The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Mar 29th, 2024, 6:10am
Pages: 1 2 3 
Send Topic Print
time variant impedence in a passive mixer (Read 39917 times)
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #15 - Apr 24th, 2009, 10:21pm
 
Hi pancho_hideboo,


Quote:
If you use PSP or QPSP of Cadence Spectre, you have to use "analogLib/port" which results in load at observation nodes.

Such loads at observation nodes are not preferable especially for interstage impedances measurements.
Here even if you increase port resistance, it can not be a solution.


I was thinking that we can leave the output open circuit (or whatever the desired load is), and use a port at the RF input. Is this setup ok for measuring Z11?

Also, I was thinking that the LO leakage to either RF or IF ports can affect the operation and hence the Zin of the mixer. If you use a voltage source at the RF input the LO-RF leakage will be zero, so can this be an accurate measure of input impedance?

Quote:
Instead I created my own impedance probe by using vsources, isources and ideal-transformer.
This impedance probe is valid not only for Cadence Spectre but also for Agilent RFDE(ADSsim).
And this impedance probe is valid not only for conventional Small Signal AC Analysis but also for Small Signal AC Analysis subjected to Master Large Signal Analysis such as HB-SS of Agilent RFDE(ADSsim) and PSS/PAC of Cadence Spectre.


Would you mind sharing some details on how you set up this probe?


thanks,
Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedance in a passive mixer
Reply #16 - Apr 24th, 2009, 10:46pm
 
aaron_do wrote on Apr 24th, 2009, 10:21pm:
I was thinking that we can leave the output open circuit (or whatever the desired load is), and use a port at the RF input.
Is this setup ok for measuring Z11?
No, you are misunderstanding.

Resistance of "analogLib/port" surely acts as load in PSS of Cadence Spectre.
When you evaluate Z-parameters using PSP, you can't ignore this unpreferable load effects.
http://www.designers-guide.org/Forum/YaBB.pl?num=1231854969/5#5

This is because PSP is slave analysis of master analysis PSS.
You yourself also mentioned this effects in PSS in your another comments.

If you create switch using Verilog-A which operations are OFF for PSS and ON for PSP, you can use PSP without loading at nodes.

Quote:
// This module is valid only for Cadence Spectre
`include "discipline.h"
`include "constants.h"
module ac_switch(vp, vn);
inout vp, vn;
electrical vp, vn;

analog begin
  if( analysis("ac", "pac", "qpac",
           "xf", "pxf", "qpxf",
           "sp", "psp", "qpsp",
           "noise", "pnoise", "qpnoise") )
     V(vp, vn) <+ 0.0;
  else
     I(vp, vn) <+ 0.0 ; // V(vp, vn) <+ 1.0T * I(vp, vn);
end // analog

endmodule


On the other hand, you can use ports directly to evaluate Z and Y parameters in conventional SP-Analysis as far as ports don't affect DC bias.

aaron_do wrote on Apr 24th, 2009, 10:21pm:
Also, I was thinking that the LO leakage to either RF or IF ports can affect the operation and hence the Zin of the mixer.
Right.
http://www.designers-guide.org/Forum/YaBB.pl?num=1193135745/3#3

aaron_do wrote on Apr 24th, 2009, 10:21pm:
If you use a voltage source at the RF input the LO-RF leakage will be zero, so can this be an accurate measure of input impedance?
Impedance Probes don't shorten any terminals. See Fig. A.2 at page.32 of http://www.rdmiddlebrook.com/downloads/GFTManual.pdf
Impedance Probes are inserted like following.

[Source_Impedance or Previous_Stages]=[Diff_Imp_Probe]=[Mixer]=[Diff_Imp_Probe]=[Next_Stages]
                                                                                                         ||
                                                                                           [Diff_Imp_Probe]
                                                                                                         ||
                                                                                                    [VCO]

Here [Diff_Imp_Probe] don't shorten [Source_Impedance or Previous _Stages].
In this way, I can evaluate Zleft_diff, Zleft_comm, Zright_diff and Zright_comm at same time without any affect to original electrical connections of circuits.

Back to top
 
« Last Edit: Apr 25th, 2009, 9:04am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #17 - Apr 27th, 2009, 1:22am
 
I understand what you mean now. Thanks for sharing.

Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedence in a passive mixer
Reply #18 - Apr 30th, 2009, 10:47pm
 
This is an Interstage Differential Impedance Probe.
Back to top
 

test_Diff_Impedance_Probe.jpg
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #19 - May 4th, 2009, 8:51pm
 
Hi pancho_hideboo,


I'm having a little trouble implementing the impedance probe. I figured I could create the probe using the opposite of the ac_switch which you showed me. So the modified code is,

Quote:
`include "constants.vams"
`include "disciplines.vams"

module ac_switch(vp,vn);
inout vp;
electrical vp;
inout vn;
electrical vn;

analog begin
 if( analysis("ac", "pac", "qpac",
          "xf", "pxf", "qpxf",
          "sp", "psp", "qpsp",
          "noise", "pnoise", "qpnoise") )
    I(vp, vn) <+ 0.0;
 else
    V(vp, vn) <+ 0.0 ; // V(vp, vn) <+ 1.0T * I(vp, vn);
end // analog

endmodule


The arrangement for the impedance probe is shown in the attachement. The two isin components have an AC and PAC magnitude of 1 A. So I should be able to simply probe the differentical voltages RIGHT and LEFT in order to see the impedance.

I tried to test the probe in a simple AC circuit and there were errors with the AC simulation. If I run the DC analysis alone there is no problem. However, when running the AC analysis, there were errors in finding the DC solution for the AC analysis. The errors are shown below. (this was after a number of lines saying zero diagonal found in Jacobian at 'net020' and 'net 020', Reordering Jacobian, Extrapolation Failed).

Quote:
Error found by spectre during DC analysis, during AC analysis `ac'.
   Matrix is singular (detected at `net017').
   No DC solution found (no convergence).  

The values for those nodes that did not converge on the last Newton iteration are given below.  Also given is the manner in which the convergence criteria were not satisfied in the following form:
           Failed test: | Value | > RelTol*Ref + AbsTol

   V(VDD) = 0 V
       residue too large: | -1.50432 A | > 7.52162 mA + 1 pA
   I(I98.L0.Lsub3:1) = 0 A
       residue too large: | 650 mV | > 3.25 mV + 1 uV
   V(I98.L2.6) = 0 V
       residue too large: | 54.4945 mA | > 272.472 uA + 1 pA
   V(I98.L2.62) = 0 V
       residue too large: | 203.77 mA | > 1.01885 mA + 1 pA
   V(I98.NM0.D0.dsxnb:int_a) = 0 V
       residue too large: | 1.01 A | > 5.05 mA + 1 pA
   I(I98.NM0.M0.LBULK:1) = 0 A
       residue too large: | -500 mV | > 2.5 mV + 1 uV
   V(I98.R1.11) = 0 V
       residue too large: | 236.06 mA | > 1.1803 mA + 1 pA
   I(I117.M0.LDRAIN:1) = 0 A
       residue too large: | -650 mV | > 3.25 mV + 1 uV
   I(I118.M0.LDRAIN:1) = 0 A
       residue too large: | -650 mV | > 3.25 mV + 1 uV


So anyway I figure that one way to fix this is to get the AC analysis to read the DC solution from the DC analysis. Is there any way to do this? Also, is there a better way to solve this problem?


thanks,
Aaron
Back to top
 

Impedance_Probe.jpg

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedence in a passive mixer
Reply #20 - May 7th, 2009, 8:49am
 
aaron_do wrote on May 4th, 2009, 8:51pm:
I'm having a little trouble implementing the impedance probe.
I figured I could create the probe using the opposite of the ac_switch which you showed me.
     [snipped]
The arrangement for the impedance probe is shown in the attachement.
The two isin components have an AC and PAC magnitude of 1 A.
So I should be able to simply probe the differentical voltages RIGHT and LEFT in order to see the impedance.
You still don't understand interstage impedance probe.

In your schematic, your switch cut signal path.
Interstage impedance probe never cut any signal path.
Interstage impedance probe is basically same as measurement method of loop gain where any signal path is never cut.
Again see Fig. A.2 at page.32 of http://www.rdmiddlebrook.com/downloads/GFTManual.pdf

As I wrote in the following, my impedance probe is composed of combination of vsources, isources and ideal-transformer.
http://www.designers-guide.org/Forum/YaBB.pl?num=1240334878/10#10

I don't use any analysis specific switch in my Interstage impedance probe.

The reason why I showed analysis specific switch by Verilog-A in the following is to avoid unpreferable load effects in PSP and QPSP.
http://www.designers-guide.org/Forum/YaBB.pl?num=1240334878/16#16
Back to top
 
 
View Profile WWW Top+Secret Top+Secret   IP Logged
lorenzo
New Member
*
Offline



Posts: 2

Re: time variant impedence in a passive mixer
Reply #21 - Aug 25th, 2010, 2:36am
 
pancho_hideboo wrote on Apr 30th, 2009, 10:47pm:
This is an Interstage Differential Impedance Probe.

I have been reading those interesting post and currently I am interested into looking inside the box to see how you implemented the differential impedance probe. I look at the Middlebrook, but I am not sure the way you derive your proble, can you share it with us in more details ? Thanks
Back to top
 
 
View Profile   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #22 - Aug 27th, 2010, 1:19am
 
Hi,


pancho_hideboo said that he used an ideal transformer. I tried doing a impedance probe using the component xfmr from analogLib in cadence, but it seems that the xfmr component also passes DC. So I ended up using an ideal inductor to sense the AC current. I think that an ideal xfmr is in a way acting as a simulation dependent component, so perhaps the AC switch is another way to do it. I never ended up getting my AC switch to work though.

This is what I finally understood from pancho_hideboo's hints, but i'm not sure if its really correct. See attachment. I forgot to write the port names, but you get the idea. Also, I guess the simulator finds it easier if you use an ideal transformer since there's no frequency dependence (i think).

cheers,
Aaron
Back to top
 

ImpedanceProbe.jpg

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
lorenzo
New Member
*
Offline



Posts: 2

Re: time variant impedence in a passive mixer
Reply #23 - Aug 27th, 2010, 2:58am
 
Thanks,
I'll give it a try, how are you supposed to derive the impedance?
I guess you just measure voltage across the left pins for Zleft and right pins for Zright?
btw, I downloaded the Middlebrook paper...jeee it will take me a while to read it all Smiley but I guess it's worth.
Cheers
Back to top
 
 
View Profile   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #24 - Aug 27th, 2010, 5:29pm
 
Hi,


yeah just measure the voltage. Anyway hopefully I didn't misunderstand anything from what pancho_hideboo said, but that's the schematic I interpreted. Seems to work...


cheers,
Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedence in a passive mixer
Reply #25 - Aug 28th, 2010, 8:34am
 
lorenzo wrote on Aug 25th, 2010, 2:36am:
I look at the Middlebrook, but I am not sure the way you derive your proble,
can you share it with us in more details ?
Attached figure with mosaic is a Hint-1.

The following is a Hint-2.
http://www.designers-guide.org/Forum/YaBB.pl?num=1231344965/1#1

The following is a Hint-3.
But in my Impedance Probe both current and voltage excitations are implemented at same time.
http://www.designers-guide.org/Forum/Attachments/summarytb.pdf

If only current excitation is invoked, we can't measure very small impedance(e.g. short) correctly.
Vice versa, if only voltage excitation is invoked, we can't measure very high impedance(e.g. open) correctly.

The followings are general notes for you.

- Always describe correct tool's name and vendor's name which you use as tool or simulator.
- Don't do multiple posts which are same content.
- Don't request source code or behavioral model without any efforts.
- There are many simulators which have analyses called as PSS, PAC and Pnoise.
- Describe in detail with using correct terminologies.
- Warnigns are different from Errors.
- ADS is not name of simulator.
- There is no tool which name is Cadence.
- Don't use Direct Plot of Cadence ADE blindly without knowing definition.
- All gains in Direct Plot of Cadence ADE are "right", "true" and "practical" voltage gain.
- Don't mix up Simulation with Post Processing. They are completely different phase.
- MATLAB are different from Simulink.
- Learn measurements using actual instruments. Not "EDA Tool Play
Back to top
« Last Edit: Aug 29th, 2010, 4:20am by pancho_hideboo »  

Hint.png
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #26 - Aug 30th, 2010, 2:10am
 
Hi pancho_hideboo,


one thing i'm confused about (among other things). Won't your probe disturb the DC conditions of the circuit? Or are your current/voltage probes only measuring the AC signal? If so could you say how this is done?

Also, you mentioned you are implementing the voltage and current exitations at the same time. I noticed you have 8 outputs, but I can only see the need for four (common/differential mode Zleft and Zright). Are you providing two options for each one? i.e. one with current exitation and one with voltage exitation?


thanks,
Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedence in a passive mixer
Reply #27 - Aug 30th, 2010, 5:04am
 
aaron_do wrote on Aug 30th, 2010, 2:10am:
Won't your probe disturb the DC conditions of the circuit?
Not at all.

aaron_do wrote on Aug 30th, 2010, 2:10am:
Or are your current/voltage probes only measuring the AC signal?
Yes.

aaron_do wrote on Aug 30th, 2010, 2:10am:
If so could you say how this is done?
By using small signal sources for conventional small signal analysis
and slave small signal analysis for master large signal periodical steady state analysis.

aaron_do wrote on Aug 30th, 2010, 2:10am:
I noticed you have 8 outputs,
but I can only see the need for four (common/differential mode Zleft and Zright).
Not correct.
Eight quantities are needed for evaluation of Zleft_com, Zleft_diff, Zright_com and Zright_diff.

Zleft_com=Vleft_com/Ileft_com
Zleft_diff=Vleft_diff/Ileft_diff

Zright_com=Vright_com/Iright_com
Zright_diff=Vright_diff/Iright_diff

Back to top
 
« Last Edit: Aug 30th, 2010, 6:45am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: time variant impedence in a passive mixer
Reply #28 - Sep 7th, 2010, 12:45am
 
Hi pancho_hideboo,


it seems that your circuit may be less complicated than I originally thought. You earlier pointed to a reference to Middlebrook's general feedback theorem, and it led me to believe that your circuit is doing some kind of null-double-injection. However, after going over your circuit (the mosaic), it seems that you are simply doing the following:

1) split the signal into common-mode and differential-mode branch.

2) For each branch, inject a small signal current and voltage (both appear to be derived from a single small-signal source.

3) Zleft_common = Vleft_common/Ileft_common (the other impedances are similarly calcualted).

Am I still misunderstanding your probe?

Also, you mentioned earlier that only injecting a current would result in inaccuracy when measuring small impedances (and vice versa). Is this some kind of rounding error in the simulator's calculations?


thanks,
Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: time variant impedence in a passive mixer
Reply #29 - Sep 7th, 2010, 2:12am
 
aaron_do wrote on Sep 7th, 2010, 12:45am:
1) split the signal into common-mode and differential-mode branch.
Correct.

aaron_do wrote on Sep 7th, 2010, 12:45am:
2) For each branch, inject a small signal current and voltage
(both appear to be derived from a single small-signal source.
Correct.

aaron_do wrote on Sep 7th, 2010, 12:45am:
3) Zleft_common = Vleft_common/Ileft_common
(the other impedances are similarly calcualted).
Correct.

aaron_do wrote on Sep 7th, 2010, 12:45am:
Also, you mentioned earlier that only injecting a current would result in inaccuracy when measuring small impedances (and vice versa).
Simply consider a case where Zleft=0 and Zright!=0.
All injected current flow into Zleft.
So we can't measure Zright.
Back to top
 
 
View Profile WWW Top+Secret Top+Secret   IP Logged
Pages: 1 2 3 
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.