| Author |
Message |
Mr. D.
Joined: 27 Nov 2003
Posts: 33
Helped: 2
Location: Switzerland
|
 02 Dec 2003 14:44 Port Impedances |
|
|
|
Hi all
S-parameters are always related to a certain entry- or exit -port impedance. How can i change these impedances?
Thanks and bye
D.
|
|
| Back to top |
|
 |
flatulent
Joined: 19 Jul 2002
Posts: 4856
Helped: 292
Location: Middle Earth
|
| 02 Dec 2003 17:00 Re: Port Impedances |
|
|
|
| There are algebraic equations that convert to values with different source and load values. I have seen them in text books, but do not remember them. Try looking through some books in your library.
|
|
| Back to top |
|
|
colombo2
Joined: 02 Dec 2003
Posts: 35
|
| 02 Dec 2003 19:01 Re: Port Impedances |
|
|
|
Mr. D.
What do you mean by changing the value of the S paramters? Do you mean convert the parameter to others like Y paramters for example or what?
Colombo2
|
|
| Back to top |
|
|
flyhigh
Joined: 12 Nov 2003
Posts: 512
Helped: 25
Location: somewhere over the rainbow
|
| 03 Dec 2003 14:16 Re: Port Impedances |
|
|
|
Hi,
S pars are defined (normalized) to some convinient characteristic impedance of ports of the circuit. It is usually 50 Ohms, but they can be renormalized to any other impedance of the ports by proper S matrix manipulation. As stated above, you can find this at many text books, it is usually covered at the very beginning of a book as it is a fundamental issue. You can also shift referent planes at which S parameters are defined - again S matrix is manipulated.
Any more details on this topic is for you to learn yourself, these are the basics of RF/microwave theory!
flyhigh
|
|
| Back to top |
|
|
caocao
Joined: 23 Nov 2003
Posts: 117
Helped: 3
|
| 04 Dec 2003 14:27 Re: Port Impedances |
|
|
|
Mr.D,do you want to change the reflected coeffience?s11=(Zl-Zc)/(Zl+Zc);
you must make the Zl~=Zc.and look for S par in textbook or Refrence
[quote="flyhigh"]Hi,
S pars are defined (normalized) to some convinient characteristic impedance of ports of the circuit. It is usually 50 Ohms, but they can be renormalized to any other impedance of the ports by proper S matrix manipulation. As stated above, you can find this at many text books, it is usually covered at the very beginning of a book as it is a fundamental issue. You can also shift referent planes at which S parameters are defined - again S matrix is manipulated.
|
|
| Back to top |
|
|
Mr. D.
Joined: 27 Nov 2003
Posts: 33
Helped: 2
Location: Switzerland
|
| 05 Dec 2003 12:27 Re: Port Impedances |
|
|
|
Thanks to u all.
I think a have to change the question. The s-parameters are not the problem, but:
How do i change the value of the impedance of a port?
I assume that it's set to 50 ohms by default, correct?
Thanks
D.
|
|
| Back to top |
|
|
Redi
Joined: 16 Dec 2002
Posts: 77
|
| 05 Dec 2003 18:29 Re: Port Impedances |
|
|
|
| Mr. D. wrote: |
Thanks to u all.
I think a have to change the question. The s-parameters are not the problem, but:
How do i change the value of the impedance of a port?
I assume that it's set to 50 ohms by default, correct?
Thanks
D. |
Hi,
The question is.. where do you want to change port impedance?
Greets,
Redi.
|
|
| Back to top |
|
|
colombo2
Joined: 02 Dec 2003
Posts: 35
|
| 05 Dec 2003 23:28 Re: Port Impedances |
|
|
|
The port impedance is set by maximum power transfer (conjugate matching). Typically 50 Ohms. However, impedance transformers are usually adopted to match and alter the port impedance. Common-Source amplifiers with degenration are a prime example where impedance transformation is adopted to suppress noise figure.
Colombo2
|
|
| Back to top |
|
|
Mr. D.
Joined: 27 Nov 2003
Posts: 33
Helped: 2
Location: Switzerland
|
| 08 Dec 2003 13:12 Re: Port Impedances |
|
|
|
thanks!
D.
|
|
| Back to top |
|
|
cheng
Joined: 01 Jan 1970
Posts: 111
Helped: 2
|
| 09 Dec 2003 12:05 Re: Port Impedances |
|
|
|
Here is my 20 cents:
Port impedance is sort of artificial. It is largely detemied by the field configuration and sometimes can yield bogus results, especially if you cascade (network level) to structures where the interface will exhibit field discontinuities. It is defined basically to avoid thinking in fields and in networks instead. In each major program I know there is always the option to renormalize the available S-parameters to different values. You simply have to go through the manual and find it. Or if you want I can send you the formula to do it manually.
When dealing with waveguides, one should resort to the GSM, where each mode carries its own impedance (hense the name) and therefore be carefull when cascading multi modal devices.
From your post it is not clear to me, what structure you are dealing with. If that is a waveguide, you should not, unless abs. necessary change impedances. Try working on a GSM level. If it is microstrip or other suitable structure the redefinition is trivial.
regards,
cheng
|
|
| Back to top |
|
|
Mr. D.
Joined: 27 Nov 2003
Posts: 33
Helped: 2
Location: Switzerland
|
| 12 Dec 2003 12:14 Re: Port Impedances |
|
|
|
thanks mr. cheng.
What we are trying to do is to verify some formulas and an equivalent simplified cirquit (so that they can be simulated in "Microwave Office") of a "discontinuity". That is a sudden jump in diameter of the inner conductor of a "coax" conductor. The assumption (and literature suggests that too) is, that the jump can be modelled by a capacitor.
The wave impedances of the conductors are not 50 ohms. The point now is that the reflection on the ports are so big (due to the different impedances of the ports and the "coax") that the refletions induced by the jump are no longer "visible".
So the goal is to change the port impedances so that we can eliminate the reflections caused by the ports.
THanks for ur help.
D.
|
|
| Back to top |
|
|
cheng
Joined: 01 Jan 1970
Posts: 111
Helped: 2
|
| 15 Dec 2003 10:37 Re: Port Impedances |
|
|
|
I am not sure I completely understand,yet, will try to provide some info. Seems that you have 2 coax cables, each different that 50 Ohms and abrupt change in between. My first thought will be to use them (even in MWO) with their respective port impedances and not normalizing to 50 Ohms. You need not normalize them to 50 Ohms unless you will measure them by Network Analyzer.If you only want to numerically extract the discontinuity I'd think of working with the GSm and not 50 Ohms. Sometimes engineers are obsessed with 50 ohms, but it is rarely necessary.
Hope it helps.
If you want you can PM me with your structure and I will have a look.
Cheng
|
|
| Back to top |
|
|
Mr. D.
Joined: 27 Nov 2003
Posts: 33
Helped: 2
Location: Switzerland
|
| 22 Dec 2003 22:20 Re: Port Impedances |
|
|
|
thanks again!
we now solved the problem. we found out, that the "renormalize"-option in the post processor does exactly what we needed: it normaizes the s-params to a certain port impedance. that was it!
thanks a lot.
bye
D.
|
|
| Back to top |
|
|
