Notes on the oscilloscope analysis

[JoshMuir]

Hi Raziz,

I'd like to start by saying I really like the project, and appreciate the in-depth analysis and discussion you have shared. I will definitely be taking notes from this on how I should present my projects in the future.

I would just like to point out for you that in your analysis of the switching node you have highlighted the 'ringing' that is present in your measurements as a flaw of your circuit design. I would like to contest that claim and explain why. In short, I believe the ringing is actually an artifact from insufficient speed from your oscilloscope.

From what I understand, you are using the Analog Discovery 2. I'll post the specs below.

Channels: 2
Resolution: 14-bit
Sample Rate: 100 MS/s
Bandwidth: 10 MHz+
Input Impedance: 1 MΩ
Voltage Range: ±25V
Max Buffer Size: 16384 Samples/channel

Now we can see the bandwidth of this scope is 10MHz, and the signal you are measuring is only 1MHz. Should be plenty right?

Typically, if you were measuring a sinusoidal waveform this would be fine. However because you are measuring a digital signal, you must look at the rise times and fall times of your signals.

From your image, the signal looks to rise (10% to 90%) within half of a tick.
Given 10 ticks is 0.5us, we can determine 1 tick is 50ns and your rise time is about 25ns.
f = 1/T = 40MHz

Given the bandwidth of your scope is only 10MHz, and rule of thumb is to have at least 3x the bandwidth of the fastest rise time you're measuring, this is insufficient.
The consequence of this is the overshoot and undershoot you are seeing from your scope, as it cannot accurately represent the signal.

Additionally, this ringing can be caused by a long ground lead creating inductance on the probes path, but that is another story.

While not so consequential in this circumstance, as you get into more advanced digital circuits, artifacts such as this can cause you to chase your tail.
This is also an important concept to understand when it comes to preventing and protecting your circuitry from noise. Digital signals with extremely fast rise times can cause you lots of trouble.

Overall I really enjoyed reading about your project, and I hope that you find this information useful! I learned this through many wasted nights on my own projects, investigating issues that didn't even exist!

[Raziz1]

Hey @JoshMuir, thank you so much for the feedback. I really appreciate you taking the time to dive into the project and share your insights. I actually had the chance to take another capture of the same switching node using a Tektronix MSO44B (up to 1.5 GHz), and as you mentioned, the ringing is pretty much gone. It looks like it was indeed due to the bandwidth limitation and likely some stray inductance from the probes.