Filter input imepdance beginners questions

[ku637]

Dear Geeks,

I have got a simple filter structure..like shown below..

INPUT    L1
------(((((()--------------------------       OUTPUT
                          |
                          | 
                         __C1
                         ---
             L2          |
------(((((()-------------------------

I would like to analyse its input impedance in Agilent ADS .. can anyone please tell me how to do it?


Input is the inductor side.. Output is the capacitor side..


Problem i face is ..

1.I connect port 1 with 50Ohm termination in between the input side inductors & port 2 on the output side across the capacitor ..
2.Now how to properly connect the GND symbol?
3.If i connect it at both input and output ports it will short the L2 inductor


Can somebody please help me...??

Regards

 

[volker_muehlhaus]

For the filter, it makes no difference if L2 is in the signal or return path, because both currents are dependent (isignal = -ireturn).

You can simulate your filter like this

or like this. Both are the same.

In my opinion, what you really want for a low pass filter is L2 between the shunt C and the output.

 

[ku637]

Dear volker,

If i understand you rightly i think what you are saying is like i shall put a single inductor whose value will be say L1+L2 and a capacitor C1 and simulate as a simple LC filter..?
is that so.
thanks for any help,

regards

 

[volker_muehlhaus]

put a single inductor whose value will be say L1+L2 and a capacitor C1 and simulate as a simple LC filter..?

Yes, that's it.

 

[tony_lth]

Is your signal differential signal? If so, you can divide it into two branch. And only analysis one branch is enough.
Onr branch includes L1, and shunt C1/2, then simulates with both 50 ohms terminal.

 

[ku637]

Actually this is for an EMI filter in the AC mains .. i need to find the input impedance of the same.. for upto 100MHz. And yes Thats for the differential mode noise signals..

 

[volker_muehlhaus]

Actually this is for an EMI filter in the AC mains ..

That's important because it means that source and load impedance are different from the 50 Ohm used in RF design.
I have no experience with EMI filter design, but would suspect that a low source impedance is assumed. Not sure about the load impedance.

 

[aggpankaj2]

I too agree with Tony and volker's view.. instead of two inductance assume it one dividing cap value by two for differential signalling..

 

[ku637]

That's important because it means that source and load impedance are different from the 50 Ohm used in RF design.
I have no experience with EMI filter design, but would suspect that a low source impedance is assumed. Not sure about the load impedance.

Thanks for pointing out..But we are assuming a 100Ohm as the line impedance..

 

[tony_lth]

You can do it with only single ended design, then copy it to each branch of balanced design. That is a dummy banlanced design, always use to filter EMI.

 

[drkirkby]

That's important because it means that source and load impedance are different from the 50 Ohm used in RF design.
I have no experience with EMI filter design, but would suspect that a low source impedance is assumed. Not sure about the load impedance.

It's common to quote the impedance at 50 Ohms, even as you say, the load is not 50 Ohms. But realistically you can't say what the load impedance will be at RF frequencies, even if you know the voltage and current at 50 or 60 Hz.

I used to work for a company that made huge filters for EMP protection. One filter I know had an attenuation of 120 dB. But this huge attenuation was achieved by the use of an inductor on a core. At low powers, as one would use on a spectrum or network analyzer, the core would work as intended. But at high current, the core would saturate and so the inductance fall, and so the attenuation fall. I always thought that was a bit of a con.

Dave