While accurate high-speed oscilloscope measurements can be challenging, using simple tips and techniques can significantly improve your measurements. This article collects some easy-to-implement methods to ensure that your time-domain measurements are fast and accurate. Although some of these methods seem basic, it is surprising that they are often overlooked.
To accurately measure the frequency response and fast rising edges, the oscilloscope and probe must have sufficient bandwidth. A good rule of thumb is that the bandwidth of the oscilloscope and probe (the probe also has bandwidth limitations) should be 3 to 5 times the highest frequency of the signal being measured. The -3dB bandwidth attenuation introduces a 30% amplitude measurement error, so the wider the oscilloscope and probe bandwidth, the better.
There is a very common situation. People use the oscilloscope produced by X company to measure with the probe produced by Y company. In fact, oscilloscopes and probes are not always interchangeable or compatible. The best practice is to use an oscilloscope and probe from the same company to eliminate any potential conflicts.
One of the most overlooked steps when making measurements with an oscilloscope is calibration. Calibration is a simple and easy way to ensure that every measurement you make is from the beginning, independent of the last measurement. Manual calibration should be performed before starting the measurement. If the oscilloscope has a self-calibration function, you should run this function before measuring. Under-compensated or over-compensated probes introduce significant errors in amplitude, rise time, and waveform distortion measurements of the signal under test.
Choosing the right probe for a specific measurement task will always result in the best measurement. Typically, a 10:1 probe is sufficient for general-purpose measurements; however, for low-amplitude signal measurements, you might want to consider using a 1:1 probe. When making high-speed measurements, you should focus on the probe capacitance. Probes with large capacitance values slow down the rising and falling edges, and can even cause oscillations in the input or output of some devices, such as high-speed op amps. Therefore, another consideration for measuring high speed circuits is the use of active FET probes. Active probes typically have a low contact capacitance (usually a few pF) and an extremely high impedance, so the active probe presents a very small load to any node under test.
The biggest mistake you can make when making a high-speed measurement is to use the probe ground clip, causing an oscilloscope probe problem. Using a ground clip is equivalent to adding a series inductor to the ground path. This series inductance and probe capacitance work together to introduce oscillation and overshoot. The best way to ground is to use the ground shield inside the probe, but you have to disassemble the probe. However, this is not difficult to do. First, loosen the plastic sheath from the probe tip and remove it from the probe. At this time, the metal grounding shield of the probe is exposed. Then, remove the ground clip from the probe and it is complete. Measurements can now be made by simply measuring a node and connecting the metal grounding shield of the probe to the closest ground point of the circuit under test. This method eliminates any series inductance and virtually eliminates all oscillations and overshoots. If you can’t find the closest ground point, circle a bus around the metal ground shield of the oscilloscope probe and ground it. You can use almost all small metal objects for grounding. I have used screwdrivers, paper clips and tweezers. My favorite is the dice, because the tip of the dice can be poked on the ground plane and can be inserted into the dense area of the PCB.
When making measurements of time difference or transmission delay, be sure to use two probes of the same length. The cable has a transmission delay of approximately 1.5 ns/ft. Cable lengths are different and can cause you trouble. For example, using a 3 ft. and a 6 ft long cable oscilloscope probe to measure the propagation delay, the cable length difference can cause an error of about 4.5 nanoseconds (ns), which is quite equivalent when it comes to distinguishing measurements in 1 ns. Big error.
Although the above tips and tricks alone are not worth noting, together they can significantly improve the accuracy of the measurement. Even if you use only a few of them in your measurements, they still ensure fast and reliable measurements every time you enter the lab.