Oscilloscope mistakes and how to avoid them

Steve Sandler, Founder, AEI Systems

September 12, 2013

Steve Sandler, Founder, AEI SystemsSeptember 12, 2013

Editor's Note: Avoiding these some oscilloscope pitfalls will get you better data with minimum effort. Steve Sandler takes a look at four common scope mistakes and gives details on how to avoid them. This particular mistake focuses on bandwidth and sample rates.

The engineering community uses oscilloscopes more than any other piece of equipment, and yet many of the published results are questionable at best. Some errors are very common, and so we can eliminate a great deal of bad data by considering a few simple, but key points. This first part of a four-part series covers: insufficient bandwidth and/or sample rate.

The majority of engineers only have one oscilloscope on their bench. This is likely due to cost. While this may be acceptable, it presumes that the one oscilloscope meets all of our time-domain measurement requirements. Let’s consider just a few of the scenarios that a typical engineer might face, and evaluate the bandwidth and sample rate required for the measurement.

Case 1: a simple high-speed CMOS logic gate
It is rare that any electronics these days would not include some type of digital gate or buffer. In this simple case, we are using a NC7SZ04 Tinylogic inverter gate, connected to a 10-MHz SMD oscillator. The 10 MHz is not significant. It is just a handy example. In using such a logic gate, we might consider the rise and fall time of the gate. The rise will impact the noise generated on the supply voltage rail and will also provide some guidance on the narrowest glitch we might need to capture if we needed to troubleshoot the circuit.

Most engineers might not guess how fast these very common gates are. In this first measurement, we connected a 30-Ohm resistor and a 0.01-uF capacitor in series with the output of the logic gate. The scope is then terminated into 50 Ohms, allowing it to perform at its maximum performance level. In this measurement, we are using a LeCroy 640Zi 4-GHz oscilloscope. The rise and fall time are shown in the oscilloscope measurement in Figure 1 with a 40 GS/s sample rate.

Figure 1
The actual rise and fall time of the logic gate output (399 pSec/518 pS).
The actual rise and fall time of the logic gate output (399 pSec/518 pS).

Reducing the sample rate to 5 GS/s results in a 16.5 percent error in the rise time measurement. Likewise, reducing the oscilloscope bandwidth to 2 GHz would also result in a 16.5 percent measurement error.

Figure 2
Measured rise and fall time of the logic gate output at 5 GS/s (466 pSec/559 pS).
Measured rise and fall time of the logic gate output at 5 GS/s
(466 pSec/559 pS).

Accurately measuring the high speed CMOS gate requires at least 10 GS/s and 2.5 GHz of measurement bandwidth. The measurement bandwidth includes the limitations of both the probe and the oscilloscope. Since we are using a 50-Ohm matched connection for this measurement, there is no probe limitation. We could use a 2.5-GHz oscilloscope.

Case 2: a linear regulator transient
A 1-kHz step load applied to a linear regulator results in a narrow voltage excursion, which exponentially recovers to the starting voltage. The current step is approximately from 50 mA to 200 mA, with a 15 nS rise time. The voltage transient, recorded at 10 MS/s, 50 MS/s, and 250 MS/s is shown in Figure 3. The transient is approximately 60 nS wide, as recorded at 250 MS/s. The excursion amplitude is approximately correct at 50 MS/s, though the fidelity of the image is poor. The minimum sample rate for high fidelity should be approximately 10/ transient width or 166 MS/s in this case, while a 20-MHz oscilloscope could easily capture the waveform.

Figure 3
A linear regulator load step response at 10 MS/s, 50 MS/s, and 250 MS/s.
A linear regulator load step response at 10 MS/s, 50 MS/s, and 250 MS/s.

Figure 3: A linear regulator load step response at 10 MS/s, 50 MS/s, and 250 MS/s.

The results of these two cases might be used as a guide for determining the oscilloscope requirements based on the application, though there is a danger in doing so as seen in Figure 4. Figure 4 shows the output voltage of a linear, low drop out (LDO) voltage regulator, which can be seen oscillating at approximately 140 MHz. If the measurement was performed using a 20-MHz bandwidth scope, consistent with the transient response requirements, such an oscillation would easily be missed.

Figure 4
A linear regulator load step response at 10 MS/s, 50 MS/s, and 250 MS/s.
A linear regulator load step response at 10 MS/s, 50 MS/s, and 250 MS/s.

Recommendations
Understand the bandwidth and sample rate you need for your measurements. You will generally require more bandwidth than you think. For most applications we recommended at least a 1-GHz bandwidth. If you are working with high-speed circuits, such as microprocessors, FPGA’s, CPU’s, or others, we recommend a minimum of 2.5 GHz and no less than 20 GS/s. FPGA manufactures currently require power supply assessments to 10 GHz. A 4-GHz scope can often be used, but at least 20 GS/s or preferably 40 GS/s are required.

The measurement bandwidth is limited by both the scope and the probe. If you are using or plan to be using a probe you will likely want to increase the scope bandwidth to account for this.

To read more go to Oscilloscope Mistakes: Part 2, Part 3 and Part 4.
(This article has also been published on EETimes.)

Loading comments...

Most Commented