Designer’s Guide Community

ggnhdk

New Member

Offline

Posts: 5

Output impedance mixer
Jan 31^st, 2008, 1:56am

Hello I'm designing a mixer and I'd like to get the output impedance on the IF access. Under ADS I know how to do this, it's used on several design guides, they use a small signal current source and they get the voltage to determine the impedance. But under cadence I don't know how to realize the same thing with spectre.

I tried to set a PAC analysis with PSS and set the current source with 1mA AC magnitude and get the voltage at the node to get the impedance with the ratio of the two values but it doesn't work. If someone has an idea it would help me a lot.

Thanks

Back to top

IP Logged

Ken Kundert

Global Moderator

Offline

Posts: 2386
Silicon Valley

Re: Output impedance mixer
Reply #1 - Jan 31^st, 2008, 9:19pm

Perhaps you can tell us where you put the ac source, what sideband you were looking at, how you set up the analysis, and why you believe it does not work.

Why did you set the AC to 1mA, that is an odd choice. Normally people use 1. Oh, and did you mag or pacmag. mag is only active in ac analyses, pacmag is active in the PAC analysis.

-Ken

Back to top

IP Logged

ggnhdk

New Member

Offline

Posts: 5

Re: Output impedance mixer
Reply #2 - Feb 1^st, 2008, 8:16am

For exemple, RF=39GHz, OL=38GHz, IF=1 GHz.
I want the coefficient reflexion on RF and IF access. The AC to 1 mA it 's the choice used in ADS in all design guides, they make the ratio Voltage/current to get the impedance.

I think I can get these parameters with a PSS+PSP analysis by setting a frequency range, the different ports with the right harmonic for each port and I get the curves S11 S22 for each power OL specified in swepp parameter in PSS.
The problem is that I have the port RF @39 and the IF @1, but it traces S11 AND S22 @39GHz +- the frequency sweep range specified in relative sweeptype. I have to launch another simulation and specify the same harmonic for the port RF and the port IF to get the right scale on X axis ie the IF +- freq range.

DO you have an idea for setting the parameters of the simulation to get the Sxx at the right scale?

EDIT: I have just changed the frequency sweep range for the PSP simulation "from 0 to 40G" to get the curves for all the frequency domain and if I compare the results to the results of Advanced Design System, the good values are for the range around the RF frequency.
I don't understand why the S22 which concerns the IF access seems to be the right result at 39GHz, not at 1 GHz...

thanks for your help ken.

Back to top

IP Logged

Ken Kundert Global Moderator Offline Posts: 2386 Silicon Valley	Re: Output impedance mixer Reply #3 - Feb 1^st, 2008, 9:02am Normally people use pacmag=1 to avoid having to compute V/I. mag and pacmag are ignored in PSP. The settings for PSP are a little confusing, and I have not used it for a while. I recommend you use "spectre -help psp" to get you going. Focus on setting the harmonics correctly for your sweeptype. -Ken
Back to top	IP Logged

ACWWong Community Fellow Offline Posts: 539 Oxford, UK	Re: Output impedance mixer Reply #4 - Feb 1^st, 2008, 9:13am Also be sure to check the "freqaxis" setting to understand what the x axis is. Often (in times of confusion) I have found it very helpful to do absin for freqaxis with an absolute sweep so I can find my feet.
Back to top	IP Logged

ggnhdk New Member Offline Posts: 5	Re: Output impedance mixer Reply #5 - Feb 3^rd, 2008, 8:07am Ok thanks for your help you gave me the solution ! In fact, I choosed the freqaxis and set IN and I get the good results at the good frequency. I din't understand the absin and the difference with the others.
Back to top	IP Logged