A question about the input match of a LNA

I'm in trouble with the input match of a LNA (Fig. 1). The impedance of Zin1 is very small because of the capacitors Cb and Cp. Cb is the parasitical capacitor of the bondwire and the package, and Cp is the parasitical capacitor of the pad and the ESD. Althought I can convert real(Zin) to 50 Ohm, impedance match and noise match can not achieve simultaneously.
How can I realize impedance match and noise match simultaneously?
**Note:1 The frequency is 1.5GHz.
2 Low power is desired, so W/L of M1 can not be very large.
3 Ls is an on-chip inductor.
Thanks.
[/img]

How can I realize impedance match and noise match simultaneously?

Click to expand...

It can be possible with LS which can drive both noise match and input match to 50 but you loose gain.
What is the gain that the transistor support and try a 3 dB lossy such that both noise and input match is met.
The best way is to have gain circles and noise circles for LS sweep and decide.

heritage said:

I'm in trouble with the input match of a LNA (Fig. 1). The impedance of Zin1 is very small because of the capacitors Cb and Cp.

Click to expand...

50 Ohm @ 1.5 GHZ =>2 pF. If Zin is small due to capacitive load, what range of capacitance do you have? Several pF mismatch on a bond wire seems too much.

/Alf

How low is the power that you want it to be?
To improve the noise you can increase the bias current

To improve the noise you can increase the bias current

Click to expand...

Blindly increasing the bias current will lead in NF degradation. It should be matched to the technology. NF is function of bias current is like a dwell. if the current is too small then also NF is high.
[/quote]

Thanks for your reply.
The LNA is differential, and all current is 2mA, so each branch is 1mA. As I known, Cp is about 200fF, Cb is about 260fF, Lb is 1~2nH. Cp, Lb, Cb make real(Zin) small.
When the bondwire is not added, the NF is 1.4dB. Then NF is 2dB after the bondwire is added and some inductors and capacitors are used to match.

Added after 9 minutes:

E Kafeman said:

heritage said:

I'm in trouble with the input match of a LNA (Fig. 1). The impedance of Zin1 is very small because of the capacitors Cb and Cp.

Click to expand...

50 Ohm @ 1.5 GHZ =>2 pF. If Zin is small due to capacitive load, what range of capacitance do you have? Several pF mismatch on a bond wire seems too much.

/Alf

Click to expand...

The impedance match (50 Ohm) is needed for the LNA. I mean the real(Zin) is smaller than 50 Ohm, so it should be converted to 50 Ohm by L or T network.

Added after 3 minutes:

kspalla said:

How can I realize impedance match and noise match simultaneously?

Click to expand...

It can be possible with LS which can drive both noise match and input match to 50 but you loose gain.
What is the gain that the transistor support and try a 3 dB lossy such that both noise and input match is met.
The best way is to have gain circles and noise circles for LS sweep and decide.

Click to expand...

The inductor Ls can't be too large, because it is on the chip. A large Ls has a large serial parasitical resistor, it will degenerate the noise performance.

?

Try with paralleling the bond wires to reduce the effect.