Connect with us

Design problem in CMOS differential Amp

Discussion in 'General Electronics Discussion' started by Alex_Bam, Mar 8, 2021.

Scroll to continue with content
  1. Alex_Bam

    Alex_Bam

    20
    1
    Sep 28, 2020
    Hello,
    It's my first attempt to design CMOS Differential amplifier in 0.35um technology using a cadence tool. I search threads related to my problem in this forum but could not able to solve my problem. So I will explain my approach in deriving dimensions of MOSFET's used in it. Moreover, I have also attached my simulation picture for your reference.

    I get help from the following reference for calculations:
    https://medium.com/@vishnuanugrahith/designing-of-a-cmos-differential-amplifier-68d3d90baae4

    Design Parameters:
    GBW= 10M CL= 10pF, Vdd=5V
    UnCox=190u A/V^2; UpCox=55u A/V^2;
    Vtn=0.8V Vtp=-1V
    Slew Rate(SR) :10V/uSec
    Vcmax=4.5V ; Vcmin=1.5V (input common mode range)


    Step 1 _ Calculating Id:
    Id=SR*CL 10V/uSec * 10pF= 100uA

    Step 2 - Calculating (W/L) for M3 and M4:
    Vds1=Vmax-vthn=4.5-0.8=3.7V
    Vds3=Vdd-Vds1=5-3.7=1.3V
    Id3=Id/2=50uA
    (W/L)= 2*Id3/UpCox*vds3=1.075

    Step 3 - Calculating (W/L) for M1 and M2:
    gm1=GBW*2*pi*CL
    = 10M*2*3.14*10pF
    =6.27*10^-4
    Id1=Id/2
    (W/L)=gm1^2 / 2*Id1*unCox
    =20.75

    Step 4 _ Calculating (W/L) for M5 and M6:
    Vx=(Vmin-Vtn)-Vdsat1
    Vdsat1= (2*Id1/UnCox * (W/L)1 )^1/2= 0.16
    Vx=0.54
    Id5=Id
    (W/L) =2*Id5/UnCox * Vx^2= 3.6

    Finally, transistor dimensions are:
    (W/L)1,2 = 1.07u / 1u
    (W/L)3,4 = 20.75u / 1u
    (W/L)5,6 = 3.6u / 1u

    Differential input Signal:
    Freq= 1 Mhz; Amp= 500 mV; Sine DC level= 2V

    Differential gain:
    = gm2 (ron_n2 || ron_p4)
    = 418.5u( 71.73K || 17.04K )
    =5.7


    1_ I am not getting the output signal according to my calculated gain? I am making mistake in my calculations?
    2_ Could anyone comment on my design and mathematical calculations?
    3- I am always struggling to find the un*Cox value of MOSFET. after searching in different threads in this and other forums. One way is to find it from betaeff=un*Cox*W/L. But un*Cox changes with changing W/L. So I am always doubtful regarding un*Cox value. Can anybody know the accurate method to find un*Cox value of MOSFET especially using Cadence?

    Thanks.
     

    Attached Files:

    • dia.JPG
      dia.JPG
      File size:
      31 KB
      Views:
      6
    • sch.JPG
      sch.JPG
      File size:
      92.9 KB
      Views:
      5
    • re.JPG
      re.JPG
      File size:
      136.8 KB
      Views:
      4
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-