- 9th October 2010, 09:51 #1

- Join Date
- Sep 2010
- Location
- bangalore
- Posts
- 48
- Helped
- 1 / 1
- Points
- 625
- Level
- 5

## IIP3 and 1db compression point(P1dB) of LNA

From the nonlinearity analysis we know that there is a proportional relationship beetween iip3 and P1dB(1dB compression point).i.e., if IIP3

is high P1dB also high.But when I simuting my LNA for various IIp3 values iam getting same P1dB.IIP3 varying from 12dBm to 19dBm but P1dB remains at -8dBm.what may be the reason.iam using cadence tool and spectre simulator.can any one help? Thanks in advance.

- 9th October 2010, 09:51

- 17th October 2010, 11:20 #2

- Join Date
- Mar 2002
- Location
- Germany
- Posts
- 25
- Helped
- 1 / 1
- Points
- 2,275
- Level
- 11

## Re: IIP3 and 1db compression point(P1dB) of LNA

Did you use a bipolar transistor and vary the collector current and nothing else?

- 17th October 2010, 13:21 #3

- Join Date
- Sep 2010
- Location
- bangalore
- Posts
- 48
- Helped
- 1 / 1
- Points
- 625
- Level
- 5

- 17th October 2010, 15:26 #4

- 17th October 2010, 16:37 #5

- Join Date
- Mar 2002
- Location
- Germany
- Posts
- 25
- Helped
- 1 / 1
- Points
- 2,275
- Level
- 11

## Re: IIP3 and 1db compression point(P1dB) of LNA

P1dB is measured at output and not at input? P1dB - at output measured - should be proportional to P-DC. If this is not the case your mosfet goes into voltage saturation or current saturation and not both simultaneously. This is due to output mismatch.

- 17th October 2010, 16:37

- 18th October 2010, 08:02 #6

- Join Date
- Nov 2006
- Location
- Shanghai
- Posts
- 410
- Helped
- 37 / 37
- Points
- 3,868
- Level
- 14

## Re: IIP3 and 1db compression point(P1dB) of LNA

i think the formula P1dB=IIP3-9.6dB is derived from Taylor series analysis

maybe in spectre simulation, the IIP3 and P1dB is derived from its transform function accordingly, not pure Taylor series.

- 18th October 2010, 14:23 #7

- Join Date
- Jun 2010
- Location
- Belarus
- Posts
- 190
- Helped
- 52 / 52
- Points
- 1,744
- Level
- 9

## Re: IIP3 and 1db compression point(P1dB) of LNA

Hi

Could you clarify more the structure of your circuit?

What is the input power of tones in iip3 analysis?

I think the proportional relationship appears only if we assume that coefficients of power series of transfer function don't change with input power level - it's not the case when we measure P1db - dc bias point may change significantly.

So may be the reason of such behaviour is that during iip3 and p1db measurements input power levels seen by circuit differs very much - so there are different sources of nonlinearity.

Regards.

- 18th October 2010, 16:10 #8

- Join Date
- Sep 2010
- Location
- bangalore
- Posts
- 48
- Helped
- 1 / 1
- Points
- 625
- Level
- 5

- 18th October 2010, 17:11 #9

- Join Date
- Jun 2010
- Location
- Belarus
- Posts
- 190
- Helped
- 52 / 52
- Points
- 1,744
- Level
- 9

- 19th October 2010, 16:19 #10

- Join Date
- Sep 2010
- Location
- bangalore
- Posts
- 48
- Helped
- 1 / 1
- Points
- 625
- Level
- 5

- 21st October 2010, 15:30 #11

- Join Date
- Mar 2002
- Location
- Germany
- Posts
- 25
- Helped
- 1 / 1
- Points
- 2,275
- Level
- 11

## Re: IIP3 and 1db compression point(P1dB) of LNA

Did you measure P-DC when measuring P1dB? Is P-DC constant or raises your drain current by probably two times or three times when you raise your input RF power? I don´t know know what derivative superposition is. Can you please explain?

Thanks.

- 27th July 2011, 23:02 #12

- Join Date
- Nov 2004
- Posts
- 93
- Helped
- 4 / 4
- Points
- 2,298
- Level
- 11

## Re: IIP3 and 1db compression point(P1dB) of LNA

I guess the question really is "why derivative superposition (DS) does not help P1dB?"

DS is used to cancel out the 2nd derivative of transconductance, and the P1dB is also caused by this 2nd derivative (generally). So it should help both the IP3 and P1dB, shouldn't it?

Any follow-up?