8

I programmed my FPGA to create a simple 1 MHz rectangle function. But when I display the resulting function on my oscilloscope it shows some oscillation after the edges.

At first I thought this might be the the Fourier components, but this doesn't seem to fit(eg. it's not symmetric).

What creates these oscillations? I have a few guesses, but no idea which, if any, is right:

  1. some unintentional LC-circut
  2. a badly configured PID inside the FPGA
  3. crosstalk

enter image description here

Stein
  • 191
  • 2
  • bad probe? correct impedance matching –  Aug 27 '15 at 23:08
  • 2
    Ho ! I still have an oscilloscope like this one. It has a vacuum tube input stage. Pretty reliable for 40 years old gear. – TEMLIB Aug 28 '15 at 00:48

4 Answers4

11

How long is the ground lead on your scope probe?

  1. your scope probe should be in the X10 position and be properly compensated.
  2. Remove the long ground lead that came with the scope probe and also remove the probe grabber hook from the front of the probe.
  3. Grab some #22 or #24 bare wire and wrap 2 or 3 turns around the exposed metal ring at the front of the probe. Tightly twist the ends of the wire loop together so as to make a stiff pigtail.
  4. Solder that ground pigtail to a ground point that is as close as possible to the signal that you want to measure. The ground pigtail should be as short as possible.
  5. Touch the probe tip to the signal that you want to observe.

Now look at your scope waveform. Chances are very good that it will look much better, without the ringing that you are seeing on the fast transistions.

In other words, this looks to be a measurement problem rather than something wrong with your circuit.

Dwayne Reid
  • 23,390
  • 2
  • 35
  • 67
7

This is normally called ringing and is common. If there is just single pulse then it is referred to as overshoot.

As you surmise it is caused by unintentional LC elements in the wiring being excited by the fast edges of the signal. Even one or two inches of wire may create enough inductance to cause this effect.

Often it is exaggerated by bad grounding of the scope probe. You should use as short a ground lead on the probe as possible. Often I can check whether it is caused by the probe just by gripping the scope probe and ground lead tightly with my hand - if the displayed trace changes then the grounding is not good enough. Your scope probe may have come with alternate tips that include ways of connecting the probe to the circuitry with a shorter ground.

This application note from Jim Williams has some good ideas and explanations of probing problems: Jim Williams AN47

As the other posters have commented the oscilloscope you are using is extremely low in bandwidth by modern standards. In general I find that 100MHz bandwidth is a good bandwidth for use with circuits with clocks up to 10-20MHz. In my professional work 1GHz is typical now for general purpose use.

Also although the signal frequency may only be 1MHz in your example the signals will transition in 1 ns or faster that can cause very high frequency effects.

Kevin White
  • 32,097
  • 1
  • 47
  • 74
0

your scope says 10Mhz right on the front; that's the 3dB bandwidth. So the 1Mhz fundamental gets through without much distortion, but only up through about the 9th or 11th harmonic... so the square wave will appear to have some ringing at the edges. this is normal. a higher bandwitdh scope would show the 1Mhz signal with crisper edges, but there will always be some ringing. it's more noticeable as you approach the bandwidth limit of the instrument.

MarkU
  • 14,413
  • 1
  • 34
  • 53
  • So you are saying the high frequency Fourier components are damped and this creates the oscillation? Why wouldn't it be symmetric? – Stein Aug 27 '15 at 23:22
  • For a 3.3v digital clock signal, it looks pretty clean - there is some ringing at the edges but not enough to make the signal ambiguous. I'm not convinced that there is a real problem here. – MarkU Aug 27 '15 at 23:23
  • I don't see the asymmetry? – MarkU Aug 27 '15 at 23:24
  • @marku I probably won't be a problem, but I'm interested in what creates it – Stein Aug 27 '15 at 23:26
  • 1
    The oscillation is only after the edges, but not before. Unlike the Fourier components: https://upload.wikimedia.org/wikipedia/commons/thumb/f/fa/Fourier_synthesis.svg/748px-Fourier_synthesis.svg.png – Stein Aug 27 '15 at 23:27
  • Good point, I was only looking at vertical symmetry. I think Dwayne's answer is better, the ringing is most likely caused by non-ideal scope probe ground lead connection. – MarkU Aug 28 '15 at 00:20
0

I believe the major cause of the edge ringing is the scope probe. With regards to the symmetry of the signal, I see it very symmetrical.

Guill
  • 2,430
  • 10
  • 6