Help getting the right transistor model for a colpitt's oscillator,866.5MHz

Discussion in 'The Projects Forum' started by Prosper J.M, May 29, 2011.

  1. Prosper J.M

    Thread Starter New Member

    May 28, 2011
    5
    0
    I have designed a colpitt's oscillator using a bjt npn transistor at 866.5MHz fundamental frequency.When I used a generic transistor model on CircuitMaker 2000,it simulated perfectly with the right frequency,866.5MHz,and self starting...I didnt want to use FETs.My problem is that when I replace the generic model of the transistor to a practical model,example MMBR901 which is supposed to be silicon,Vmax 15 v,Imax 30mA,2.5 Ghz hifreq...it simulates but with a lower frequency!

    I googled and did some searching for atleast 2 days before now and am thinking it has something to do with the James Earl effect but not sure.

    Anyways,I'd appreciate if someone could tell me whats wrong and why the change as well help me with how to correct it via mathematics and not just trial and error tuning.I have to solder this circuit in the end,so it has to be right.
    Thanks
     
  2. Prosper J.M

    Thread Starter New Member

    May 28, 2011
    5
    0
    Screenshots
     
    • osc.jpg
      osc.jpg
      File size:
      35.8 KB
      Views:
      30
    • osc.png
      osc.png
      File size:
      15.9 KB
      Views:
      17
  3. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Does your generic transistor have junction capacitances and a finite fT? When you changed to the MMBR901, did you include these parameters in your calculations?
    When you solder it, parasitic capacitances will also be present. Are you including those in your simulation?

    What is the James Earl effect?
     
  4. Prosper J.M

    Thread Starter New Member

    May 28, 2011
    5
    0
    @Ron..thanks for responding.James M. Early effect...decrease in the base-collector depletion region,hence increase of collector current with increasing collector voltage...havent studied it much,heres a link: http://en.m.wikipedia.org/wiki/Early_effect

    And by "generic",I mean ideal,so the generic transistor exhibits ideal characteristics : no or negligible junction capacitances,zero forward transit time if thats what you meant by fT.Everything is ideal,and I think you pinned it,that may actually be the problem : internal capacitances leading to non-ideal transit time!But what gets me is the specs say its fit for up to 2.5ghz,so all that must already be compensated for...right?!

    If you were making such an oscillator and needed that much a frequency reading,what type of transistor would you recommend that wont force me to get unconventional in the design?
     
  5. Adjuster

    Well-Known Member

    Dec 26, 2010
    2,147
    300
    The resonant circuit tuning your oscillator terminates in a very small capacitance at the collector end (C2, 0.2pF). This is probably comparable with, or even less than the output capacitance of even a good RF transistor. There is no way that the transistor can "compensate" for this, it is an effect you need to allow for. In a practical layout, other parasitic capacitances could also be of this order, so it may be unrealistic to try to set the resonating capacitance this low.

    In fact, the frequency you are trying to obtain is not far from the limit of what is possible with a "lumped" design, and you will have to accept significant effects from parasitic capacitances (and inductances). The physical layout of circuits like this is quite difficult.
     
  6. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I know what the Early effect is. I don't know what the James Earl effect is (unless it has something to do with James Earl Jones).:D
    The junction capacitances affect the resonant frequency of the tank circuit. I believe they are independent of fT, the transit frequency, which is the frequency where β=1. I suppose fT will also reduce the oscillation frequency, if it is near fT.
    I'm no expert on high frequency oscillators. I just know that you have to include the transistor model in your calculations if you want to predict frequency-dependent phenomena.
    Many high frequency transistors are characterized by their s-parameters, which may be more useful in designing an oscillator, but I know too little about the subject to be helpful.
     
  7. Prosper J.M

    Thread Starter New Member

    May 28, 2011
    5
    0
    Yeah I think I'll have to include the transistor capacitances.The frequency is very high to try and neglect them,probably use a hybrid model,PI model or something.Only problem is the simulator am using doesnt give these values so I can include them in my calculations.
     
Loading...