+ Post New Thread
Results 1 to 7 of 7
  1. #1
    Member level 1
    Points: 1,679, Level: 9
    Achievements:
    7 years registered

    Join Date
    Mar 2009
    Posts
    35
    Helped
    1 / 1
    Points
    1,679
    Level
    9

    Digital Oscilloscope to view high frequency signal

    Hi,

    I read your reply in one of the posts and thought you might be able to help me. I am using Keysight 2.5GHz oscilloscope to view a 121.5MHz signal across a 15pF capacitor. The data sheet says it should be a square wave but I am getting a sine wave of 121.5MHz. The probe bandwidth is 500Mhz with 10MG and 9.5pF capacitance.

    Could you advise what am I doing wrong?

    Kind regards

    •   Alt16th March 2017, 04:44

      advertising

        
       

  2. #2
    Newbie level 3
    Points: 21, Level: 1

    Join Date
    Apr 2017
    Posts
    3
    Helped
    0 / 0
    Points
    21
    Level
    1

    Re: Digital Oscilloscope to view high frequency signal

    Maybe the problem is the bandwidth. When you have a square wave you have energy in all odd harmonics of its fundamental frequency. So your bandwidth is not large enough and only pass the firsts two harmonics.



  3. #3
    Advanced Member level 5
    Points: 34,085, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,600
    Helped
    1653 / 1653
    Points
    34,085
    Level
    45
    Blog Entries
    7

    Re: Digital Oscilloscope to view high frequency signal

    It is not your bad test method. I assume you have a std 10:1 probe with long ground wire and probe clip. the probe input is say 3~10pF //10M but the cable capacitance at roughly 60pF /m for ultra low C coax instead of typ 50 ohm coax at 100pF/m is a problem with probe ground inductance.

    Next your std 10cm ground clip at 1mm is 100nH (102) and 150nH if 0.5mm thick or 70nH if 5cm long. Now calculate the resonant frequency of LCR and you get a sine wave.

    So your probe must be SMA from a pad from x Ohm source to 50ohm 1:1 coax direct to scope or try spring wire 5mm long wrapped around probe barrel and remove clip to expose <5mm tip to get maybe 300 MHz bandwidth . coax quality and proper probe use is critical. If source is 50 Ohm CMoS and you make a 10:1 50 ohm pad, the scope must also be terminated to 50 ohms for textbook waveforms with compensation cap added as required.

    I don't know your probe but it needs to be 50ohm terminated to prevent LC resonance , then scale accordingly.
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975



    •   Alt10th April 2017, 17:14

      advertising

        
       

  4. #4
    Full Member level 4
    Points: 2,774, Level: 12
    Achievements:
    7 years registered

    Join Date
    Nov 2009
    Location
    United States
    Posts
    237
    Helped
    68 / 68
    Points
    2,774
    Level
    12

    Re: Digital Oscilloscope to view high frequency signal

    Indeed, your probe's impedance and reactance is so similar to the impedance of the signal load you are measuring on (15pF cap) that it is significantly affecting the waveform.
    With these sorts of measurements, you may have to resort to an active, FET input probe for a more accurate image. The lead length and grounding tips mentioned earlier are good to follow and may help you to get closer to not affecting your measurement with your measuring device, but these sorts of sensitive measurements are often done with the active FET probe to minimize measurement error



  5. #5
    Advanced Member level 5
    Points: 34,085, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,600
    Helped
    1653 / 1653
    Points
    34,085
    Level
    45
    Blog Entries
    7

    Re: Digital Oscilloscope to view high frequency signal

    its not that simple with a long ground wire and it's actually more complex than this and much higher impedance than the signal imedance, unless he has 10cm bread board wires
    \
    Last edited by SunnySkyguy; 11th April 2017 at 03:43.
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975



    •   Alt11th April 2017, 03:36

      advertising

        
       

  6. #6
    Super Moderator
    Points: 237,282, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    41,064
    Helped
    12540 / 12540
    Points
    237,282
    Level
    100

    Re: Digital Oscilloscope to view high frequency signal

    The real circuit is even more complex because the coaxial cable is intentionally lossy, using a resistance wire as conductor.

    The 100 nH ground clip + around 10 pF probe input capacitance circuit is however sufficient to explain why harmonics of the 120 MHz signal are attenuated. The probe manual fine print would probably tell that 500 MHz bandwidth can be only achieved with a very short (e.g. coaxial or "bayonet" tip) ground connection.

    500 MHz or GHz bandwidth with "flying wire" connection is only possible with an active probe.



  7. #7
    Advanced Member level 5
    Points: 34,085, Level: 45
    Achievements:
    7 years registered Created Blog entry Created Album pictures
    SunnySkyguy's Avatar
    Join Date
    Sep 2007
    Location
    Richmond Hill, ON, Canada
    Posts
    6,600
    Helped
    1653 / 1653
    Points
    34,085
    Level
    45
    Blog Entries
    7

    Re: Digital Oscilloscope to view high frequency signal

    They do have a passive 500Mhz probe option but they don't tell you how to use and the equiv. circuit has compensation more complex than I have shown.

    Obvious to me is to use tip and barrel in a test socket or twin wire tips <<1cm
    A good design question lists your overall requirements™ The best question deserves a better answer. ™
    ... Tony Stewart EE since 1975



--[[ ]]--