: "Boundary Setup Exercise: Radiation"— Presentation transcript:
1Boundary Setup Exercise: Radiation 1. The Boundary radio button should remain selected.2. From the list of available boundaries, select Radiation.3. Leave the Graphical Pick option set to Face.4. Click in the graphical window to touch the air volume surrounding the structure on the three faces indicated. You may wish to rotate to facilitate your selection.NOTE: Had this model been constructed with the air solid sitting on top of the substrate solid, instead of containing the substrate solid, we would have to pick specific faces on three sides of the substrate object as well.5. In the Name field, type in “absorbing”, and click the Assign button.6. The boundary should appear in the boundary list at left.NOTE: We have assigned a Radiation boundary over where the microstrip port will need to be! This will be superceded in a step in part 2 of this exercise, following the Source discussion.3.4. (Back, right side, top)22.214.171.124.
2Boundary Setup Exercise: Ground Plane 1. The Boundary radio button should remain selected.2. From the list of available boundaries, select Perfect E.3. Leave the Graphical Pick option set to Face.4. Either rotate the model view to bring the lower face to the front, and click on it, or click as though touching the lower face of the air volume and use the “N” key to shift focus deeper to the lower surface of the air volume and substrate.5. In the Name field, type in “ground_plane”, and click the Assign button.6. The boundary should appear in the boundary list at left.NOTE: Since this is being assigned a Perfect E boundary, we could have allowed the automatic “outer” boundary to take care of this face if we wished.126.96.36.199.2.5.
3Boundary Setup Exercise: Symmetry Plane 1. The Boundary radio button should remain selected.2. From the list of available boundaries, select Symmetry.3. Leave the Graphical Pick option set to Face.4. Click on the face of the model which bisects the microstrip trace and coax. Once a face is selected, the options for the Symmetry boundary appear below the graphical view. Click again in the model to select the cut faces of the ‘thru_hole_in_wall’ and “coax_outer” cylinders as well. (You may wish to zoom in to assure you have the correct faces selected.)NOTE: Again, if we had defined our air volume to sit atop rather than to contain the substrate, we would need to select the substrate face too.5. In the parameter space for the boundary, click the radio button for Perfect H type symmetry (E-fields tangential to surface).6. In the Name field, type in “mag_symmetry”, and click the Assign button.7. The boundary should appear in the list at left.THIS CONCLUDES PART 1 OF THE BOUNDARY SETUP EXERCISE. DO NOT EXIT THE BOUNDARY/ SOURCE MANAGER.188.8.131.52.184.108.40.206.5.
4HFSS Ports: A Detailed Look The Port Solution provides the excitation for the 3D FEM Analysis. Therefore, knowing how to properly define and create a port is paramount to obtaining an accurate analysis.Incorrect Port Assignments can cause errors due to......Excitation of the wrong mode structure...Bisection by conductive boundary...Unconsidered additional propagating modes...Improper Port Impedance...Improper Propagation Constants...Differing phase references at multiple ports...Insufficient spacing for attenuation of modes in cutoff...Inability to converge scattering behavior because too many modes are requestedSince Port Assignment is so important, the following slides will go into further detail regarding their creation.We will now spend some time going into detail on each parameter in the port setup, as proper port assignment is vital to getting good HFSS results. The exercise will be continued following this examination.This particular slide can be passed over quickly, as it merely lists different errors that can be caused by inappropriate port assignment.
5HFSS Ports: Setup Interface Name FieldPorts are always named “portN”. Box also includes Assign, Clear, and Options buttons.Lumped Gap Source Port OptionActivating enables Port Impedance entry fields.The parameter fields for a port are magnified in the above view, with different parameters highlighted.The instructor may wish to comment that the Impedance Multiplier field is not really ‘tied’ to the port specifically, but an overall project settings which is made visible when a port is selected.These parameters and their use are described in detail in the slides following.Mode Entry FieldSet port mode solution requirements. Set polarization. Shows impedance and calibration definitions applied, if any.Impedance and Calibration Line Fields‘Edit Line’ dropdown allows setting, clearing, and relating Imped. and Calib. lines.Impedance Multiplier FieldUse if symmetry planes intersect ports.
6HFSS Port Selection: Standard or Gap Source? When would you choose to use a Gap Source Port over a Standard Port?When the model has tightly-spaced linesWhen ‘backing’ the port would be too disruptive of internal fieldsWhen a port reference location is difficult to determine using a Standard portWhen you’d like to use a voltage gap, but want S-parameter outputCriteria for selecting a standard vs. gap source type port.Gap Source Ports (blue)
7A Microstrip Port Assignment (includes air above substrate) HFSS Ports: SizingA port is an aperture through which a guided-wave mode of some kind propagatesFor transmission line structures entirely enclosed in metal, port size is merely the waveguide interior carrying the guided fieldsRectangular, Circular, Elliptical, Ridged, Double-Ridged WaveguideCoaxial cable, coaxial waveguide, square-ax, Enclosed microstrip or suspended striplineFor unbalanced or non-enclosed lines, however, field propagation in the air around the structure must also be includedParallel Wires or StripsStripline, Microstrip, Suspended StriplineSlotline, Coplanar Waveguide, etc.A Coaxial Port AssignmentFirst look at port sizing possibly including area ‘outside’ of the transmission line structure itself. This is a frequent point of confusion for those used to using circuit tools (or even our own Ensemble and Spicelink tools, where the ends of conductors are selected as ‘ports’ and ‘terminals’ with no other area designation).Additional sizing guidelines follow.A Microstrip Port Assignment (includes air above substrate)
8HFSS Ports: Sizing, cont. The port solver only understands conductive boundaries on its bordersElectric conductors may be finite or perfect (including Perfect E symmetry)Perfect H symmetry also understoodRadiation boundaries around the periphery of the port do not alter the port edge termination!!Result: Moving the port edges too close to the circuitry for open waveguide structures (microstrip, stripline, CPW, etc.) will allow coupling from the trace circuitry to the port walls!This causes an incorrect modal solution, which will suffer an immediate discontinuity as the energy is injected past the port into the model volumePort too narrow (fields couple to side walls)Examples of too-narrow and too-short ports to illustrate impact of sizing on solved fields.Specific sizing guidelines for different line types follow.Port too Short(fields couple to top wall)
9HFSS Ports: Sizing Handbook I Microstrip Port Sizing GuidelinesAssume width of microstrip trace is wAssume height of substrate dielectric is hPort Height GuidelinesBetween 6h and 10hTend towards upper limit as dielectric constant drops and more fields exist in air rather than substrateBottom edge of port coplanar with the upper face of ground plane(If real structure is enclosed lower than this guideline, model the real structure!)Port Width Guidelines10w, for microstrip profiles with w h5w, or on the order of 3h to 4h, for microstrip profiles with w < hwh6h to 10h10w, w hor5w (3h to 4h), w < hMicrostrip sizing guidelines, standard ports.Note that the rule changes for ports in which the trace is narrower than the dielectric height. For these high impedance lines, the port outline should be proportional to the dielectric rather than to the line width.Instructor should caution that these are good ‘guidelines’ or ‘rules of thumb’ but not guaranteed in all cases. If the user is designing a rather non-standard transmission line structure, he or she should do a ports-only solution first with a few different port dimensions to assure that the solution is not highly dependant upon the port outline selected.Note: Port sizing guidelines are not inviolable rules true in all cases. For example, if meeting the height and width requirements outlined result in a rectangular aperture bigger than /2 on one dimension, the substrate and trace may be ignored in favor of a waveguide mode. When in doubt, build a simple ports-only model and test.
10HFSS Ports: Sizing Handbook II Stripline Port Sizing GuidelinesAssume width of stripline trace is wAssume height of substrate dielectric is hPort Height GuidelinesExtend from upper to lower groundplane, hPort Width Guidelines8w, for microstrip profiles with w h5w, or on the order of 3h to 4h, for microstrip profiles with w < hBoundary Note: Can also make side walls of port Perfect H boundarieswh8w, w hor5w (3h to 4h), w < hPort sizing guidelines for stripline. Stripline ports are constrained by ground planes top and bottom, and are less sensitive to sizing than microstrip.Instructor should caution that these are good ‘guidelines’ or ‘rules of thumb’ but not guaranteed in all cases. If the user is designing a rather non-standard transmission line structure, he or she should do a ports-only solution first with a few different port dimensions to assure that the solution is not highly dependant upon the port outline selected.
11HFSS Ports: Sizing Handbook III Slotline Port GuidelinesAssume slot width is gAssume dielectric height is hPort Height:Should be at least 4h, or 4g (larger)Remember to include air below the substrate as well as above!If ground plane is present, port should terminate at ground planePort Width:Should contain at least 3g to either side of slot, or 7g total minimumPort boundary must intersect both side ground planes, or they will ‘float’ and become signal conductors relative to outline ‘ground’Approx 7g minimumLarger of 4h or 4gghPort sizing guidelines for slotline. Note inclusion of air above and below substrate, unless slotline has a ground palane beneath substrate!!!Instructor should caution that these are good ‘guidelines’ or ‘rules of thumb’ but not guaranteed in all cases. If the user is designing a rather non-standard transmission line structure, he or she should do a ports-only solution first with a few different port dimensions to assure that the solution is not highly dependant upon the port outline selected.
12HFSS Ports: Sizing Handbook IV CPW Port GuidelinesAssume slot width is gAssume dielectric height is hAssume center strip width is sPort Height:Should be at least 4h, or 4g (larger)Remember to include air below the substrate as well as above!If ground plane is present, port should terminate at ground planePort Width:Should contain 3-5g or 3-5s of the side grounds, whichever is largerTotal about 10g or 10sPort outline must intersect side grounds, or they will ‘float’ and become additional signal conductors along with the center strip.Larger of approx. 10g or 10sLarger of 4h or 4gshgPort guidelines for CPW. CPW, because of its vast range of possible ‘aspect ratios’, is among the hardest of lines to obtain a good port solution for. More experimentation may be necessary, especially in cases of thick, lossy dielectrics (e.g. to feed a spiral inductor on silicon).Another rule of thumb can be to add (2g+s) and call it a total factor, W, then make the port width no less than 3W. This results in having sufficient side grounds, but is based on the overall CPW size rather than its aspect ratio.Instructor should caution that these are good ‘guidelines’ or ‘rules of thumb’ but not guaranteed in all cases. If the user is designing a rather non-standard transmission line structure, he or she should do a ports-only solution first with a few different port dimensions to assure that the solution is not highly dependant upon the port outline selected.
13HFSS Ports: Sizing Handbook V; Gap Source Ports Perfect EGap Source ports behave differently from Standard PortsAny port edge not in contact with metal structure or another port assumed to be a Perfect H conductorGap Source Port Sizing (microstrip example):“Strip-like”: [RECOMMENDED] No larger than necessary to connect the trace width to the ground“Wave-like”: No larger than 4 times the strip width and 3 times the substrate heightThe Perfect H walls allow size to be smaller than a standard port would beHowever, in most cases the strip-like application should be as or more accurateFurther details regarding Gap Source Port sizing available as a separate presentationPerfect HPerfect HPerfect EPerfect HStandard port sizing having been reviewed, gap source port sizing options should also be discussed.In general Gap Source ports should be used in a “Strip-like” configuration.Perfect HPerfect HPerfect E
14HFSS Ports: Spacing from Discontinuities Structure interior to the modeled volume may create and reflect non-propagating modesThese modes attenuate rapidly as they travel along the transmission lineIf the port is spaced too close to a discontinuity causing this effect, the improper solution will be obtainedA port is a ‘matched load’ as seen from the model, but only for the modes it has been designed to handleTherefore, unsolved modes incident upon it are reflected back into the model, altering the field solutionRemedy: Space your port far enough from discontinuities to prevent non-propagating mode incidenceSpacing should be on order of port size, not wavelength dependentPortExtensionThe need for port extensions. Instructor may wish to refer back to the comment made at the end of the Modeler presentation, where (as a geometry suggestion) it was indicated that “port extensions need not be TOO long”. Too many people assume that the port extension should be based on wavelength, but since evanescent modes decay exponentially, this is not the case.
15HFSS Ports: Single-Direction Propagation Port on Exterior Face of ModelStandard ports must be defined so that only one face can radiate energy into the modelGap Source Ports have no such restrictionPosition Standard Ports on the exterior of the geometry (one face on background) or provide a port cap.Cap should be the same dimensions as the port aperture, be a 3D solid object, and be defined as a perfect conductor in the Material Setup moduleNeed for a backer or a port on background. Instructor may wish to refer again to the Modeler exercise which included creation of a port backer. Instructor should also clarify that this is necessary for Standard ports only, not gap source ports.Port Inside Modeled Air Volume; Back side covered with Solid Cap
16HFSS Ports: Mode Count Ports should solve for all propagating modes Ignoring a mode which does propagate will result in incorrect S-parameters, by neglecting mode-to-mode conversion which could occur at discontinuitiesHowever, requesting too many modes in the full solution also negatively impacts analysisModes in cutoff are more difficult to calculate; S-parameters for interactions between propagating and non-propagating modes may not converge wellWhat if I don’t know how many modes exist?Build a simple model of a transmission line only, or run your model in “Ports Only” mode, and check!You can alter the mode count before running the full solution.Degenerate mode ordering is controlled with calibration lines (see next slide)Circular waveguide, showing two orthogonal TE11 modes and TM01 mode (radial with Z-component). Neglecting the TM01 mode from your solution would cause incorrect results.Discussion of port modes. Too few may miss real results; too many impairs convergence.This subject segues nicely into the requirements for calibration lines.
17HFSS Ports: Degenerate Modes Degenerate modes have identical impedance, propagation constantsPort solver will arbitrarily pick one of them to be ‘mode(n)’ and the other to be ‘mode(n+1)’Thus, mode-to-mode S-parameters may be referenced incorrectlyTo enforce numbering, use a calibration line and polarize the first mode to the lineOR, introduce a dielectric change to slightly perturb the mode solution and separate the degenerate modesExample: A dielectric bar only slightly higher in permittivity than the surrounding medium will concentrate the E-fields between parallel wires, forcing the differential mode to be dominantIf dielectric change is very small (approx or less), impedance impact of perturbation is negligibleIn circular or square waveguide, use the calibration line to force (polarize) the mode numbering of the two degenerate TE11 modes. This is also useful because without a polarization orientation, the two modes may be rotated to an arbitrary angle inside circular WG.Use of calibration line for force ordering of degenerate modes. Note that the ‘trick’ of using a dielectric bar to separate the modes slightly for some structures is also discussed.Calibration lines also provide phase referencing, as discussed in following slides.For parallel lines, a virtual object between them aids mode ordering. Note virtual object need not extend entire length of line to help at port.
18HFSS Ports: Phase Calibration A second purpose of the calibration line is to control the port phase referencesThe 2D port eigensolver finds propagating modes on each port independentlyThe zero degree phase reference is chosen at a point of maximum E-field intensity on the port face.This occurs twice, with 180 degrees separation, for each 360 degree cycleTherefore the possibility exists for the software to select inconsistent phase references from port to port, resulting in S-parameter errorsAll port-to-port S-parameter phases, e.g. S21, will be off by 180 degreesSolution: The calibration line defines the preferred direction for the zero degree reference on each port.Which of the above field orientations is the zero degree phase reference? Calibration Line defines...Second use of calibration lines as phase reference assignment. Since most users will be constructing models in which only single-mode propagation is desired, this is actually a more common usage for calibration lines than degenerate mode ordering.
19HFSS Ports: Impedance Definitions HFSS provides port characteristic impedances calculated using the power-current definition (Zpi)Incident power is known excitation quantityPort solver integrates H-field around port boundary to calculate current flowFor many transmission line types, the power-voltage or voltage-current definition is preferredSlot line, CPW: Zpv preferredTEM lines: Zvi preferredHFSS can provide these characteristic impedance values, as long as an impedance line is identifiedThe impedance line defines the line along which the E-field is integrated to obtain a voltageOften it can be identical to the calibration lineFor a Coax, the impedance line extends radially from the center to outer conductor (or vice versa). Integrating the E-field along the radius of the coaxial dielectric provides the voltage difference.In many instances, the impedance and calibration lines are the same!Impedance Line usage is introduced, with a brief discussion of the assumptions HFSS makes in solving port impedances.In many cases an impedance line is not necessary in the model. However, since it can often match a calibration line which may be necessary (to enforce the correct phase reference from port to port), it takes little effort to apply it as well by simply copying the Calibration line.
20HFSS Ports: Impedance Multiplier When symmetry is used in a model, the automatic Zpi and impedance line-dependant Zpv and Zvi calculations will be incorrect, since the entire port aperture is not represented.Split the model with a Perfect E symmetry case, and the impedance is halved.Split the model with a Perfect H symmetry case, and the impedance is doubled.The port impedance multiplier is just a renormalizing factor, used to obtain the correct impedance results regardless of the symmetry case used.The impedance multiplier is applied to all ports, and is set during the assignment of any port in the model.Whole Rectangular WG(No Symmetry)Impedance Mult = 1.0Half Rectangular WG(Perfect E Symmetry)Impedance Mult = 2.0Again, the Impedance Multiplier is not really a specific “port’s” setting, but an overall project setting, but its use is associated with ports and therefore discussed herein.Half Rectangular WG(Perfect H Symmetry)Impedance Mult = 0.5...and for Quarter Rectangular WG?(Both Perfect E and H Symmetry)Impedance Mult. = 1.0
21Source Setup Exercise: Part 2 We will now complete the Setup Boundaries/ Sources exercise already begun, by adding the two necessary Ports to the problemPorts will use both calibration and impedance lines, but will require only one mode on each terminalConclude the exercise already begun by assigning the two ports to the model. Note that in the pages following, ports were intentionally assigned so that the coax port calibration/impedance line points UP and the microstrip one points DOWN, yet both go from “signal trace to ground”. This is to get users thinking about the field propagation in each port, not simply in terms of Cartesian directions.
22Source Setup Exercise: Coaxial Port 1. Select the Source radio button.2. The source list should set by default to Port.3. Zoom in on your model, or otherwise orient it so you have clear visual access to the extended face of the coaxial line. Click on the face to select it. The port parameter entry fields will now appear.4. Leave the port name as “port1”, and the number of modes as “1”.5. Check the box for Use Impedance Line. This enables the Edit Line dropdown menu beside it. In the Edit Line dropdown, pick Set...6. The side window will prompt you to Set Impedance Start. Define a starting point for the impedance line by clicking in the graphical window to snap to the vertex on the inner conductor, at the topmost point where it intersects the symmetry plane. (A first click may be necessary to activate the window before selecting vertices.)7. Click the Enter button in the side window to confirm your point selection. The window now shifts to request the vector information. (proceed to next page)220.127.116.11.2.4.5.
23Source Setup Exercise: Coaxial Port, cont. The interface will now show a vector from the starting point you defined to the origin. (This is merely a default ‘guess’ at the intended endpoint.) The side window shows vector entry fields.8. In the graphical window, snap to the point radial from the starting point (on the outer conductor radius, at the topmost intersection with the symmetry plane). The vector fields should update to reflect a Z-directed vector.9. Press the Enter button on the side window to confirm the vector end point. The side window interface closes, and the completed impedance line is displayed as a red vector with the letter “I”.10. Check the box for Use Calibration Line. In the enabled Edit Line dropdown to the right, pick Copy Impedance. The vector will now update to include a “C” indication.11. Press the Assign button to complete the port creation. The boundary list will now update to show “port1”18.104.22.168.
24Source Setup Exercise: Microstrip Port Rotate and resize the graphical window so that you have visual access to the microstrip termination end of the model.1. Source radio button and Port type are already selected.2. Click on the 2D rectangle provided for the microstrip port face. If the entire face of the air volume is selected, use the Next Behind menu pick or “N” hotkey to shift the selection.3. Leave the port name as “port2”, and the number of modes as “1”.4. Use Impedance Line should remain checked from the prior port assignment. In the Edit Line dropdown, pick Set...5. The side window will prompt you to Set Impedance Start. Define a starting point for the impedance line by clicking in the graphical window to snap to the vertex on the trace, at the point where it intersects the symmetry plane.6. Click the Enter button in the side window to confirm your point selection. The window now shifts to request the vector information. (proceed to next page)22.214.171.124.3.4.
25Source Setup Exercise: Microstrip Port, cont. The interface will now show a vector from the starting point you defined to the prior port’s ending point. The side window shows vector entry fields.7. In the graphical window, snap to the point where the ground plane intersects the symmetry face.8. Press the Enter button on the side window to confirm the end point. The side window interface closes, and the completed line is displayed.9. Use Calibration Line should already be checked. In the enabled Edit Line dropdown to the right, pick Copy Impedance.10. Before assigning the port, we need to set the Impedance Multiplier for the model. Enter a value of 0.5.11. Press the Assign button to complete the port assignment. You will receive an overlap warning, because the port overlays the earlier “radiation” boundary. After the overlap warning message is dismissed “port2” will show in the boundary list.12. We are now done with boundary assignment. To verify our assignment, pick Boundary Display from the Model menu.126.96.36.199.9.10.
26Source Setup Exercise: Verifying Boundaries HFSS will now perform only the initial meshing necessary to subdivide the problem into tetrahedra, so that actual boundary application to the finite element mesh can be viewed.13. The list of assigned boundaries is on the left. Note that it contains both boundaries we created, plus the boundaries “i_pinn” and “outer”. The “i_pinn” boundaries were assigned as a result of assigning a finite conductivity metal -- copper -- to the pin objects. The “outer” boundary is applied to any surface of the model we did not otherwise define. Highlight “outer” in the boundary listing.14. Press the Toggle Display button. The mesh on the selected boundary is displayed, indicating the surfaces on which this boundary is applied. Note that it provides the Perfect E definition on the outer conductor of the coax, on the outer conductor of the thru hole, and on the front face of the model which represents the metal module wall.15. If you wish you may continue displaying additional boundaries. When you are through, press the Close button to return to the Setup Boundaries window. There, pick Exit from the File menu and save when prompted. (The overlap warning will repeat on exit.)13.14.15.