+ Post New Thread
Results 1 to 8 of 8
  1. #1
    Member level 5
    Points: 530, Level: 4

    Join Date
    Jun 2014
    Posts
    89
    Helped
    0 / 0
    Points
    530
    Level
    4

    How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    I am using Agilent Advanced Design System 2011.10 for my circuit design. I am getting the "no convergence error" for my designed circuit to obtain HB2 Two tone swept power. I have used "OPTIONS" setting also. Even though I used "OPTIONS" settings also, am getting the " no convergence error".

    How to overcome this convergence problem for my designed circuit, latter I am able to obtain IIP3 measurement.

    Herewith I attached the Schematic of my circuit, HB2 tone swept power simulation generated convergence error file, and solving convergence using "OPTIONS" setting file.

    •   AltAdvertisment

        
       

  2. #2
    Super Moderator
    Points: 50,445, Level: 54

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    12,407
    Helped
    2464 / 2464
    Points
    50,445
    Level
    54

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    You have many capacitors and inductors, but few resistors.

    I have seen a simulator get bogged down when two or more capacitors are connected in series.

    Try putting a low-ohm resistor in series with every capacitor and inductor. This creates a time constant which might give the simulator an easier time.

    Also try different time step values. What frequency does your circuit operate at? See if it helps to set the time step so there are between 50 and 1,000 steps within this period.


    1 members found this post helpful.

    •   AltAdvertisment

        
       

  3. #3
    Member level 5
    Points: 530, Level: 4

    Join Date
    Jun 2014
    Posts
    89
    Helped
    0 / 0
    Points
    530
    Level
    4

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    Quote Originally Posted by BradtheRad View Post
    Try putting a low-ohm resistor in series with every capacitor and inductor. This creates a time constant which might give the simulator an easier time.
    Also try different time step values. What frequency does your circuit operate at? See if it helps to set the time step so there are between 50 and 1,000 steps within this period.
    My circuit operating frequency is 3.5 GHz. Can you advice, what time steps I have to use for avoiding convergence and what low -ohm resistors I have to use for avoiding convergence.



    •   AltAdvertisment

        
       

  4. #4
    Super Moderator
    Points: 50,445, Level: 54

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    12,407
    Helped
    2464 / 2464
    Points
    50,445
    Level
    54

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    The simulator tries to achieve convergence. I think its report is telling you that it gets a 'no convergence' error.

    The report mentions L17 as a cause of error. Try putting a 1 ohm resistor inline with it.

    Or there is a chance the simulator is confused by two parallel inductors. Try disconnecting either L17 or L13.

    The report states that convergence was able to be achieved when it added resistors in the circuit. The values given are suitable for a start. (I was thinking 1 ohm because it is easy to type, and it is not likely to interfere with your circuit's operation.)

    If the above steps do not succeed, then try simplifying your circuit. Remove some sections. If the simulation runs okay, then you can put the sections back again.


    1 members found this post helpful.

  5. #5
    Member level 5
    Points: 530, Level: 4

    Join Date
    Jun 2014
    Posts
    89
    Helped
    0 / 0
    Points
    530
    Level
    4

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    Quote Originally Posted by BradtheRad View Post
    The report mentions L17 as a cause of error. Try putting a 1 ohm resistor inline with it.

    Or there is a chance the simulator is confused by two parallel inductors. Try disconnecting either L17 or L13.

    The report states that convergence was able to be achieved when it added resistors in the circuit. The values given are suitable for a start. (I was thinking 1 ohm because it is easy to type, and it is not likely to interfere with your circuit's operation.)

    If the above steps do not succeed, then try simplifying your circuit. Remove some sections. If the simulation runs okay, then you can put the sections back again.
    As you said, I followed it. Now it is not showing no convergence, but it is not showing anything in the graph, blank graph it is showing.
    In the graph, it is said us increase the maximum order. Even increased the order also, I am getting blank IIP3 graph.

    Here with I attached Schematic and PDF of the generated IIP3 graph and hpeesofim file.



    •   AltAdvertisment

        
       

  6. #6
    Super Moderator
    Points: 50,445, Level: 54

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    12,407
    Helped
    2464 / 2464
    Points
    50,445
    Level
    54

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    Sorry, I'm not familiar with Agilent. About all I can suggest is that you make sure you set the timeframe of the graph so it matches the timeframe of the simulation run.

    Also try the example circuits, or demo graph circuits, that show you how to create graphs.

    As for suggesting a timestep, multiply your frequency times 100.

    Thus 3.5 GHz x 100, equals 350 GHz.

    Take the reciprocal, and you get 2.9 pica-Seconds for the timestep.

    Try 3, or 2, or 1 pSec.


    1 members found this post helpful.

  7. #7
    Member level 5
    Points: 530, Level: 4

    Join Date
    Jun 2014
    Posts
    89
    Helped
    0 / 0
    Points
    530
    Level
    4

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    Quote Originally Posted by BradtheRad View Post
    As for suggesting a timestep, multiply your frequency times 100.

    Thus 3.5 GHz x 100, equals 350 GHz.

    Take the reciprocal, and you get 2.9 pica-Seconds for the timestep.

    Try 3, or 2, or 1 pSec.
    As you suggested timestep. Where I have to set timestep 3, or 2, or 1pSec. In schematic, we have only sweep plan and course plan, there we can set start power and end power.
    I have got got, where I have to set the time step of 3, or 2, or 1 pSec.



  8. #8
    Advanced Member level 5
    Points: 28,273, Level: 41
    BigBoss's Avatar
    Join Date
    Nov 2001
    Location
    Turkey
    Posts
    4,078
    Helped
    1233 / 1233
    Points
    28,273
    Level
    41

    Re: How to overcome CONVERGENCE problem for our CKT using Agilent ADS2011.10

    Quote Originally Posted by hari_preetha View Post
    I am using Agilent Advanced Design System 2011.10 for my circuit design. I am getting the "no convergence error" for my designed circuit to obtain HB2 Two tone swept power. I have used "OPTIONS" setting also. Even though I used "OPTIONS" settings also, am getting the " no convergence error".

    How to overcome this convergence problem for my designed circuit, latter I am able to obtain IIP3 measurement.

    Herewith I attached the Schematic of my circuit, HB2 tone swept power simulation generated convergence error file, and solving convergence using "OPTIONS" setting file.
    How did you get this schematic ??? I have never seen such a nonsense schematic in my life ....
    Your problem is not a convergence problem, your circuit is definitely wrong.Biasings are wrong, inputs. outputs are wrong, there are XXX nH at the sources, resistive matching...

    There are tens of error..

    The drain of second FET is connected to VDD and it's connected to the next stage at the same time.So its output is short circuited...
    There are XXX nH inductors ( If I see correctly ).. What are your trying to do ?? Oscillator ???
    Have your ever checked the bias voltages and currents ?? You got many things to do...
    Before jumping into such complex designs, read something about LNAs and follow up some tutorials,presentations, etc. about circuit simulation and design techniques..


    1 members found this post helpful.

--[[ ]]--