PLL to double frequency

Discussion in 'Wireless & RF Design' started by Art, May 29, 2014.

  1. Art

    Thread Starter Distinguished Member

    Sep 10, 2007
    785
    61
    Hi Guys,
    This isn't strictly radio, but I was looking at one of the relatively cheap rubidium clocks on eBay
    that are supposed to maintain an ultra precise frequency reference for clocking and synchronising test equipment.

    The more expensive of them has a programmable output 2-20MHz AC wave,
    and I want a higher frequency wave that is also AC.

    My question is about whether you can set the rubidium clock to output 10MHz,
    use an external PLL to double the frequency, and output either AC or DC?
    Or does the PLL output AC, and if you want DC you have to buffer it?

    Also I'm guessing that doubling a frequency will have a negative effect on accuracy,
    especially if accuracy is quoted as a value like + or - pulses per time period.
    Cheers, Art.
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
  3. Art

    Thread Starter Distinguished Member

    Sep 10, 2007
    785
    61
    Thanks, it looks like the circuit I want is simply called a "Frequency Doubler".
    Some of them use logic chips to produce a digital signal, I'll just have to watch I use an AC one,
    and buffer it later if I want a digital signal.
     
  4. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Hello,

    You can make a simple digital frequency doubler using some gates:

    [​IMG]

    Hope this helps.

    Bertus
     
  5. Art

    Thread Starter Distinguished Member

    Sep 10, 2007
    785
    61
    Using a digital circuit would ruin a sinusoidal wave form that I want to maintain,
    and turn it into pulse width modulated DC.
    I want the AC input maintained, and the additional option of converting to DC would
    only mean buffering it through a transistor or similar.
     
  6. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Old style is to use an RF amplifier stage,with the input tuned to f1,& the output circuit tuned to 2f1.
    You can easily double,or triple the input frequency in one stage.

    The resonant output circuit restores the sine wave at the new frequency after the original frequency signal is distorted in the doubler/tripler,to create harmonics.

    You could restore the sinewave in a similar way if you use another method which gives you a square wave.

    A pulse waveform is not really DC--------it has both an AC & a DC component,& much of the time it behaves like normal AC.
     
  7. Art

    Thread Starter Distinguished Member

    Sep 10, 2007
    785
    61
    Face slap. I think that should have somehow been obvious.
    The first harmonic at the doubled frequency.

    Is the reason the DC signal like an AC signal because the transistors in the
    chips doing the switching have their own rise and falls slopes, or another reason?



     
  8. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Sort of!-------The very act of switching between different levels introduces an AC component.

    If you were to look at a series of rectangular pulses with an Oscilloscope which has an FFT function,you will see a large spike at the Pulse Repetition Frequency (PRF),with sidebands around it corresponding to PWM (if any is applied),then various harmonics of the PRF,up to the maximum frequency components,which will be determined by the rise & fall times of the switching devices.

    Pure DC would show nothing.

    Many people have the habit of thinking of a rectangular wavetrain as DC,maybe from the old term "pulsating DC" used to describe rectified AC in power supplies,but even this waveform has a substantial AC component.

    "DC" is a term best reserved for pure DC like out of a battery,or filtered DC power supply unit.
     
  9. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Although it seems counter-intuitive,the harmonic at twice the fundamental is the second harmonic.

    If there is a "first harmonic" it must be the fundamental,but the term is not used to refer to that frequency.

    It is consistent with other mathematical usage,where if you multiply
    "a" by itself once,the result is "a to the second power".
     
  10. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
  11. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Hi Art, sometimes it's a good idea to solves problems from the back. Ie; what do you need the "20MHz AC sine" for? And do you need AC and DC (re your DC buffer question)?

    And how accurate does the freq need to be?

    Most things that make use of a precision freq reference expect a 10MHz input, which is standard for the rubidium modules output anyway. It would be rare for something to need a precision 20MHz input.
     
  12. Art

    Thread Starter Distinguished Member

    Sep 10, 2007
    785
    61
    Hi again Guys,
    Thanks for all the replies, and I will look into the diodes as well.

    Hi RB,
    I have a weird frequency counter that wants a 14 MHz reference,
    and ideally I want to clock a transceiver directly (which will require doubling a frequency),
    as well calibrate instruments, and clock a micro with it to run the real time clock display.
    I have the serial software to talk to it, but ideally it's a job for a micro to make the thing a stand alone device.


     
    THE_RB likes this.
  13. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    The paper on diode doublers that Bertus suggested is very good. I suspect for your application, that you would have to follow it with a filter to attenuate the fundamental and then an amplifier to compensate for the 10db loss in power.
    If you wanted to buy a diode doubler you could look at Minicircuits.com. They have surface mount models that are less than $10. (I like Minicircuits, they are the Wal-mart of RF components.)
     
Loading...