Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

1 db Compression Point in Amplifier Design

Status
Not open for further replies.

skynet

Member level 5
Joined
Apr 18, 2002
Messages
80
Helped
2
Reputation
4
Reaction score
1
Trophy points
1,288
Activity points
638
1 db compression point

I am newbie in RFIC design. I want to know about 1 db compression compression of in LNA, got anyone that can explain to me what is 1db compression point and it effect toward LNA design, if someone got good material on this chapter, I would be appreaciate it
 

DJ

Advanced Member level 4
Joined
Jul 22, 2004
Messages
118
Helped
2
Reputation
4
Reaction score
1
Trophy points
1,298
Activity points
1,385
db compression ampli

Just search P1dB in thr Forum. We taled about that issue alot.
 

BigBoss

Advanced Member level 5
Joined
Nov 17, 2001
Messages
5,007
Helped
1,507
Reputation
3,012
Reaction score
1,346
Trophy points
1,393
Location
Turkey
Activity points
30,204
large 1db compression point

In LNA designs , P1dB compression point that is nonlinearity figure of merit is not used. Instead of it, IP3 figure of merit is used to characterize of LNA performance.

IP3 is a very important to investigate the behaviour of a LNA that is under large signal.
 

dsjomo

Full Member level 2
Joined
Oct 24, 2003
Messages
149
Helped
21
Reputation
42
Reaction score
2
Trophy points
1,298
Location
Taiwan
Activity points
1,979
testing one db compression point

??? Is 1-dB point not an important figure of merit ???...... this is the 1st time i hear such an assertion.

There are some out-band blocker specification to the whole receiver, while IP3 is measured as an in-band characteristic with 2-tone test, an out-band interferer(also know as BLOCKER) could degenerate the signal gain of LNA and lower down the sensitivity.

For most case, the difference between IIP3 and input 1-dB compression point is about 8~10 dB. With some special tenique, we can improve IIP3. But the difference between IIP3 and 1-dB point may not be 8~10dB. If you desided not to take 1dB point as a spec. when you design the system, watch out if you pass the out-band blocker test or not...

As the original question about "IIP3 circle"....... I've seen one paper mentioning this strange methodology, but i suggest you take a glance at K.L.Feng 's paper which discussing IM3 characteristic of common emitter stage. And you'll find that -- different type matching network will give rise to different IM3 characteristic. It depends on what is the impedance looking into the signal source at freqency (w2-w1). For most case of LNA, there is a DC blocking capacitor at input port, and this cap. limits the impedance looking backward. Feng proposed a solution in another IEEE paper .

As for 1-dB point, some stretagy i know to improve it is as follows-- DC current, trasistor sizing, inductive RFC bias choke, dual bias feed (proposed by Japanese, but this is a modified version of PA linearizing technique),etc...
or.....just put a resistor shunting at the input port, this is the fastest and easiest way. 8)
 

DJ

Advanced Member level 4
Joined
Jul 22, 2004
Messages
118
Helped
2
Reputation
4
Reaction score
1
Trophy points
1,298
Activity points
1,385
1 db compression in rf ic design

The question was related to the meaning of P1dB. dsjomo talked about Ip3 too and I would like to add something.
P1dB and Ip3 come from different nonlinearities. P1dB is measured when the input signal saturates the device by 1dB. Looking at the I/V curve the RF signal might cut at the saturation and also at I=0. The fundamental harmonic is suppressed but other harmonics rise. Ip3 is measured in small signal somewhere at the middle of the I/V curve and the internal nonlinearity create the 2f1-f2 and 2f2-f1. So the P1dB nonlinearity comes from the cut off of the RF signal at the two ends, but the IP3 nonlinearity comes from the internal non linearity of the device. Usually we take Ip3 as 10dB more than P1dB but it is not correct. Agilent has advanced Phemt with about 14 dB different. Also one can simulate and find that for different load circuit there is different numbers.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top