0
\$\begingroup\$

I am sending the clock signal of a DE0-Nano FPGA to one of its GPIO pins. As I increase the frequency, the signal becomes more and more distorted. Is this an issue with the GPIO pins having small buffers? My scope has a bandwidth of 200MHz and my probes are also rated for 200MHz. What might be the reason for this?

Ultimately I want to modulate data on top of this clock signal.

1) Is there a way I can output a clean clock signal at a high frequency like 100MHz?

2) Can I modulate a clean square wave of data on top of this clock signal?

10kHz

enter image description here

100kHz

enter image description here

1MHz

enter image description here

10MHz

enter image description here

100MHz

enter image description here

\$\endgroup\$
  • 3
    \$\begingroup\$ Quick test: run all the signals at 20 ns/div and look at the rising edge. I bet they all look the same. \$\endgroup\$ – John Birckhead Jul 10 '17 at 21:05
  • \$\begingroup\$ What type of modulation? \$\endgroup\$ – John Birckhead Jul 10 '17 at 21:18
  • \$\begingroup\$ Your question is not how to output the clock, but how to correctly capture it. Please look up the Nyquist sampling theorem - you will learn that to see a clean 100 MHz signal with likely 1-ns edges, you would need at least a 500-MHz or 1 GHz scope, with at least 2 Gsps sampling rate. \$\endgroup\$ – Ale..chenski Jul 10 '17 at 21:34
  • \$\begingroup\$ My scope samples at 1GS/s and is rated to have a bandwidth of 200MHz \$\endgroup\$ – user2562609 Jul 10 '17 at 21:36
  • \$\begingroup\$ For the modulation I want to output either a low or high voltage for every rising clock edge. \$\endgroup\$ – user2562609 Jul 10 '17 at 21:37
1
\$\begingroup\$

Fist, a 200 MHz oscilloscope and a 200 MHz probe will only give you a system bandwidth of 141 MHz (ish).

We talk about calculating "system bandwidth" on our blog here: https://community.keysight.com/community/keysight-blogs/oscilloscopes/blog/2016/09/01/what-is-oscilloscope-system-bandwidth-and-how-do-i-find-the-bandwidth-of-the-scope-probe

Basically, the bandwidths compound.

Second, a bandwidth of 141 MHz means you can a 141 MHz sine wave with 3dB or less of attenuation. So, for a 100 MHz clock, it's not going to look like a square wave.

This app note actually looks at a 100 MHz clock example on page 7 and talks about how to evaluate your oscilloscope's bandwidth for your app.

http://literature.cdn.keysight.com/litweb/pdf/5989-5733EN.pdf

The ringing you're seeing is probably some inductive or capacitive effects of your probing system.

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.