.

Arnand

.

You need to state your question more clearly.

Are you asking how to design an antenna so

as to optimally match into a 50-ohm load?

There are some good reasons why this has

become the reference impedance in cabling

and instrumentation, but of course, there is

nothing  'natural' about this choice. On the

other hand, the minimum loss is a coaxial

cable occurs at about 70 to 75 ohms.

.

Barrie

Thank you sir for your reply.

I am asking about what inference should we conclude from  this term Reference Impedance.

This term comes into picture when we simulate the antenna design.

Also I am curious to know as how it varies ?

I have been facing a problem as of getting variable reference impedance for the same strip feed line.

Please answer.

I also don't understand your question. 

Do you mean that reference impedance changes when you simulate?

What's your reference impedance's meaning?

Generally, the impedance of an antenna's feedline is 50 ohm.

But input impedance of an antenna is different from the antenna structure.

So, we need matching networks always.

I think you should make your question clear.  

Reference impedance is a number that is used to define S parameters. If you have S parameters and you want information about voltages and currents (or vice versa), you need the reference impedance.

When you calibrate a vector network analyzer (VNA), you are defining the positions of measurement reference planes and the reference impedances of those planes. The value of a reference impedance thus defined comes from the physical object (a certain object in your 'cal kit') that you use for calibration. Manufacturers of cal kits often claim that the reference impedance set by their cal kit is real (typically 50 Ohms) and frequency-independent. In some cases, the object's impedance has a frequency-dependent imaginary part, and the calibration algorithm tries to take the (supposedly frequency-independent) real part alone as the reference resistance.

But in general the reference impedance can be complex and/or frequency-dependent. If, for example, you use the TRL (thru-reflect-line) algorithm for calibration, the resultant reference impedance becomes equal to the characteristic impedance of the line standard in the cal kit, which is almost always complex and frequency-dependent. At times (typically in on-wafer measurements), the characteristic impedance itself (and hence the reference impedance too) is unknown. I'm not going to discuss here what to do then…

So the reference impedance comes primarily from a physical object. But it is also possible to use a different value than the one you set by calibration. Such a need typically arises, again, when you use TRL and your original reference impedance is complex and frequency-dependent. Frequency-dependent complex reference impedance is inconvenient for multiple reasons:

o You cannot save the S parameters in a touchstone (.s2p) file because it can only accommodate real and frequency-independent reference resistance.

o Not all simulators can handle complex frequency-dependent reference impedance.

o There are different, incompatible definitions of S parameters when the reference impedance is complex. (Do you know your simulator adopts which definition?)

For these reasons, you'll want to save your S parameters with a real reference resistance (typically 50 Ohms) whenever you move your data from one system to another. You can numerically change the reference impedance of your S parameters to an appropriate value. Thus, the reference impedance may or may not have a physical origin. This is in contrast with quantities like the "characteristic impedance" (of a line standard) or the "load impedance" (of a load standard), which are attributes of physical objects.

Although I don't understand your question clearly either, I hope I have answered at least part of your question.

Shuhei

Are you looking for smith charts? http://www.antenna-theory.com/tutorial/smith/smith_chart.jpg

The Smith Chart is a fantastic tool for visualizing the impedance of a transmission line and antenna system as a function of frequency. Smith Charts can be used to increase understanding of transmission lines and how they behave from an impedance viewpoint. Smith Charts are also extremely helpful for impedance matching.

I thank you all for answering.

But I am still confused.

I am attaching a image to explain what I want to ask.

Please see this image.

I just wanted to know as of why the reference impedance is varying.

And basically what does it actually means.

Is that just a 17ghz high pass filter with harmonics or s11 reflections carrying on?

Impedance of All passive rf component changes as frequency changes.

It is a very natural phenomenon.

I think that your very high impedance is because of resonance.

If you want to know more, you should refer to books about RF basic such as Pozar's Microwave engineering.

Try doing a log plot of freq, that freq response is due to the capacitor. Obviously enough, as ω tends to zero, |Z| tends to infinity, and vice versa. Worse, it implies, quite wrongly, that at frequencies above 20 GHz the impedance is almost constant, and at frequencies below about 15 GHz the impedance is changing too fast to measure, so try a log plot.

Impedance of RF Components do changes on varying frequency.

My basic question is-

What is reference impedance?

How does the simulation software calculates the reference impedance for a given filter design?

Moreover, I am confused as my same design is showing constant matching impedance of 69 ohms whereas when I try to  match it the reference impedance is changing as in the image I have shown in my above reply.

Anand,

Why do you think the quantity you plotted is a "reference impedance," rather than an impedance of a physical thing, such as a filter or whatever you are dealing with?

Reference impedance is not an attribute of a physical object (unless you choose such a value), and you have to name it. Many microwave simulators has a default value of 50 Ohms.  When you simulate transmission lines or waveguides with an EM simulator using the so-called "wave ports" or "waveguide ports," the natural choice of reference impedance would be the complex, frequency-dependent characteristic impedance of the port, which is, unfortunately, not unique and multiple definitions exist.  Again, simulators give you the choice.  You choose from a certain characteristic impedance of your choice, 50 Ohms, or some other value you like.  The default choice will depend on your simulator.

Shuhei,

I am saying that as Reference impedance as the graph reading is of the same quantity.

I am pretty convinced with your answer but still how come my same design is showing two results at the same time.

On one result my Reference Impedance is coming out to be 69 Ohms but on the other design its varying as I have attached in the figure in my previous comment.

Anand,

I don't fully understand what you are tying to tell. It will help if you could provide more specific information and state your question more clearly.  For example, what are you trying to do, what are your designs, which simulator did you use, how did you plot your graphs, what did you expect and why?

Shuhei,

I have been designing Substrate Integrated waveguide filters and while trying to match that with 50 Ohm impedance, I was getting the response I have shown in the image that I had attached previously.And the plots that I get are obtained on simulating the design.

On Y- Axis the value is Reference Impedance, where as on X- Axis frequency is shown.

Can you suggest me more matching techniques that I can use.

Anand, tell me the name of the software. Is it HFSS?

Then the reference impedance is just an impedance that S matrices are referred to (transmission line impedance, port impedance and other names). In physical experiment, you usually take it as VNA port impedance, which is normally a very good reference. VNA's software make all the calculations with respect to that impedance, and actual deviations are corrected with numerical calibration. I e, it is not a reference impedance on the plot. Correct me someone If I am wrong.