Ironically, even as the 10x passive probe is the most commonly used probe with an oscilloscope, it is often misunderstood and used incorrectly. This Application Note explains how to get the most out of it by avoiding common artifacts.
Best Practices for 10x Probes
There are five important best measurement practices to consider when using the 10x probe. Follow these guidelines to get the best performance with a 10x probe and avoid common artifacts:
Always compensate the probe using the calibration (Cal) reference signal on the front panel of every oscilloscope.
Always try to use a tip geometry that minimizes the loop inductance. A small spring ground tip or a coaxial connection to the DUT will reduce the ringing artifacts and extend the probing bandwidth to the highest frequency.
Always try to use a coaxial connection to the DUT to reduce radio-frequency (RF) pickup from the local environment.
When probing a low-impedance source, consider adding a 200-Ω series resistance to damp out the ringing from the tip loop inductance and the input capacitance of the 10x probe.
Always remember that the input impedance of the 10x probe is a 9.5-pF capacitance; it is not 10 MΩ.
Adjusting Probe Compensation
The first and most important step when using a 10x probe is adjusting its compensation. Even for signals with a 20-MHz bandwidth or with a rise time as long as 50 ns, compensation is important to minimize the probe’s distortion of the signal.
To check probe compensation, connect the probe to the Cal reference signal available at terminals on the oscilloscope’s front panel. Figure 1 shows an example of the connection in the front of an HDO4096 oscilloscope.
At the Cal terminal is a 1-Vpk-pk, 1-kHz square wave with an intrinsic rise time of about 3 ns and a source impedance of 800 Ω. The high impedance is a safety feature, guarding against accidental connection of the ground clip of the probe to the Cal signal output pin. In such a case, current flow is limited to about 1 mA, posing no danger to probe, oscilloscope, or operator.
This 1-kHz signal with a short rise time is available on every oscilloscope as a reference signal to adjust compensation for a 10x probe. The oscilloscope probe compensation is adjusted with a small set screw located in the end of the probe that plugs into the oscilloscope (Figure 2).
When the compensation is adjusted correctly, the measured square wave will look like a flat square wave. But when the probe is under- or over-compensated, the flat edge will be distorted (Figure 3).
Compensation is important even for low-bandwidth signals, considering that the impact of compensation can be seen in this 1-kHz square wave. Before using any 10x passive probe, always check its compensation with the square-wave Cal signal. If the square wave is not flat as shown at center of Figure 2, adjust the compensation screw until it flattens out.
The Importance of Tip Loop Inductance
The main impediment to the performance of a 10x passive probe is tip inductance. The highest bandwidth measurement is achievable only when using a coaxial connection at the probe’s tip. Whenever the signal and return path are separated to make contact to the DUT, two problems arise:
It introduces an inductive discontinuity at the probe tip, which will cause ringing, and
the probe tip acts like an antenna and will pick up RF noise from the environment.
The inductance of a circular loop with a 1” diameter is about 85 nH. Depending on the size of the loop created by the tip and its ground return strap, the tip’s loop inductance can be as high as 200 nH. The consequence of this higher tip loop inductance is to reduce the bandwidth of the 10x probe and to introduce a parallel LC resonance peak in the transfer function.
The height of that resonance peak will depend on the source resistance of the DUT signal. The lower the source resistance, the higher the Q and the higher the peak in the transfer function. Figure 4 shows the transfer function with a 200-nH loop inductance in the tip and a 50-Ω source resistance. Included is the measured response of a fast edge from a 50-Ω source with a large tip loop inductance.
However, when the source impedance of the DUT is reduced, the Q of the LC circuit composed of the tip inductance and the 9.5 pF input capacitance of the 10x probe increases. This increases the peak of the transfer function and contributes to higher ringing (Figure 5).
In this example, a 5-V power rail with very low output resistance was switched off and on. The transient response shows a lot of ringing at about 80 MHz. This is very close to the simulated transfer function peak at about 100 MHz, based on assuming a 200-nH tip loop inductance and the 9.5 pF input capacitance of the probe.
It is always important to minimize tip loop inductance. However, if the leads are spread apart to contact the DUT pads, there will always be some tip loop inductance. When the source impedance is low, the Q can be high and there may be artificial ringing from the probe. One additional method to fix this problem is by adding a series damping resistor to the probe tip.
Even though there is a 9-MΩ resistor in series with the probe tip, it is easy to forget that it has a 9.5-pF capacitor across it. At 100 MHz, the 9.5-pF capacitor has an impedance of just 250 Ω.
The value of the damping resistor should be large enough to provide a Q of about 1, but not so large as to overly lengthen the response time. For a Q of 1, the series resistance should be:
Decreasing the Q with the series damping resistor will also decrease the bandwidth and increase the shortest rise time that can be measured. When the bandwidth is 100 MHz, the shortest rise time that can be measured is about 3.5 ns.
Figure 6 shows an example of adding a 220-Ω series resistance to the tip of the 10x probe measuring the 5-V switching supply. This is slightly higher than the critically damped resistance, so there should be no peaking in the transfer function or ringing in the transient response. The simulated transfer function suggests a -3-dB point of about 100 MHz, which suggests the shortest rise time that can be measured is 3.5 ns.
While it is always important to minimize the tip inductance in a 10x passive probe, any tip inductance may induce artificial ringing when measuring a low-impedance power rail. Adding a series resistor in the tip of about 200 Ω will reduce this artifact, but it will also decrease the probe’s bandwidth to 100 MHz or lower.
If your rail-probing application requires a bandwidth higher than 100 MHz, consider using a probe such as Teledyne LeCroy’s RP4030 active power-rail probe, which does not have this limitation.
Practical Considerations for Reducing Tip Loop Inductance
To reduce the tip loop inductance of a 10x passive probe, consider using the two adapters that are typically supplied with them.
Large tip loop inductance will always reduce the 10x probe performance. It is always important to reduce the tip loop inductance as low as possible. Two adapters which are usually supplied with every probe will reduce the loop tip inductance. The worst case is to use the 3” supplied ground return lead. A better option is the small spring tip adapter connected to the ground return, and even better is a coaxial adapter (Figure 7).
Whenever possible, use a coaxial connection to the circuit board or the DUT. This reduces the tip inductance, reduces the RF noise pickup, and gives the highest bandwidth for the 10x probe. When you cannot use a coaxial connection to the DUT, use the short spring return path tip on the probe. When you cannot use the short spring tip as the ground return, be aware that the probe will have limited bandwidth, be more sensitive to RF pickup, and may show a ringing artifact when measuring signals with short rise times.
RF Pickup and Tip Loop Inductance
Any separation of the signal and return paths in a probe effectively results in an antenna. Thus, a larger tip loop inductance means increased RF sensitivity. When probing a signal on a board which might also have large near-field radiated emissions, such as with a switchmode power supply, it is difficult to discern actual measured voltage on the board from noise related to RF pickup.
One way to gauge ambient noise is with a second 10x probe with its tips shorted together, used specifically as a pick-up coil. When placed in proximity with the 10x probe measuring the voltage on the conductors, it can give a rough measure of the local RF noise. Figure 8 shows the two measured signals, one on the power rail conductor and one of the local RF pick up.
It is remarkable how noisy many boards are in the near field. A signal with a frequency of 100 MHz has a wavelength of 10 feet. This means all measurements a few inches from the board are near field. However, the presence of strong near-field emissions does not necessarily mean there will be strong far-field emissions. Many near-field sources drop off very quickly with distance.
When the RF-pickup signal is a significant fraction of the measured voltage on the conductor, be careful interpreting the measured voltage as a real signal. It could just be RF-pickup noise and not related to the actual voltage on the DUT.
The solution to this problem of RF pick-up is to use a coaxial-geometry connection from the 10x probe to the DUT. A connection of coaxial nature will minimize sensitivity to RF pickup.
Many 10x passive probe kits include a coaxial-to-PCB adapter for the 10x probe. The adapter can be soldered into the PCB on 100-mil-centered test points and provides a coaxial connection from the pads on the board to the 10x probe. Figure 9 shows an example of this adapter and the impact it has in reducing the RF-pickup noise on this rail.
A 10x passive probe is the workhorse probe used for many routine measurements. Its application sweet spot is for signals with bandwidths < 100 MHz and voltage ranges less than 400 V, when the smallest change to measure is >100 mV. Special care must be taken to perform reliable and accurate measurements outside of this range. Applying the best practices covered in the Application Note can increase the accuracy and reliability of measurements.