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

Oscilloscope trigger timing issue?

  1. Jul 9, 2012 #1
    I'm using an arbitrary waveform generator that is generating a 1 MHz ramp waveform. I'm triggering on a marker sent out by the waveform generator which is tied to the same memory location as the start of the waveform. So although there is supposed to be a small latency (1-3ns) between the marker and the waveform it should be roughly fixed in relation to the start of the ramp.

    I'm measuring this output on a 2 GS/s oscilloscope with a 200 MHz bandwidth, but when I trigger on the marker it seems that the spacing between the marker and the ramp can vary as much as 15ns, so when captured by the trigger the waveform continually shakes around on the order of that magnitude. The marker has a fall time of less than the min resolution of the oscilloscope which is 2.5ns. I have the input to the oscilloscope (1 M ohm input impedance) connected in parallel with a 50 ohm termination because the waveform generator is 50 ohm back terminated.

    What I'm unsure of is whether this shaking/poor timing is coming from the waveform generator itself or from the oscilloscope's inability to trigger properly. Since the AWG can produce much higher frequency waveforms I suspect it is the oscilloscope but I can't come up with a reason why. Any ideas or suggested reading would be appreciated.
  2. jcsd
  3. Jul 9, 2012 #2
    Trigger jitter on a decent 2 GS/s scope should be in the 10s of ps. Are you using BNC connections between waveform gen and scope? BNC 50 ohm feedthrough termination?
    Why not try outputting a square wave instead of a ramp so that you can see signal integrity and timing relationship of both signals.
  4. Jul 9, 2012 #3
    I'm using coaxial/sma connectors and the terminations are coaxial terminations. I'll check out a square wave tomorrow but if jumping due to the trigger is only supposed to be around 10 ps then it seems like it probably is the waveform generator. Which at least narrows it down. There are some software settings that I can adjust so I'll probably try to fiddle with those too.
  5. Jul 9, 2012 #4
    I want to confirm that all you are doing is connect two output directly from the generator to the scope and you already see the problem. One of the input to the scope trigger is the pulse associate to the start of the ramp. The second output is the ramp.

    If that is the case.
    1) Make sure you don't have a high frequency filter on the scope trigger.
    2) Change the polarity of the trigger and see whether you solve the problem. You did not mention which edge is the reference edge of the generator, the pulse width of the trigger pulse might not be constant and if you trigger on the wrong edge, you will have problem.
    3) You try adjusting the trigger threshold to see whether you can stablize the ramp?
    4) You sure the trigger output of the generator is specified to drive 50ohm? Try take off the 50ohm if all else fail.

    If your setup is directly from the generator to the scope alone, I cannot imagine any other problem.
  6. Jul 9, 2012 #5
    I've adjusted probably every setting on my oscilloscope and triggered on rising and falling edges and I have also tried it with the 50 ohm terminations removed. Also it is just a simple sma connection from the waveform generator to the oscilloscope with the terminations connected with a bnc T-connector at the oscilloscope. So I'll probably email the company to see what can be done.
  7. Jul 10, 2012 #6


    User Avatar
    Science Advisor

    I had this problem with a 200 MHz oscilloscope which was placed on top of a power supply.

    Because both devices had plastic boxes, the magnetic field from the power supply transformer was affecting the oscilloscope trace.

    Just moving them apart fixed the problem.
  8. Jul 12, 2012 #7
    If you have some extra long BNC, you might try placing a tee at the generator, run a short patch from that to the trigger (with 50ohms term at the scope), then run a longer patch from the tee to a scope input (again with a 50 ohm term at the scope). If the scope is stable, this should give you a rock-steady delay.
    Last edited: Jul 12, 2012
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook