+ Post New Thread
Results 1 to 10 of 10
  1. #1
    Full Member level 4
    Points: 1,318, Level: 8

    Join Date
    Jan 2018
    Posts
    219
    Helped
    1 / 1
    Points
    1,318
    Level
    8

    Loadpull for Differential PA

    There are few posts on this but I just clarify it properly for my problem.

    https://www.edaboard.com/showthread....ht=ideal+balun
    https://www.edaboard.com/showthread....ht=ideal+balun
    https://www.edaboard.com/showthread....ht=ideal+balun

    Click image for larger version. 

Name:	setup_diff_PA.png 
Views:	25 
Size:	25.7 KB 
ID:	155703

    First of all I would like to know whether the set up is correct in the figure.

    Here is my concern. If I have single ended configuration, I will have Zopt from loadpull as, say, Ropt+j*Xopt. And the impedance looking into the output of this single ended is ,let, Rl-j*Xl.

    Now, I have a Differential configuration. I connected the outputs of the two single-ended configurations to ideal_balun. So the impedance looking into the ideal_balun is 2*( Rl-j*Xl) (in this case, output side)

    should the value I get from the loadpull, ideally, be 2*(Ropt+j*Xopt)?

  2. #2
    Advanced Member level 5
    Points: 17,262, Level: 31
    pancho_hideboo's Avatar
    Join Date
    Oct 2006
    Location
    Real Homeless
    Posts
    2,661
    Helped
    714 / 714
    Points
    17,262
    Level
    31

    Re: Loadpull for Differential PA

    Simply your thoughts are all wrong.

    Differential mode can not exist in your circuit.
    Surely learn basic linear circuits theory before EDA Tool Play.



    •   AltAdvertisement

        
       

  3. #3
    Advanced Member level 5
    Points: 30,550, Level: 42
    BigBoss's Avatar
    Join Date
    Nov 2001
    Location
    Turkey
    Posts
    4,481
    Helped
    1346 / 1346
    Points
    30,550
    Level
    42

    Re: Loadpull for Differential PA

    -Ideal Balun does not do Impedance Transformation.It acts simply as Diff-to-Single Ended so the Impedances ( Input/Output) are same..
    -This structure has high Output Impedance ( I don't know the details but looks like it) so it's not convenient for PA.A PA should have more or less an Output Impedance closer to Load to match easily..( I presume the PA will work @+20dBm or more )



    •   AltAdvertisement

        
       

  4. #4
    Super Moderator
    Points: 261,125, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    45,612
    Helped
    13876 / 13876
    Points
    261,125
    Level
    100

    Re: Loadpull for Differential PA

    1. Your circuit is flawed and must be corrected, you have shorted the output.
    2. Not sure if the corrected circuit is exactly equivalent to "two single-ended configurations". If it is, your calculation would be correct.
    3. I don't see an advantage of using a balun in simulation. Why not simply use a differential load?



    •   AltAdvertisement

        
       

  5. #5
    Advanced Member level 5
    Points: 17,262, Level: 31
    pancho_hideboo's Avatar
    Join Date
    Oct 2006
    Location
    Real Homeless
    Posts
    2,661
    Helped
    714 / 714
    Points
    17,262
    Level
    31

    Re: Loadpull for Differential PA

    Quote Originally Posted by FvM View Post
    3. I don't see an advantage of using a balun in simulation.
    Why not simply use a differential load?
    For load value control in Cadence Spectre, we have to use balun for differential load.



  6. #6
    Full Member level 4
    Points: 1,449, Level: 8

    Join Date
    Jan 2019
    Location
    Ireland
    Posts
    206
    Helped
    86 / 86
    Points
    1,449
    Level
    8

    Re: Loadpull for Differential PA

    I inferred that you are doing mmwave from your previous posts. If that is the case, that current source will eat away unnecessary voltage headroom. I read somewhere that people don't use the current source beyond 5GHz (I have forgotten the paper). Just ground your sources. You will get a pseudo-differential pair but larger voltage headroom.

    And as someone already pointed out, your outputs are shorted. Also use 'dcblock' and 'dcfeed' else you will can run into convergence issues while doing HB-PSS (atleast I did in 45nm).

    But you don't even need all of that. Connect your DC sources directly to the center tap of the transformers. True for both input and output.

    What you get from the loadpull is the impedance that the output port wants needs to have. Its is as simple as that.


    1 members found this post helpful.

    •   AltAdvertisement

        
       

  7. #7
    Full Member level 4
    Points: 1,318, Level: 8

    Join Date
    Jan 2018
    Posts
    219
    Helped
    1 / 1
    Points
    1,318
    Level
    8

    Re: Loadpull for Differential PA

    Quote Originally Posted by vivekroy View Post
    And as someone already pointed out, your outputs are shorted. Also use 'dcblock' and 'dcfeed' else you will can run into convergence issues while doing HB-PSS (atleast I did in 45nm).
    How are you saying that output is shorted, is it because the C terminal of Ideal_balun connected to ground?. Also, I already connected 'dcblock' and 'dcfeed', where else should they be connected?

    Quote Originally Posted by vivekroy View Post
    Connect your DC sources directly to the center tap of the transformers.
    I don't think there is center tap transforment in analogLib. Should I make something like this?
    Click image for larger version. 

Name:	centertap.PNG 
Views:	8 
Size:	9.4 KB 
ID:	156007



  8. #8
    Advanced Member level 5
    Points: 17,262, Level: 31
    pancho_hideboo's Avatar
    Join Date
    Oct 2006
    Location
    Real Homeless
    Posts
    2,661
    Helped
    714 / 714
    Points
    17,262
    Level
    31

    Re: Loadpull for Differential PA

    Quote Originally Posted by circuitslave View Post
    How are you saying that output is shorted
    Can you understand differential circuit ?
    https://www.edaboard.com/showthread.php?386764#2

    Quote Originally Posted by circuitslave View Post
    I don't think there is center tap transforment in analogLib.
    Should I make something like this ?
    No. There is no coupling between L.
    Can you understand actual transformer ?

    Ideal transformer is not mutual L.

    Surely learn very basic things before EDA Tool Play.



  9. #9
    Super Moderator
    Points: 261,125, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    45,612
    Helped
    13876 / 13876
    Points
    261,125
    Level
    100

    Re: Loadpull for Differential PA

    How are you saying that output is shorted
    A wire between the differential output nodes is usually understood as short. Review the post #1 schematic.



  10. #10
    Full Member level 4
    Points: 1,449, Level: 8

    Join Date
    Jan 2019
    Location
    Ireland
    Posts
    206
    Helped
    86 / 86
    Points
    1,449
    Level
    8

    Re: Loadpull for Differential PA

    Quote Originally Posted by circuitking View Post
    I don't think there is center tap transforment in analogLib. Should I make something like this?
    Click image for larger version. 

Name:	centertap.PNG 
Views:	8 
Size:	9.4 KB 
ID:	156007
    Look at the symbol of a balun very carefully and you will see what we mean.



--[[ ]]--