Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

RF pulse detection with Power Detectors

  1. Jan 10, 2019 #1
    How can I use power detectors for RF pulse detection? RF pulse has a symmetric Gaussian-like shape.
    Is the output of power detectors continuous line or point?
     
    Last edited by a moderator: Jan 10, 2019
  2. jcsd
  3. Jan 10, 2019 #2

    berkeman

    User Avatar

    Staff: Mentor

    Welcome to the PF. :smile:

    There are a number of different power detector types. Which type are you asking about? What is the repetition rate of the pulses (the duty cycle)? What power levels are involved? What RF frequencies are involved?

    And is this really homework? If not, I can move your thread to the technical forums.
     
    Last edited: Jan 10, 2019
  4. Jan 10, 2019 #3
    I'm studying how to catch pulse width without oscilloscope. Unknown pulse is from Q-switched laser which has a symmetric Gaussian-like shape.
    If a pulse is sinusoidal wave form, the problem becomes quite simple. I can find exact frequency because Fourier transform of sinusoidal wave form has finite number of frequency.
    220188-4c434af793251cf579c175a0c06e941e.png

    But Fourier transform of Gaussian is Gaussian which has infinite number of frequency. That means each wave of frequency has different power. Is this right?
    220189-4cc5c7a7e42d8bbeaa047027090d5d7c.png

    This is the graph of power detector what I consider.
    220190-ff135dc0bb58918174afa72d52ee1194.png
    I can't come up with which value is output from power detector. I am confused that the output of power detector is just one value or a continuous liner value.

    Unfortunately, I don't know the type of power detector. Also I will use just one pulse so the duty cycle is not taken into consideration.

    And.. if possible would you kindly explain me relation between RF pulse detection and power detector or advise me a web site?

    P.S)When I posted this thread, I didn't knew about specific categories. I should have posted on the technical forum.
     

    Attached Files:

  5. Jan 10, 2019 #4

    berkeman

    User Avatar

    Staff: Mentor

    If you want to detect the pulse width, you would use an envelope detector, not a power detector. Let me know the frequencies of interest and I can suggest an instrument.
    (Thread moved from the schoolwork forums to the technical forums)
     
  6. Jan 12, 2019 at 5:19 PM #5

    tech99

    User Avatar
    Gold Member

    If you use a power detector, the output voltage is proportional to input power irrespective of wave shape.
    A typical power detector for lower frequencies is a square law device, like a diode. At optical frequencies a photo diode performs the same function. If you apply a sine voltage wave, the output is its square. If we average this over a period of time we obtain the mean square of the input voltage, and so it represents the mean square value of the input voltage waveform, which is the power.

    If you have a pulsed waveform, the average power can be found by measuring the output voltage with an averaging circuit, such as a resistor followed by a shunt capacitor, giving a long time constant. Then you measure the peak power, now using a small resistor with the capacitor, so that RF cycles are smoothed but the envelope fully charges the capacitor, so now we see the peak power in the envelope.
    Using average and peak power you can find the mark/space ratio, or more accurately, duty cycle, as we do not know the pulse shape.
     
  7. Jan 14, 2019 at 1:31 AM #6
    I have a Gaussian like shaped pulse waveform. It has 0.3ns pulse width, and it means 1.45Ghz half-band width in my calculation. If signal has 13dBm then what is output voltage from power detector? Is this enough information to calculate output or not? It's ok if I could get approximate value.
    upload_2019-1-14_16-30-7.png
    upload_2019-1-14_16-30-42.png
     

    Attached Files:

  8. Jan 14, 2019 at 12:07 PM #7

    tech99

    User Avatar
    Gold Member

    I don't like the fact that you have a DC offset on the power meter calibration. Best to re-draw the graph with 2.1 volts as the zero and reading upwards.
    If 13dBm is the average power, then a power meter with an averaging CR circuit (or just a slow analogue meter) would read around 0.9 volt depending on carrier frequency.
    If 13dBm is the peak power, then a power meter fitted with a peak reading capacitor will similarly read 0.9V. In this case I don't know what average reading you will get.
     
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?