---
+ Post New Thread
Results 1 to 12 of 12
  1. #1
    Newbie level 4
    Points: 1,535, Level: 8

    Join Date
    Aug 2004
    Posts
    6
    Helped
    0 / 0
    Points
    1,535
    Level
    8

    Differences between SiGe process and CMOS

    Dear all, I was used to be a CMOS designer, but now have to try SiGe process. May I have all the friends' advice about SiGe? For example, the guidelines, the important difference, or somethins about layout, etc. 3X a lot to my dear friends!

  2. #2
    Full Member level 4
    Points: 3,489, Level: 13
    hrkhari's Avatar
    Join Date
    Mar 2004
    Posts
    224
    Helped
    6 / 6
    Points
    3,489
    Level
    13

    Re: SiGe Process VS CMOS

    Hi

    SiGe process offers much more higher gain with higher transit frequency, and with less loading parasitics. If you are using the npn transistor the base is doped with Ge which provides higher current gain. The layout requires much more precision in symmetry since there is a current flow at the base, copared to CMOS where there is no current flow at the Gate

    Rgds


    1 members found this post helpful.

    •   Alt11th October 2004, 07:36

      advertising

        
       

  3. #3
    Advanced Member level 5
    Points: 27,958, Level: 40

    Join Date
    Oct 2001
    Posts
    4,014
    Helped
    1097 / 1097
    Points
    27,958
    Level
    40

    Re: SiGe Process VS CMOS



    1 members found this post helpful.

  4. #4
    Full Member level 1
    Points: 4,277, Level: 15
    lokeyh's Avatar
    Join Date
    Sep 2004
    Posts
    106
    Helped
    52 / 52
    Points
    4,277
    Level
    15

    Re: SiGe Process VS CMOS

    For CMOS, you can change within and length. But for SiGe, there is less flexibility as there are fixed sizes and the layout of the bipolar is almost standardized.



    •   Alt12th October 2004, 02:52

      advertising

        
       

  5. #5
    Full Member level 2
    Points: 2,317, Level: 11

    Join Date
    Dec 2003
    Posts
    138
    Helped
    7 / 7
    Points
    2,317
    Level
    11

    SiGe Process VS CMOS

    How can I acquire SiGe's design kit?



    •   Alt13th October 2004, 09:01

      advertising

        
       

  6. #6
    Full Member level 4
    Points: 2,775, Level: 12
    selvaraja's Avatar
    Join Date
    Jun 2004
    Posts
    191
    Helped
    6 / 6
    Points
    2,775
    Level
    12

    Re: SiGe Process VS CMOS

    from layout point of view you must carefull signal path ..avoid making turns for high speed signals ...you less worry about drc rule compare to CMOS ...cause most foundry provide fix layout for SiGe HBT (bipolar).



  7. #7
    Full Member level 3
    Points: 2,414, Level: 11

    Join Date
    Oct 2004
    Posts
    166
    Helped
    1 / 1
    Points
    2,414
    Level
    11

    Re: SiGe Process VS CMOS

    which technology is the most popular in industry now? And how about their application? ex. CMOS, BiCMOS, Bipolar, or GaAs. Anyone can give an summary?

    Quote Originally Posted by hrkhari
    Hi

    SiGe process offers much more higher gain with higher transit frequency, and with less loading parasitics. If you are using the npn transistor the base is doped with Ge which provides higher current gain. The layout requires much more precision in symmetry since there is a current flow at the base, copared to CMOS where there is no current flow at the Gate

    Rgds



  8. #8
    Member level 5
    Points: 1,918, Level: 10

    Join Date
    Oct 2004
    Location
    Dallas, Texas, USA
    Posts
    90
    Helped
    11 / 11
    Points
    1,918
    Level
    10

    SiGe Process VS CMOS

    Of course CMOS is the populat one because it is cheap



  9. #9
    Member level 5
    Points: 1,918, Level: 10

    Join Date
    Oct 2004
    Location
    Dallas, Texas, USA
    Posts
    90
    Helped
    11 / 11
    Points
    1,918
    Level
    10

    SiGe Process VS CMOS

    BEing able to work with SiGe process is very advantageous, notice 2 thing, first the transconductance
    gm=Ic/Vt it changes linearly with the bias current,
    and at least 20 times higher than CMOS counterpart for the same bias current.
    And small signal gain
    gm.ro=Ic/Vt*Vaf/Ic=Vaf/Vt as you can see this value is independent of the biasing condition (as oppose to CMOS), Vaf early voltage in a good process is arround 100 and Vt at room temperature is 26 mV. So the small signal gain that you can get from a BJT is on the order of 4000 (72dB). In practice of course this going to be smaller but it explain the potential of the device. The only bad thing is the low input resistance, but you have
    learn proper design structure to circumvent this (so it is doable). Once you fully undestand the capacbility of the device, especially in analog domain, you will start to appreciate it more more :)
    is arround



  10. #10
    Full Member level 1
    Points: 2,569, Level: 11

    Join Date
    Jul 2004
    Location
    21.402, -157.739
    Posts
    99
    Helped
    15 / 15
    Points
    2,569
    Level
    11

    Re: SiGe Process VS CMOS

    I had posted this paper, which may be pertinent to this discussion, in another thread.

    http://www.edaboard.com/ftopic92246.html

    "CMOS and SiGe Bipolar Circuits for High-Speed Applications"
    by Werner Simburger, et. al., GAAS IC Symposium 2003.

    Abstract—Recently, CMOS has been demonstrated to be a
    viable technology for very-high-bit-rate broadband and wireless
    communication systems up to 40Gb/s and 50 GHz. Advances in
    device scaling and doping-profile optimization have also resulted
    in SiGe bipolar transistors with impressive performance, including
    cut-off frequencies of more than 200 GHz. This paper presents
    advances in circuit design which fully exploit the high-speed
    potential of a 0.13 µm CMOS technology up to 50GHz and of a
    high-performance SiGe bipolar technology up to 110GHz operating
    frequency. The combination of advanced circuit techniques
    and a state-of-the-art fabrication-process technology results in
    continuing the upward shift of the frequency limits.



  11. #11
    Junior Member level 1
    Points: 1,610, Level: 9

    Join Date
    Oct 2004
    Posts
    17
    Helped
    0 / 0
    Points
    1,610
    Level
    9

    Re: SiGe Process VS CMOS

    By my personal experience, you'll really appreciate the high gm of
    Sige HBT, high current capability so that reduce the parasistic in the
    same current requirement. You may often need degenerated resistor
    to help the linearity, current mirror matching, lower the noise current.
    In LNA design, due to the higher gm, the noise should be smaller inherently.
    And, lower LO power requirement for switching function in Mixer. In PA design,
    SiGe provide higher power effency due to higher gain.



  12. #12
    Newbie level 6
    Points: 1,511, Level: 8

    Join Date
    Nov 2004
    Posts
    13
    Helped
    0 / 0
    Points
    1,511
    Level
    8

    Re: SiGe Process VS CMOS

    In terms of Ft:

    what you achieve with SiGe in a given technology node, you achieve it with 1 to 2 genetation later with RFCMOS.

    Remind that RFCMOS is not a pure CMOS technology. It requires several metal layers (6-7) for inductors and special process steps/ and design to reduce noises and increase reliability.

    Regarding flexibility on dimensioning and transistor optimization, Sige bipolars can be modelled with "Hicum" physical and scalable model:

    have a look on Xmod products: www.xmodtech.com

    Regards,



+ Post New Thread
Please login