VCO design Help!!

Discussion in 'General Electronics Chat' started by Moyango, Dec 27, 2012.

  1. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    Hi all:

    I have a problem need to be solved urgently. If you could help. I will be very appreciated. Thanks.

    Problem is a design voltage control oscillator. These are the requirements:

    - transistors (bjt and mosfet) and other passive elements like resistors, inductors, capacitors. are allowed.
    - Filters are forbidden.
    -wo=275Khz range is from 0.8w0 to 1.2wo.
    - no jitter and no phase noise requirements so no need any PLL or varactors
    - A design only with resistors and transistors quite posible! they say
    -higher frequency is quite welcome.

    this is a discrete time design. i find some sort of VCOs. I decided to design a ring oscillator. but it does not work; inverters are not inverting. all values keep constant instead of oscillating.

    [​IMG]
     
  2. WBahn

    Moderator

    Mar 31, 2012
    17,757
    4,800
    1) Your image wouldn't load. Please attach it to your post instead of linking to it on a third-party site.

    2) Why is it so urgent that someone solve your homework assignment for you?

    3) What efforts have you made to date to solve it yourself? Describe them, and the results, in some detail.

    4) This almost certainly should have been posted in the Homework Help section.
     
  3. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    http://i48.tinypic.com/30ldulk.png[​IMG]

    "urgent" was overstatement I accept but I did not mean "do my work". I need only some little help from experienced people. I already designed a circuit a ring oscillator and finally got some result but now problem is this: my waveform is not proper square... do you have any idea to fix the shape of output...

    http://i45.tinypic.com/15g8i6f.png
    [​IMG]
     
  4. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Try feeding the output into some buffer stage, that might help with the corners.
    Also, I am not quite sure how good the models of the cmos transistors are, so you might need to try it with a real chip.
    EDIT: didn´t notice that buffer, the picture is too large.
    Also, if my calculations are correct its oscillating at about 16MHz, which is all the cmos chips are good for, at this frequency it won´t get any better.
    Try to move the oscillation to that 275khz range, the corners will look much better.
     
    Moyango likes this.
  5. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    first of all, in this project cmos is forbidden I can use only transistors and R,L,C

    and it is actually interesting you say "275Khz" because in project requirement ,our asisstance put "275Khz or more" limitation...

    I am wonderin how you calculate the oscillation frequency, it is quite right.

    can you suggest any thing different than buffering for sharp square?
     
  6. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I understood your requirement "-wo=275Khz range is from 0.8w0 to 1.2wo." as 275khz center frequency, tunable between 0.8*275khz and 1.2*275khz.
    In your plot, you have time axis in microseconds, so you can easily calculate that 1/(4.92us-4.85us) is about 16MHz.
    You can try replacing the 4007 cmos transistors with some small signal mosfets or bjts and see what it does, the arrangement can stay the same, only for bjts you need to control the bottom transistors with a current source.

    As for the square shape, in reality no wave is perfectly square, so I wouldn´t worry about those corners. Anyway if a buffer doesn´t help with them, you need a better/faster buffer.
     
  7. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    thanks for help and I did not see what I already read there "wo=275Khz range is from 0.8w0 to 1.2wo" I thought you say "275Khz" for another reason. sorry about that.

    I see I have to use faster transistors but unfortunetly I do not know how I find desired transistors, because I do not know enough kind of transistors...
     
    Last edited: Dec 28, 2012
  8. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    If you connect the ring without the 100p caps, you will probably get an almost sinewave oscillator at the maximum frequency. Also you dont need better transistors, you need some transistors. Ask your tutor if you can use the transistors inside the 4007, since those are just a few transistors in a package, I wouldn´t call that a dedicated IC.
     
  9. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    CD4007 is CMOS.:confused:
     
  10. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    True, I don´t understand why OP starts with a chip he is not supposed to use, but on the other hand 4007 is to me a special case as it is more of a discrete transistor array than an IC as people normally think of it.
     
  11. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Yeah, maybe he means he can't use the VCO in CD4046.:D
     
    kubeek likes this.
  12. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    :D
    ok CD4007 is orginally cmos but in here I only use nmos and pmos model of it. But you are right I should replace of them with actual mosfets. Thanks for warning me. This is the first time I face cmos thing....
     
  13. Moyango

    Thread Starter New Member

    Dec 3, 2012
    7
    0
    Could you suggest any fast mosfet or bjt type??
     
  14. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    You can make a vco with an emitter-coupled oscillator, using BJTs. If you make it so that all transistors are non-saturating, 275kHz is no problem for general-purpose BJTs (e.g., 2N3904, 2N3906). If you select higher speed ones (e.g., MPSH10, MPSH81 for non-saturating transistors, 2N2369 and 2N5771 for saturating), you can easily get to tens of Megahertz if you know what you are doing.:D
    The down side of the emitter-coupled oscillator is it takes quite a few parts to make a good design, and also get logic-level outputs.
     
  15. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I simulated this circuit with 2N6804 and 2N6568 (arbitrary choice from the models available, it´s up to you to look them up), with 3x5pF and 10V supply it is running at ~16Mhz, so just get some small mosfets with low gate capacitance and try it out for real (or get the models). Adding a resistor to the supply changes the frequency nicely, so biasing the control mosfet (it may as well be just a single mosfet for the three pairs) should be easy to do.
     
Loading...