Designer’s Guide Community

purplewolf

Senior Member

Offline

Posts: 103

Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Jun 22^nd, 2010, 3:40am

Hello,
I am trying to simulate the OIP3/IIP3 of class B RF power amplifier using Cadence Spectre ..
I have tried the following analysis
1)QPSS with one large and one moderate tone
2)PSS with two large tones
3)PAC +PSS

I sweep the input power from -60dbm to 30 dbm whereas my amplifier inout referred one db (Pin 1dB) is at 20 dbm.
So, according to theory my IIP3 must be 20+9.6 or greater than that factor but I am getting IIP3 of 14 dBm ???

Secondly , does large signal means that Spectre analysis the circuit interms of harmonic balance techniques(strongly non-linear) while moderate or small signal means analysis interms of volterra series expansion (weakly non-linear)..

Do i need to provide different power levels if i choose one tone as large while other as moderate??

What exactly IIP3 OIP3 means in terms of class-B power amplifies? Does the theory in books regrding LNA,Mixer IM products apply here?

I have tried every possible way to solve this issue and searched lots of forums and manuals but dont know how to fix it??

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Reply #1 - Jun 22^nd, 2010, 4:09am

Are your questions focusing on very specific vendor's simulator, Cadence Spectre ?
If so, post to simulator's board not design's board.

purplewolf wrote on Jun 22^nd, 2010, 3:40am:

I sweep the input power from -60dbm to 30 dbm whereas my amplifier inout referred one db (Pin 1dB) is at 20 dbm.
So, according to theory my IIP3 must be 20+9.6 or greater than that factor but I am getting IIP3 of 14 dBm ???

What scheme do you use for input and output in Class-B amplifier ?
Class-B Push-Pull using 180deg-Hybrid or transformer ?
If so, your result of IIP3 is relative reasonable, since 20+9.6-12=17.6dBm is close to 14dBm.

purplewolf wrote on Jun 22^nd, 2010, 3:40am:

while moderate or small signal means analysis interms of volterra series expansion (weakly non-linear)..

No. Both moderate and small signals are not volterra series expansion.
And small signal in PSS/PAC or QPSS/QPAC is pure linear analysis around periodical or quasi-periodical steady state operation points.

purplewolf wrote on Jun 22^nd, 2010, 3:40am:

Do i need to provide different power levels if i choose one tone as large while other as moderate??

No.

But valid signal level for small signal in PSS/PAC and QPSS/QPAC is 10dB smaller than P1dBin(Input Refered 1dB Gain Compression Point) as Class-A amplifer.

You had better set level of moderate signal in QPSS smaller than P1dBin(Input Refered 1dB Gain Compression Point) as Class-A amplifer especially for Shooting-Newton QPSS.

See threads after following.
http://www.designers-guide.org/Forum/YaBB.pl?num=1262828862/18#18

Back to top

« Last Edit: Jun 22^nd, 2010, 8:48am by pancho_hideboo »

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #2 - Jun 22^nd, 2010, 4:35am Its a single ended design with input matching,stability , harmonic tank and output matching network.
Back to top	IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Reply #3 - Jun 22^nd, 2010, 4:39am

purplewolf wrote on Jun 22^nd, 2010, 4:35am:

Its a single ended design with input matching, stability, harmonic tank and output matching network.

What is it ?

Use the followings. Don't use HB-SS or PSS/PAC.

(1) Two-tones HB based on one fundamental frequency in Agilent ADS
(2) Two-tones CR(=HB) based on one fundamental frequency in Agilent GoldenGate
(3) Two-tones HB-PSS in Cadence Spectre

(4) Two-tones HB based on two fundamental frequencies in Agilent ADS
(5) Two-tones CR(=HB) based on two fundamental frequencies in Agilent GoldenGate
(6) Two-tones HB-QPSS in Cadence Spectre

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #4 - Jun 22^nd, 2010, 4:52am A Parallel RC network in Series with the gate to ensure stability . The input matching is done after stability network.
Back to top	IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Reply #5 - Jun 22^nd, 2010, 4:54am

purplewolf wrote on Jun 22^nd, 2010, 4:52am:

A Parallel RC network in Series with the gate to ensure stability . The input matching is done after stability network.

I got it.
See the followings.
http://www.designers-guide.org/Forum/YaBB.pl?num=1218635777/5#5
http://www.designers-guide.org/Forum/YaBB.pl?num=1272344339/3#3

purplewolf wrote on Jun 22^nd, 2010, 3:40am:

I am trying to simulate the OIP3/IIP3 of class B RF power amplifier using Cadence Spectre ..
I have tried the following analysis
1)QPSS with one large and one moderate tone
2)PSS with two large tones
3)PAC +PSS

Do you get same result as IIP3 in any of these three analyses.

When I have spare time, I will check P1dBin and IIP3 for Class-B amplifier using Agilent ADSsim although I also have Cadence Spectre.

See Class-B Power Amplifier in the following.
http://edocs.soco.agilent.com/display/ads2009/About+Amplifier+DesignGuide
Class B > Spectrum, Gain, TOI and 5thOI Points, and PAE vs. Power ; "HB2TonePAE_Pswp_ClassB.dds"

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #6 - Jun 22^nd, 2010, 5:10am but what does stability network has to do with IIP3 and OIP3?? the amplifier is already stable
Back to top	IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #7 - Jun 22^nd, 2010, 5:10am yes i get the same result for all three analysis...
Back to top	IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Reply #8 - Jun 22^nd, 2010, 5:14am

purplewolf wrote on Jun 22^nd, 2010, 5:10am:

but what does stability network has to do with IIP3 and OIP3??
the amplifier is already stable

There is no relation generally

purplewolf wrote on Jun 22^nd, 2010, 5:10am:

yes i get the same result for all three analysis...

It is very strange.

I will check P1dBin and IIP3 for Class-B amplifier using Golden Standard RF Circuit Simulator Agilent ADSsim.
If you show me netlist of your Class-B amplifier, I can resimulate it using various simulators.

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #9 - Jun 22^nd, 2010, 6:12am i have generated the netlist but which file you would be needing to simulate it...and where to find the netlist file?
Back to top	IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #10 - Jun 22^nd, 2010, 7:10am here the netlist file... change extension to .scs
Back to top	input.jpeg IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #11 - Jun 22^nd, 2010, 11:01am here are the files.
Back to top	qpss_two_tone.jpeg IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #12 - Jun 22^nd, 2010, 11:03am thats the other
Back to top	pac_pss.jpeg IP Logged

purplewolf Senior Member Offline Posts: 103	Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre Reply #13 - Jun 22^nd, 2010, 2:50pm this is a device in umc 90nm... i think its parameters dont appear in netlist.. how to get it??
Back to top	IP Logged

RFICDUDE

Community Fellow

Offline

Posts: 323

Re: Class B RF Power Amplifier ,IIP3/OIP3 Spectre
Reply #14 - Jun 22^nd, 2010, 4:03pm

Dude, you can't share proprietary foundry models and/or parameters.
Unless you have legal permission from the foundry to do so (usually requires a three way nondisclosure agreement).

Sometimes the EDA company has existing agreements with foundries, so in these cases you can share netlists with someone from the EDA company and they can simulate your netlist because they have or can obtain the foundry models.

Back to top

IP Logged