Measuring 1 ns pulse accurately w/ 500 MHz Scope

Discussion in 'General Electronics Chat' started by newbie217, Nov 3, 2009.

  1. newbie217

    Thread Starter Active Member

    Apr 12, 2009
    Hi all,

    I am trying to accurately measure a short, 1 ns square wave (0 to 5 V) that is coming in as an input signal.

    The scope I am using is rated at 500 MHz bandwidth. This translates to a period of 2 ns. For a 50% duty cycle, the ON pulse would be 1 ns, exactly what I am trying to measure.

    I know the topic of measurement and signal integrity is vast and complicated, but for my intended purposes (ignoring stray capacitance / board parasitics for the moment, should my scope be able to accurately recreate the input signal? I've read literature where they say a scope's bandwidth is equal to the -3 dB low-pass filter roll-off point, or 30% error point? Does that mean that for my 2 ns square wave, at 50% duty cycle, I would have 30% error in my measurement? Is this why scope manufactures recommend the bandwidth of the scope be 3-5x greater than the signal of interest?

    Thanks for the help!
  2. someonesdad

    Senior Member

    Jul 7, 2009
    The basic problem you're facing is locating the requisite edges of the pulse. You're using a low pass filter (the scope's input channel) that's pretty much only going to see a reduced amplitude version of a sine wave at the fundamental frequency. You can do some simple numerical experimentation (using numpy in python is great for such things) to get a feel for what problems you'll have. Don't forget to factor in the effects of the probe(s) you use, such as rise times and cable lengths.

    Personally, I've always used the rule of thumb I made up years ago that I shouldn't mess with square waves of frequencies more than 1/10th of the scope's bandwidth without expecting some potential measurement problems.

    Another problem is: supposing you do get a measurement? How do you check it and verify its accuracy?

    It's possible that the tool you want for this is a time interval counter. I know HP made some of these in decades past and that some of the folks I used to work with used them for accurate time interval measurements. You might also want to talk to some experimental nuclear physicists, as they have needs of dealing with fast pulses in particle counting experiments. Optical physicists and engineers also deal with measuring short pulses.
  3. russ_hensel

    Well-Known Member

    Jan 11, 2009
    Is it a repettive wave. Saying it has a 50% duty cycle implies it is. If so rounding of the wave does not matter, you can measure either the period or the frequency.