Designer’s Guide Community

mustadem

Junior Member

Offline

Posts: 11

P1dB simulation
Apr 03^rd, 2014, 1:00am

Hi everyone,

Could anybody help me with how to obtain P1dB point from PSS simulation. My LNA is a wide band LNA that operates at multibands of 60, 70 and 80GHz with a bandwidth of about 25GHz. How do I set the beat frequency for the simulation. �I have access to the attached document (www2.ece.ohio-state.edu/~bibyk/ece822/SpectreRF_0728.pdf‎) but their frequency of simulation is at lower frequency. When I chose different beat frequency for the simulation, I got different results and in the results, the output power is lower than the input power which is not a reflection of the gain of the LNA.

I would appreciate you comments.
Thanks

Note: the attachement is too large so here is the link
www2.ece.ohio-state.edu/~bibyk/ece822/SpectreRF_0728.pdf‎

Back to top

IP Logged

aaron_do Senior Fellow Offline Posts: 1398	Re: P1dB simulation Reply #1 - Apr 3^rd, 2014, 2:47am Hi, as long as the input frequency is some harmonic of the beat frequency you should be fine, including setting the beat frequency equal to the input frequency (just click auto). If you want a better answer, you need to post more information. regards, Aaron
Back to top	there is no energy in matter other than that received from the environment - Nikola Tesla IP Logged

mustadem

Junior Member

Offline

Posts: 11

Re: P1dB simulation
Reply #2 - Apr 3^rd, 2014, 7:02am

Thanks Aaron.

I am sending a few snapshots to show how I am doing the setup.

The LNA is a three stage LNA and frequency of the input signal set to 70GHz as shown in the attachment.

The settings for the pss simulation is also shown. The first setting is shown in one of the attached snapshots. In the second setting, I changed the beat frequency from 70GHz to 10GHz.

The result of the 70GHz case shows no intercept while that of the second shows intercept.

So is my simulation correct?
Is it normal to have different results with different beat frequency?

I appreciate feedback from everyone.

Thanks

Back to top

P1dB.rar

IP Logged

sheldon

Community Fellow

Offline

Posts: 751

Re: P1dB simulation
Reply #3 - Apr 3^rd, 2014, 8:01pm

Mustadem,

� The information you are referring to is quite old. Since you are
working at quite high frequencies, using lumped element components
is difficult. So it may be better to use the harmonic balance engine.
Recent versions of Spectre RF releases include information,
application notes, on simulating LNA with Harmonic Balance �
analysis in the documentation. In addition, there have been some
enhancements to accelerate the simulation of P1dB in the most
recent release that you might find useful.

� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � Best Regards,

� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � Sheldon

Back to top

IP Logged

aaron_do Senior Fellow Offline Posts: 1398	Re: P1dB simulation Reply #4 - Apr 3^rd, 2014, 8:02pm Why don't you post the pictures directly? A lot of people won't want to download files (which could contain anything...). Aaron
Back to top	there is no energy in matter other than that received from the environment - Nikola Tesla IP Logged

mustadem Junior Member Offline Posts: 11	Re: P1dB simulation Reply #5 - Apr 4^th, 2014, 3:40am aaron_do wrote on Apr 3^rd, 2014, 8:02pm: Why don't you post the pictures directly? A lot of people won't want to download files (which could contain anything...). Aaron
Back to top	Screenshot-1_010.png IP Logged

mustadem Junior Member Offline Posts: 11	Re: P1dB simulation Reply #6 - Apr 4^th, 2014, 3:43am mustadem wrote on Apr 4^th, 2014, 3:40am: aaron_do wrote on Apr 3^rd, 2014, 8:02pm: Why don't you post the pictures directly? A lot of people won't want to download files (which could contain anything...). Aaron
Back to top	Screenshot-2_005.png IP Logged

mustadem Junior Member Offline Posts: 11	Re: P1dB simulation Reply #7 - Apr 4^th, 2014, 3:45am aaron_do wrote on Apr 3^rd, 2014, 8:02pm: Why don't you post the pictures directly? A lot of people won't want to download files (which could contain anything...). Aaron
Back to top	P1dB_70G.png IP Logged

mustadem Junior Member Offline Posts: 11	Re: P1dB simulation Reply #8 - Apr 4^th, 2014, 3:48am aaron_do wrote on Apr 3^rd, 2014, 8:02pm: Why don't you post the pictures directly? A lot of people won't want to download files (which could contain anything...). Aaron Hi Aaron, I zipped the files bcos I think it's impossible to send multiple files on this forum. I hope this four pictures provide useful information. Thanks
Back to top	P1dB_10G.png IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: P1dB simulation
Reply #9 - Apr 6^th, 2014, 4:22am

Hi,

so for your 70 G plot, I guess you're saying the gain is too low? You should be able to match small signal simulations with PSS at low power. As Sheldon mentioned, you should probably be using the harmonic balance engine for PSS but that probably isn't the reason for the gain mismatch. How did you simulate your small signal gain?

For the 10 G plot, are you plotting the correct harmonic (i.e. the 70G harmonic).

Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

mustadem

Junior Member

Offline

Posts: 11

Re: P1dB simulation
Reply #10 - Apr 6^th, 2014, 5:34am

aaron_do wrote on Apr 6^th, 2014, 4:22am:

Hi,

so for your 70 G plot, I guess you're saying the gain is too low? You should be able to match small signal simulations with PSS at low power. As Sheldon mentioned, you should probably be using the harmonic balance engine for PSS but that probably isn't the reason for the gain mismatch. How did you simulate your small signal gain?

For the 10 G plot, are you plotting the correct harmonic (i.e. the 70G harmonic).

Aaron

Yes I meant that the gain is too low (it doesn't correspond to the small signal gain) and there is even no input intersect of the two lines.
For the small signal gain, I used s parameter simulation for it. the s21 is the small signal gain.
For the 10GHz plot, I was actually plotting the 10GHz harmonic. So another question is that regardless of the beat frequency, I should always plot the curve for the operating freq i.e 70GHz in this case

Back to top

IP Logged

mustadem

Junior Member

Offline

Posts: 11

Re: P1dB simulation
Reply #11 - Apr 6^th, 2014, 5:49am

sheldon wrote on Apr 3^rd, 2014, 8:01pm:

Mustadem,

� The information you are referring to is quite old. Since you are
working at quite high frequencies, using lumped element components
is difficult. So it may be better to use the harmonic balance engine.
Recent versions of Spectre RF releases include information,
application notes, on simulating LNA with Harmonic Balance �
analysis in the documentation. In addition, there have been some
enhancements to accelerate the simulation of P1dB in the most
recent release that you might find useful.

� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � Best Regards,

� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � Sheldon

Thanks Sheldon,

By just changing the engine to harmonic balance, the results are similar. I am still trying to see if I can get the recent release of Spectre RF that discusses pss simulation and with specific reference to LNA

Thanks

Back to top

IP Logged

RFICDUDE Community Fellow Offline Posts: 323	Re: P1dB simulation Reply #12 - Apr 6^th, 2014, 2:15pm You probably should, if you haven't already, verify the linear gain is "as desired" before trying to debug the P1dB simulation. Have you run linear simulations (AC or S-parameter) to verify the small signal circuit gain is greater than 0dB? If the linear gain is this low it may explain the results you are getting.
Back to top	IP Logged

mustadem

Junior Member

Offline

Posts: 11

Re: P1dB simulation
Reply #13 - Apr 7^th, 2014, 2:41am

RFICDUDE wrote on Apr 6^th, 2014, 2:15pm:

You probably should, if you haven't already, verify the linear gain is "as desired" before trying to debug the P1dB simulation.

Have you run linear simulations (AC or S-parameter) to verify the small signal circuit gain is greater than 0dB?

If the linear gain is this low it may explain the results you are getting.

I have run AC and s-parameter simulations. The gain from the AC simulation is slightly lower than that of the s-parameter but they are both higher than 6dB of gain from 60 to 80GHz

Back to top

IP Logged

aaron_do Senior Fellow Offline Posts: 1398	Re: P1dB simulation Reply #14 - Apr 7^th, 2014, 6:08pm You should be able to get a match between PSS and SP assuming you are comparing Pout/Pav in the PSS analysis with S21...
Back to top	there is no energy in matter other than that received from the environment - Nikola Tesla IP Logged