Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
it depends on your application, for example if you are exciting a waveguide you use a waveport, if you are exciting a terminal (e.g. a dipole) you use lumped port.
I would use only lumped ports when using wave ports are not possible. Wave ports use a 2D solver on the port cross-section to obtain the modes that will propagate down a transmission line of infinite extent. They can be used for many geometries such as waveguides, stripline, CPW, microstrip, coax, etc.
Lumped ports can be used to feed balanced antennas such as dipoles or spirals. They can also be used when geometries are constrained so that the port must be internal to the problem.
How about in case of the signal line locates inside substrate. In this case, we can't use waveport? Who can explain for us more about what is the general ideal for lumped port excitation?
If you are referring to a stripline, where the signal line is embedded inside two dielectrics, then yes a waveport can be used for this.
If you are referring to a transmission line that begins inside a substrate and inside a model, then you can use a lumped port for this. However, I have seen a waveport used for interior ports in some cases if the waveport is backed with a PEC to let the solver know which direction the energy should propagate.
The general idea of a lumped port is that you draw a rectangle from the signal line to ground and assign a voltage/integration/calibration line and input impedance. The S-parameters are calculated directly at the port location and cannot be de-embedded.
Ok but there's something more not so clear to me...
I've taken a model from the HFSS tutorial manual on a patch feeded by a coax probe. There was used a waveport and the solution was performed as Driven Modal.
Then for a question of completeness I've made a Driven terminal solution with the waveport and I've obtained the same results(very close).
Then I've substituted the waveport with the lumped port and I've tried to solve again in Driven Modal solution.
Before I've simulated the original structure just with the port substitution and then I've added an air box to enclose the coax feed so I could use the lumped port in the classical method:Inside the model and not at the boundary...
So for the lumped port in the first case I've obtained some 'little' trasnlation in resonance frequency but a 5 dB change in S11...not acceptable in terms of accuracy...
For the second case, with the airbox, I've found an absolutely inconsistent solution...
There's someone who can definitively tell us what's the correct way to use this terrible Lumped ports?
Thanks...
The concept of "lumped ports" is a basic one and can be implemented in various methods (MoM, FEM, TLM, etc.). For example, NEC uses a voltage lumped gap source. In HFSS, you draw a lumped gap source using a 2D rectangle and assigning it as a port. Set the impedance and calibration line. A small difference in resonant frequency is to be expected when comparing to a wave port such as coax since they are not presenting the same impedance characteristics to the model. They should be used when you cannot use a wave port.
It sounds like there may be a problem with the boundary definitions in the air box case. I would double-check all of the boundary condition assignments. The two solutions should be in fairly good agreement.
In general, one should use wave ports wherever possible. The reason for this is that for any type of technology (stripline, u-strip, coax, etc.) a lumped port is only an approximation of the actual. For instance, take a u-strip. Most (but not all) of the E-field lines extend from the center conductor directly down to the ground plane. However, there are also many field lines, "fringe" lines, that take a more curling or indirect path from the center conductor to the ground. Thus, the quasi (as opposed to full) TEM solution for u-strip. We can think of the lumped port as being the idealized (non-fringing) solution whereas the wave port also adds the fringe lines. Therefore, the waveport will always be the more accurate.
There are also other discontinuities inherrent with the use of lumped ports. In the coax example, you added an air box to keep the center conductor from shorting to the outter conductor. having this air box will add a discontinuity in the model that may not correspond to real life.
With all this said, there are some very real reasons for using lumped ports, but they should be used sparringly.
hi,
I am using hfss for past 5,6 months. Initially i used to get confused whether to use lumped or wave ports. After some time i found wave ports give good results. So use wave ports whenevrr possible. (But one needs to know how to use with uS ,CPW etc).But still iiam not very clear abt lumped ports. ppl (and manual) say internal ports etc im not very clear still. can any1 give a clear example where lumped port is absolutely necessary...
cu
cv
The strigent requirement for proper integration line as well as the assignment of the proper port impedance one has to perform in HFSS in order to use a lumped port suggests to me that, ALWAYS use a waveport whenever possible !
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
