Designer’s Guide Community

baab

Senior Member

Offline

Posts: 178
EA

Re: S11 and Z11 problem for input impedance matching
Reply #30 - Mar 17^th, 2014, 7:06am

Hi Aaron.

I replaced almost all components by TSMC 0.13um except for the inductor in parallel with 1K resistor.

However, this is my new problem.

1. Can I use an inductor from TSMC 0.13 and connected it in parallel with a external resistor 1K?
2. If #1 is possible, then I need to use the inductor with inductance > 1uH. As you said, it seems to big. How can I solve this/

Back to top

IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: S11 and Z11 problem for input impedance matching
Reply #31 - Mar 17^th, 2014, 6:01pm

Hi baab,

the only way you can get that 1 uH inductor is to have it off-chip. But you don't need to do that. Instead you should change your matching network. You seem to have chosen a very complicated matching network. How did you choose the values? Why don't you just use one shunt inductor to replace the 1u inductor, and then a series capacitor? If you do this, the value of the inductance will be much more reasonable. Also, the 1kohm resistor can be absorbed into the parasitic resistance of the inductor.

regards,
Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

baab Senior Member Offline Posts: 178 EA	Re: S11 and Z11 problem for input impedance matching Reply #32 - Mar 18^th, 2014, 2:52am Hi Aaron. I did exactly the circuit you suggested. However, because the parasitic resistance too small, the voltage gain S21 is so low. I don't have the picture right now. I will post it later.
Back to top	IP Logged

aaron_do Senior Fellow Offline Posts: 1398	Re: S11 and Z11 problem for input impedance matching Reply #33 - Mar 18^th, 2014, 6:21am Hi baab, that doesn't sound right. If S21 was not good then you most likely chose your component values wrongly. regards, Aaron
Back to top	there is no energy in matter other than that received from the environment - Nikola Tesla IP Logged

baab

Senior Member

Offline

Posts: 178
EA

Re: S11 and Z11 problem for input impedance matching
Reply #34 - Mar 20^th, 2014, 1:53am

Hi, Aaron.
This is my circuit.
I used the circuit as you suggested. However, my problem is
about drain inductor. It should be larger than 1uH to get a good S12, S21, S11, S22. If I change Ld to the one with much smaller inductance then all S parameters will be degraded. I know that the impedance matching networks are not correct.
However, that will completely change my circuit.
Please help me with the following.
1. Is the output matching network above the one you suggested?
2. What are the criteria in choosing coupling capacitors? These coupling capacitors are parts of input and output matching networks. I think we need to choose these capacitors first.
(my frequency is about 1.5GHz)

Back to top

LNA11.png

IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: S11 and Z11 problem for input impedance matching
Reply #35 - Mar 20^th, 2014, 6:09pm

Hi baab,

you made quite a big error there. In your circuit, L0 is shorting the drain of M1 to ground (i.e. the biasing point is wrong).

For the circuit I'm trying to describe, you should remove L17 and use L0 in its place. That way you won't need the 1uH inductor. Also, as you say, C1 will become part of the matching network.

regards,
Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

baab

Senior Member

Offline

Posts: 178
EA

Re: S11 and Z11 problem for input impedance matching
Reply #36 - Mar 20^th, 2014, 10:29pm

Hi, Aaron.

You are right. The simulator caused an error and I put a 1ohms in series with that inductor only for simulation purpose.
I thought that that is only a problem with simulation. In real world, that is not problem at all.

I'd like to ask about coupling capacitors. What are criteria in choosing these ones? Assuming that my frequency is about 1.5GHz. I consulted some papers and calculated the impedance Zc = 1/wC of that capacitors at interested frequencies and the results are different.
Zc ranges from 1ohm to 120 ohms. That is quite big.

Back to top

IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: S11 and Z11 problem for input impedance matching
Reply #37 - Mar 21^st, 2014, 2:02am

Hi baab,

Quote:

The simulator caused an error and I put a 1ohms in series with that inductor only for simulation purpose.
I thought that that is only a problem with simulation. In real world, that is not problem at all.

What do you mean? In the real world it is still a problem. You can't put an inductor where L0 is...

Quote:

I consulted some papers and calculated the impedance Zc = 1/wC of that capacitors at interested frequencies and the results are different.
Zc ranges from 1ohm to 120 ohms. That is quite big.

It depends on your design. The capacitors should be large enough not to cause any significant performance loss. But you don't want them to take up too much area.

regards,
Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

baab

Senior Member

Offline

Posts: 178
EA

Re: S11 and Z11 problem for input impedance matching
Reply #38 - Mar 21^st, 2014, 2:22am

Thank you, Aaron.

Quote:

What do you mean? In the real world it is still a problem. You can't put an inductor where L0 is...

In real world, L0 is not purely inductive. Therefore, there is resistance and no DC short there. This is what I thought.

Quote:

It depends on your design. The capacitors should be large enough not to cause any significant performance loss. But you don't want them to take up too much area.

Yes, that is what I confused. Based in layout, I think I can estimate the upper limitation for these capacitors. However, how about lower limit? I am still not sure how to calculate it.
The larger the capacitance the smaller its reactance and the better performance.
How to know that a capacitance causes significant performance loss or not?

For example, my frequency is 1.5GHz and C = 1pF. Is that good?

Back to top

IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: S11 and Z11 problem for input impedance matching
Reply #39 - Mar 21^st, 2014, 6:39pm

Hi baab,

Whether L0 is purely inductive or not it is gonna mess up your DC operating point.

Taking C0 as an example, it is in series with L2. L2 and C0 will resonate at some frequency, below which the combination will act like a capacitor rather than an inductor. So make sure the resonance is much lower than your operating frequency. Just make sure there are problems with your s-parameters.

Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

vedat

New Member

Offline

Posts: 1

Re: S11 and Z11 problem for input impedance matching
Reply #40 - Apr 29^th, 2014, 1:38pm

@baab, an inductor placed in that particular will give a short. I understand that practically speaking an inductor has got some resistance but it doesn�t matter when we�re talking about voltages greater than 1V. It�s much like connecting two terminals of a battery with a simple wire, it has some resistance too but it is negligible in that case. I�ve seen in this thread that people are trying to adjust S21 just by tuning one component, it is important for people to note that it won�t work that way, you need to adjust the values of all the components in the path.

Back to top

IP Logged

totowo

Junior Member

Offline

Posts: 13
SH, PRC

Re: S11 and Z11 problem for input impedance matching
Reply #41 - Nov 16^th, 2014, 6:25pm

The reply is too long......I have no patience to read it all.
But there is some important points:

1. Z11 means the input impedance when I2=0, which means the other port OPEN. (just think about the definition of Z-param matrix)

2. S11 could use to calculate the input impedance (actually, ZM1) with a true 2nd port in simulation. In cadence simulation, the Gamma_L (or ρ_L in some pictures above) is always equal to zero, which means, Gamma_s (or ρ_L) is always equal to S11.

3. for a 2-port network, Y/Z paramter matrix is defintely uncorrelated to PORT impedance in simulation because it's the network characteristic paramter . But S parameter matrix is correlated with PORT impedance because cadence use Z0=Zs(or ZL) to uniform the matrix.

Back to top

IP Logged