+ Post New Thread
Results 1 to 10 of 10
  1. #1
    Newbie level 5
    Points: 88, Level: 1

    Join Date
    Jan 2017
    Posts
    8
    Helped
    0 / 0
    Points
    88
    Level
    1

    ft of HBT transistor and Class A operation

    Dear All,


    I have seen some papers on power amplifier design. They are biasing the transistor near their peak ft and saying that transistor is biased near class A operation region. I did not understand conceptually that if we bias the transistor near their peak ft then, we are in class A region. How is the the different biasing classes related to ft ? what if we want to design class B then how to decide on biasing point vs ft.

    The title of papers are
    1) A 110–134-GHz SiGe Amplifier With Peak Output Power of 100–120 mW.

    2)A Wideband High-Efficiency 79–97 GHz SiGe Linear Power Amplifier with > 90 mW Output.

    Any explanation is most welcome.

  2. #2
    Super Moderator
    Points: 45,198, Level: 51

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    11,243
    Helped
    2222 / 2222
    Points
    45,198
    Level
    51

    Re: ft of HBT transistor and Class A operation

    It is possible to increase bias to class B so the bias is continually ON. Then each transistor is operating more like class A. However this wastes power.

    There is something interesting that you can do with class A. As you raise the bias in class A, you increase DC gain, and also any AC signal riding the DC voltage. It's a means to obtain maximum power (as long as you avoid clipping.)
    Class A offers low distortion. CLass A is easy to work with because you only have the single bias to care about.

    Class B is more complicated. It has two transistors which alternately turn on & off (in simple terms). You need to care about matching gain of two transistors, and applying two biases, adjusting two biases, etc.


    1 members found this post helpful.

    •   Alt12th February 2017, 01:39

      advertising

        
       

  3. #3
    Advanced Member level 5
    Points: 23,493, Level: 37

    Join Date
    Nov 2001
    Location
    Earth
    Posts
    3,246
    Helped
    992 / 992
    Points
    23,493
    Level
    37

    Re: ft of HBT transistor and Class A operation

    There is a probably a misunderstanding.There is no relationship between class of an amplifier and fT.Class of an amplifier is related to its bias point.( essentially conduction current )
    fT is almost function of DC current of the transistor and it rises by current increasing then reaches a peak after that drops smoothly.Therefore there is optimum current for max. fT
    I guess they meant to say.


    1 members found this post helpful.

    •   Alt12th February 2017, 10:48

      advertising

        
       

  4. #4
    Newbie level 5
    Points: 88, Level: 1

    Join Date
    Jan 2017
    Posts
    8
    Helped
    0 / 0
    Points
    88
    Level
    1

    Re: ft of HBT transistor and Class A operation

    Quote Originally Posted by BigBoss View Post
    There is a probably a misunderstanding.There is no relationship between class of an amplifier and fT.Class of an amplifier is related to its bias point.( essentially conduction current )
    fT is almost function of DC current of the transistor and it rises by current increasing then reaches a peak after that drops smoothly.Therefore there is optimum current for max. fT
    I guess they meant to say.

    I asked the author of the paper and he mentioned that if you are biasing the transistor at high biasing current then you are automatically in class A operation region otherwise you can be in class AB, B or C.

    In 0.13um SiGe process the IHP has not specified the maximum current of transistor. They said for optimum results from the transistor you can bias the transistor near peak ft current density. I want to design Class AB or B power amplifier using ft versus biasing current plots. I do not know how to design the class B power amplifier without knowing maximum current of the transistor.



    •   Alt12th February 2017, 12:15

      advertising

        
       

  5. #5
    Advanced Member level 2
    Points: 4,062, Level: 15
    Achievements:
    7 years registered

    Join Date
    May 2009
    Posts
    547
    Helped
    130 / 130
    Points
    4,062
    Level
    15

    Re: ft of HBT transistor and Class A operation

    Actually, IHP gives the current density at maximum ft in their datasheets, and consider your dc collector current at "class b", you'll suffer from a very low ft. It's very useful to simulate ft. vs ic. By the way, in the PDKs of IHP, there are a lot of diagrams with plots of ft,fmax vs. ic


    1 members found this post helpful.

  6. #6
    Newbie level 5
    Points: 88, Level: 1

    Join Date
    Jan 2017
    Posts
    8
    Helped
    0 / 0
    Points
    88
    Level
    1

    Re: ft of HBT transistor and Class A operation

    I tried to design the PA in class B operation region. It seems at high frequency the transistor has almost no gain in Class B region. That is why people are biasing the transistor near peak ft.



    •   Alt15th February 2017, 21:44

      advertising

        
       

  7. #7
    Advanced Member level 5
    Points: 23,493, Level: 37

    Join Date
    Nov 2001
    Location
    Earth
    Posts
    3,246
    Helped
    992 / 992
    Points
    23,493
    Level
    37

    Re: ft of HBT transistor and Class A operation

    Quote Originally Posted by searchforknowledge View Post
    I tried to design the PA in class B operation region. It seems at high frequency the transistor has almost no gain in Class B region. That is why people are biasing the transistor near peak ft.
    Class-B amplifiers are not generally designed as single ended.Instead, practical Power Amplifiers are designed as Class-AB Push-Pull configuration.Ft is found to be as function of DC current and you should plot Ft vs. Idc in a simulator.If you use Cadence ADE Spectre, you may use calculator utility to plot this curve.



  8. #8
    Advanced Member level 5
    Points: 23,493, Level: 37

    Join Date
    Nov 2001
    Location
    Earth
    Posts
    3,246
    Helped
    992 / 992
    Points
    23,493
    Level
    37

    Re: ft of HBT transistor and Class A operation

    I prepared a test-bench for you in Cadence ADE by using Spectre simulator.If you abel to access to this simulator, you can do the same thing to find the FTmax.
    As you see, FT has a max. value where the Collector/Base current is optimal.You can use this function to find the max. FT.

    This is FT. vs Ibase ( of course Icoll) curve of BFP420 of Infineon and it has exactly FTmax ( 25GHz) as mentioned in its datasheet

    Code:
    unityGainFreq(mag((IF("/Q0/C") / IF("/Q0/B"))))]
    Click image for larger version. 

Name:	FT_Test.png 
Views:	15 
Size:	17.0 KB 
ID:	136213Click image for larger version. 

Name:	FT_vs_Ibase.png 
Views:	11 
Size:	68.0 KB 
ID:	136214


    1 members found this post helpful.

  9. #9
    Advanced Member level 2
    Points: 4,062, Level: 15
    Achievements:
    7 years registered

    Join Date
    May 2009
    Posts
    547
    Helped
    130 / 130
    Points
    4,062
    Level
    15

    Re: ft of HBT transistor and Class A operation

    Quote Originally Posted by searchforknowledge View Post
    I tried to design the PA in class B operation region. It seems at high frequency the transistor has almost no gain in Class B region. That is why people are biasing the transistor near peak ft.
    Exactly, RF PAs are mostly biased in weak class AB in a push-pull configuration for linear PAs



  10. #10
    Newbie level 5
    Points: 88, Level: 1

    Join Date
    Jan 2017
    Posts
    8
    Helped
    0 / 0
    Points
    88
    Level
    1

    Re: ft of HBT transistor and Class A operation

    I am sorry for so late reply. I was checking the threads again and found your simulation in Cadence. I did the simulation for ft/fmax in ADS. However, I got ft/fmax greater than the one specified in the IHP design kit. There is about 70-80 GHz difference. I think at frequency the model becomes invalid so we can not use the formulation at high frequency. We have to extract the ft/fmax using some low frequency. By the way, I have the design kit with ft/fmax 300/450 GHz.



--[[ ]]--