Designer’s Guide Community

AnalogAroma

Community Member

Offline

Posts: 32
Germany

Problem in simulating a Diff. Switching PA in SpectreRF
Apr 23^rd, 2009, 5:50pm

Hi Everybody,

I am facing some problems in simulating a differential switching power amplifier where the tail current is modulated by the baseband signal. I mean, the baseband signal will generate a ref current which will linearly control the tail current which will be up-converted to the RF frequency using a LO signal. Similar to a single balanced mixer. The attached pic further clarfies the PA.

Typical simulation techiniques in SpectreRF are suitable for a PA succeeding a mixer where RF input is coming from the mixer.

But, in my case, LO osc input does not have any baseband envelope .
And the low freq baseband input cannot use RF port symbol !!!

So, typical way of calculating 1dB compression point is not suitable.

Is there any way to overcome this problem in SpectreRF ?
Any trick ?

Back to top

Switching_PA_sim_issue_in_SpectreRF.jpg

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Problem in simulating a Diff. �Switching PA in SpectreRF
Reply #1 - Apr 23^rd, 2009, 7:25pm

I can't find out any Design Issue in your questions. �Your questions are simple measurement issue of RF block.

I can't understand what you are confusing.
Your "Switching PA???" is no more than TX-UpMixer or AM-Modulator which have two inputs, LO and BB.
So there is no problem in treating your case by two-tones(LO, BB) or one-tone(LO) steady state analysis.

I don't think your circuit is Switching PA. http://www.designers-guide.org/Forum/YaBB.pl?num=1214811603

AnalogAroma wrote on Apr 23^rd, 2009, 5:50pm:

Typical simulation techiniques in SpectreRF are suitable for a PA succeeding a mixer where RF input is coming from the mixer.
But, in my case, �LO osc input �does not have any baseband envelope.
And the low freq baseband input cannot use RF port symbol !!!

What do you mean ?

AnalogAroma wrote on Apr 23^rd, 2009, 5:50pm:

So, typical way of calculating �1dB compression �point is not suitable.

How do you define 1dB compression �point ?
"LO_Input_Level v.s. Output_Level with BB_Input=Constant Level(DC or Sine)"
or
"BB_Input_Level v.s. Output_Level with LO_Input=Constant Level(Sine)" ?

How do you define output power for modulation output ?
"Averaged Power" or "PEP(Peak Envelope Power)" ?

Usually I simulate simply non-modulation case where BB is DC.
Here I use swept BB of DC source with one-tone(LO) steady state analysis.

If I use sinusoidal signal for BB, I invoke�two-tones(LO, BB) steady state analysis.
Here you have to choose which output power you should evaluate. "Averaged Power" or "PEP(Peak Envelope Power)".
This selection is dependent on your needs or interests.

Attached figure is Instantaneous Envelope Power of ASK signal.
http://www.designers-guide.org/Forum/YaBB.pl?num=1209812189

AnalogAroma wrote on Apr 23^rd, 2009, 5:50pm:

Is there any way to overcome this problem in SpectreRF ?

Is there any problem in treating your case by Cadence Spectre ?, although I don't use Cadence Spectre.

Back to top

« Last Edit: Apr 24^th, 2009, 4:51am 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

Ken Kundert Global Moderator Offline Posts: 2384 Silicon Valley	Re: Problem in simulating a Diff. �Switching PA in SpectreRF Reply #2 - Apr 23^rd, 2009, 11:53pm SpectreRF does not require the use of ports as sources. It works just as well with voltage sources and current sources. -Ken
Back to top	IP Logged

pancho_hideboo Senior Fellow Offline Posts: 1424 Real Homeless	Re: Problem in simulating a Diff. �Switching PA in SpectreRF Reply #3 - Apr 24^th, 2009, 4:30am Ken Kundert wrote on Apr 23^rd, 2009, 11:53pm: SpectreRF does not require the use of ports as sources. It works just as well with voltage sources and current sources. Any vendor's RF simulator except for very old one such as "SuperCOMPACT" does not require the use of ports as sources.
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

Andrew Beckett Senior Fellow Offline Life, don't talk to me about Life... Posts: 1742 Bracknell, UK	Re: Problem in simulating a Diff. �Switching PA in SpectreRF Reply #4 - Apr 24^th, 2009, 6:39am Correct, but it's a common misperception amongst users of RF simulators, perhaps because examples are often using ports, and so there is sometimes a mistaken assumption that you must use them.
Back to top	IP Logged

AnalogAroma Community Member Offline Posts: 32 Germany	Re: Problem in simulating a Diff. Switching PA in SpectreRF Reply #5 - May 6^th, 2009, 10:55am Tape out date is over. So far, I could not respond to all the feedback. Within several days I will try to describe the problems encountered in the recent tape-out of power amplifier in greater detail.
Back to top	IP Logged

AnalogAroma Community Member Offline Posts: 32 Germany	Re: Problem in simulating a Diff. Switching PA in SpectreRF Reply #6 - May 11^th, 2009, 10:38am Although I had attached a CMOS based PA image, in reality SiGe HBT based PA was designed.
Back to top	IP Logged

pancho_hideboo Senior Fellow Offline Posts: 1424 Real Homeless	Re: Problem in simulating a Diff. �Switching PA in SpectreRF Reply #7 - May 12^th, 2009, 3:54am AnalogAroma wrote on May 11^th, 2009, 10:38am: Although I had attached a CMOS based PA image, in reality SiGe HBT based PA was designed. What do you want to say ? Whatever process you use for designing your PA, simulation methods are same.
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