| Author |
Message |
lijulia
Joined: 05 Jan 2003
Posts: 78
|
 29 Sep 2004 22:48 w-cdma lna |
|
|
|
|
I design cascode LNA for wcdma, and added the bypass switch between the input and output , but the s11 & s22 change a lot from high-gain mode. how to solve this problem?
Thanks!
|
|
| Back to top |
|
 |
dhlee
Joined: 20 Apr 2004
Posts: 41
Helped: 4
|
30 Sep 2004 1:56 lna bypass mode |
|
|
|
|
| In my opinion, at high-gain mode lowering the reflection coeff. of LNA may be meaningless. Because the input signal is large enough at the bypass mode, the power loss induced by the poor VSWR of LNA does not affect system performance.
|
|
| Back to top |
|
|
djalli
Joined: 10 Nov 2001
Posts: 887
Helped: 15
Location: 1600 Pennsylvania Avenue, Washington DC 20500
|
30 Sep 2004 6:51 bypass mode loss |
|
|
|
|
Can you post the schematic. I am interested!
thank you
djalli
|
|
| Back to top |
|
|
lijulia
Joined: 05 Jan 2003
Posts: 78
|
30 Sep 2004 21:40 wcdma & pdf |
|
|
|
|
| sure, the input power is high for bypass mode is high. But the performance of the duplexer won't be good if LNA input is not matched to 50ohm, right?
|
|
| Back to top |
|
|
Google
AdSense
|
30 Sep 2004 21:40 Ads |
|
|
|
|
|
|
| Back to top |
|
|
vfone
Joined: 10 Oct 2001
Posts: 2327
Helped: 326
|
30 Sep 2004 21:53 lna bypass design |
|
|
|
|
This document will help you:
http://cp.literature.agilent.com/litweb/pdf/5988-4278EN.pdf
rgds
|
|
| Back to top |
|
|
dsjomo
Joined: 24 Oct 2003
Posts: 154
Helped: 15
Location: Taiwan
|
04 Oct 2004 11:50 wcdma lna |
|
|
|
|
I guess, you tide your switch to an high-Q output node, and input impedance varies because output suffers process variation or inaccurate modelling or parasitics.
For a transistor operating under high frequency, Gamma-in is not only a function of S11 but also a function of S22. And so does Gamma-out, this is because of the Cbc capacitance. High-Q node is easy to be affected by inacurrate modelling or parasitics. So both Gamma-in and Gamma-out present totally different value compared to the designed value.
The followings are my suggestions......
1) Lower down the node Q which you tide your switch to. For example, shunting a resistor.
2) Current steering method. But whatch out the low gain linearity. For WCDMA, this mighy not be a applicable way.
3) um...... this might be the most effective way, but here i can only give you some hints.
Hint #1 : For a cascode LNA the upper stage is common-base stage AMP.
Hint #2: The mag. of S12 of common-base AMP is vary small.
Hint #3: Common-base AMP could be treated as isolated 2-port network.
Hint #4: For an isolated 2-port, Gamma-in is no more a funcion of S22.
Wish these could help you. 
|
|
| Back to top |
|
|
dhlee
Joined: 20 Apr 2004
Posts: 41
Helped: 4
|
05 Oct 2004 10:49 inductorless lna wcdma |
|
|
|
|
lijulia said:
sure, the input power is high for bypass mode is high. But the performance of the duplexer won't be good if LNA input is not matched to 50ohm, right?
==> What performance of duplexer did you refer to?
the insertion loss?? filtering characteristics??
I think neither maybe affect the system performance in high input power.
BR.
|
|
| Back to top |
|
|
