🚀 In today’s Newsletter, I want to discuss some basic lab skills when using an oscilloscope.
📅 But, before that… Have you already signed up for the New Frenetic Product Launch event, live on November 2nd at 5 pm CET? We'll present our innovative procedure for the design of Magnetic Components, with a technical demonstration and a Q&A session. You can register for free here!
🔎 Let’s go back to our topic… After spending hundreds of hours in front of an oscilloscope screen, I’m always amazed by all the basic questions that still come up to my mind.
Questions like “Is this ringing observed true?”. Meaning, “Is this oscillation caused by the circuit, and the scope accurately depicts the waveform, or have I fallen into a measurement trap, where the oscillation is due to the probe?”
This is a very good question, and its answer comes down to parasitic elements. You need to have the scope input, as well as the probe structure modelled to tackle the question. I’m going to address a specific problem that every Power Electronics engineer will face in the lab to make things interesting and, in parallel, present the necessary theory to support each logic step taken to get to the answer.
⚠️ Let’s start with… The problem
Figure 1. A full bridge configuration
Assume that I’m probing “V1” voltage in Figure 1 to see the waveform of the full bridge. Then you see a bit of ringing on the rising/falling edge of the pulse.
“Is it because of the actual circuit or because of your scope?”
💡 The rise time – bandwidth correlation
Is the bandwidth of your 100kHz power converter, 100KHz? Please don’t say yes… 😜
If the power world was operating with sinewaves, then the former would be correct, but almost everything is square waves when it comes to SMPS. Power switches are either ON or OFF. The real question is how fast this switching happens. In other words, what is the rise/fall time of square voltage pulses.
Figure 2. Typical square pulse with some ringing observed in a full bridge
In Figure 2, a typical pulse with some ringing is shown. Rise time is the time it takes for the voltage to rise from 10->90% of its final value. In our case, the time from 40->360V. Essentially that’s the switching time of the switching element of the bridge in Figure 1.
🤔 Doesn’t this remind you the step response of a low pass, RC filter? If we forget the oscillations and some high order effects, notice the little bumps and imperfections during rise time, Figure 3 is what we get. That’s the basis of the following analysis and the results closely approximate reality, resulting in no need to evaluate more complex models.
Figure 3. Example of an RC low pass filter
It can be easily shown that in a low pass,1-pole, RC filter this holds true:
The RC filter presents a pole at:
From (1), (2):
So, the bandwidth of the filter (BW):
⚙️ From (4) we know the bandwidth of the filter. The BW is simply the frequency where the output is -3db down, or simply the output signal amplitude is 0.707 that of the input signal, looking at the output frequency response. Since we are talking about step responses of a circuit resembling a simple RC low pass filter, the BW for now is just in direct relationship with the rise time.
One important point here is that oscilloscope manufacturers suggest a x3-5 scope-probe BW to measure with less than 3% amplitude error. That is true for both sinewaves and square pulses. If the BW of the scope-probe system is the same as the signals BW, then in case of a sinewave we would observe a -30% less amplitude than the actual. In case of a square wave, an overdamped behavior would be observed at the pulse edges.
Figure 4. Passive x10 probe and oscilloscope simulation with variable Llead
Let’s take a look at the probe-scope system in Figure 4. Notice L1 (“Llead”) plays the role of the earth/GND alligator clip wire, a typical passive probe comes with. It doesn’t matter if the inductance is in the ground wire or the signal wire because it’s connected in series with the probe-scope circuitry.
A x10 probe, divides (why not call it ➗ 10 ?) the input voltage by a factor of 10. This can be seen from resistive divider R3 and R5 and the capacitor divider C3 and (Cline +C4+C5). Cline is the capacitance of the lossy transmission line O1 (that is the coax cable of the probe), with a capacitance of 85pF with a 71pF/m, 1.2m long coax cable. Notice C4 is just a trimmer cap, used to match total capacitance of Cline+C4+C5 to be 1/9 of that of C3, to preserve the 1/10 division ratio.
All this fuss is to show you the simulation result in Figure 5.
Figure 5. Probe frequency response with variable Llead
Roughly speaking, if you measure a random detachable ground lead, you’ll get 10nH/cm. So that’s where the 200nH inductance comes from. In green we see the flat response of the probe far exceeding the 100MHz probe rating. Notice the -20db which translates to a 1/10 division of the input signal.
☝️ But when the earth lead comes into place, resonance happens at around 90-100MHz. When it comes to resonance, remember that the amplitude change effects start appearing a decade earlier at around 10MHz. At 30MHz the amplitude change is still small though at 0.62dB away from the -20dB mark.
This means that a sinewave with a frequency of 30MHz will actually be depicted practically unchanged by the oscilloscope. The same roughly applies to a square wave with a 30MHz bandwidth. Applying a controlled rise time pulse (V1 in Fig. 4) to the probe-scope simulation we get Figure 6.
Figure 6. Controlled rise time step input response with a 200nH earth lead
💥 The verdict
Whenever probing with a x10 or even x100 probe (similar results) a power node, as shown in Figure 1, if the rise time of the waveform is over 35ns then ringing isn’t caused be the probe-scope system (less than 2% overshoot) and is caused by the circuit under test. So, any ringing depicted is real!
Since this knowledge was acquired from experience and through reading, any feedback is valuable. Mistakes happen all the time, the point is to be open to address them😉.
😎 Hope you like it, stay tuned!
🚀 Join our live event on November 2nd, and uncover the latest Frenetic secrets!
Doug Ford, 2009, “The secret world of probes”, Silicon Chip, October, pp. 16-23
Tektroniks, 2009, “Understanding Oscilloscope Bandwidth, Rise Time and Signal Fidelity” Available at: people.ece.ubc.ca/robertor/Links_files/Files/TEK-Understanding-Scope-BW-tr-Fidelity.pdf
Teledyne Lecroy, 2013, “Passive Probe Ground Lead Effects”, Available at: cdn.teledynelecroy.com/files/appnotes/passive_probe_ground_lead_effects.pdf
Tektronix, 1997, “ABCs of Probes”, Available at: web.mit.edu/6.101/www/reference/ABCprobes_s.pdf
Aditya Rao, Melinda Piket-May, Eric Bogatin, 2021, “Bandwidth of Signals: What is Important, Rise Time or Slew Rate?”, Signal Integrity Journal, 4 May, Available at: signalintegrityjournal.com/articles/2092-bandwidth-of-signals-what-is-important-rise-time-or-slew-rate
Eric Bogatin, 2018, “Back to Basics: Bandwidth and Rise Time”, Signal Integrity Journal, 14 May, Available at: signalintegrityjournal.com/blogs/12-fundamentals/post/853-back-to-basics-bandwidth-and-rise-time
