1
\$\begingroup\$

I have a Siglent SDS 1102CML+ oscilloscope. I would like to inspect SPI signals: SCK (probe 1) and MOSI (probe 2).

Why aren't these graphs rectangular like on logic analyzers?

Do I need to configure my oscilloscope in a different way?

I need to mention this model does not support protocol inspection like SPI, I2C, etc.

My probes and channels are configured in ×10 mode.

enter image description here

\$\endgroup\$
5
  • 2
    \$\begingroup\$ At each sample point in time, a logic analyzer can record and display one of two possible values: 0 or 1. So it'll always look square. An oscilloscope records and displays one of 2^8 = 256 or 2^12 = 4096 or whatever (depending on how many bits of resolution its ADC has) values. \$\endgroup\$
    – TypeIA
    Commented Sep 28, 2022 at 11:39
  • \$\begingroup\$ @TypeIA You should write your answers in the answer section so it can be vetted by the community, edited for clarity, indexed, accepted, and not push actual answers down. \$\endgroup\$
    – pipe
    Commented Sep 28, 2022 at 11:49
  • \$\begingroup\$ What is the clock signal speed? Please check again that probes are at 10x setting, and go through all waveform acquire and display settings. \$\endgroup\$
    – Justme
    Commented Sep 28, 2022 at 11:58
  • \$\begingroup\$ @pipe It's not a complete answer: one should read the existing answers below, which I don't wish to duplicate or plagiarize. It really was a comment: a particular way of looking at the situation that I find helpful, in addition to the nice answers below. \$\endgroup\$
    – TypeIA
    Commented Sep 28, 2022 at 12:05
  • \$\begingroup\$ Usually a 50 or 100 Mhz Oscilloscope displays MOSI or MISO and CLK signals without ripples. Most SPI devices look for a voltage above half amplitude of the SPI pulses . Some SPI peripherals can't detect erratic MOSI, MISO or CLK signals but ASICs can. \$\endgroup\$
    – Amit M
    Commented Sep 28, 2022 at 16:25

3 Answers 3

Reset to default
6
\$\begingroup\$

It looks like your oscilloscope enables Sinc-interpolation by default, which is a horrible default in my opinion. Either that or you accidentally enabled it yourself.

I have a feeling that you are using a sample rate that is much too low to see the information you're looking for - rise time and ringing.

Try disabling interpolation, and increasing the bandwidth to at least ten times higher.

\$\endgroup\$
4
\$\begingroup\$

In real life, there are no square signals.

Everything has a rise time, a slope. Lots of things ring and bounce.

For understanding the digital signal, you just pretend they are square!

That is, unless the signals are showing problems, then you pay attention to the analogue reality.

In your particular case, perhaps your scope probes aren't tuned, and you are introducing the ringing with the probes. Many oscilloscopes have a test signal and you can adjust the probes. Digital oscilloscopes have many other settings such as various interpolations -- for which see pipe's answer.

The non-squareness of your signal doesn't look particularly awful, but as commenters note, the ringing in advance of change is probably an artifact of the scope, not real in the signal.

\$\endgroup\$
2
  • 3
    \$\begingroup\$ I would say that a signal that shows ringing before the actual change is highly unusual... Time-travelling oscilloscope! \$\endgroup\$
    – pipe
    Commented Sep 28, 2022 at 11:16
  • \$\begingroup\$ @pipe you looked at it more closely than I did! edited. \$\endgroup\$
    – jonathanjo
    Commented Sep 28, 2022 at 11:29
4
\$\begingroup\$

Thanks you all for the useful answers and tips. I used option auto to adjust my oscilloscope and not everything was set properly. I made following corrections:

  • Compensated both my probes,
  • Went through acquire options and changed interpolation from sinx to x and sample rate from 25MSa to 100MSa,
  • Last step was to properly set trigger level as 50% of MOSI signal for example.

Basically this video (https://www.youtube.com/watch?v=PDkfjKyD8Ms) explains perfectly why I had this issue. I used the oscilloscope wrong!

Result is following. It looks better and there is no ringing: enter image description here

\$\endgroup\$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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