How to tune (compensate) x10 oscilloscope probes

For passive probes, in order to minimize the capacitive load on the device under test, most probes use an x10 (also known as 10:1) attenuator. This can often be adjusted or compensated to improve the frequency response. There are two types of compensation: low frequency (LF) and high frequency (HF). Some probes have only LF compensation, while others have both types. Oscilloscope probes are shipped from the factory with HF compensation and do not need to be adjusted, but if you wish to use a different probe on your oscilloscope, you may need to adjust its HF compensation.

Figure 1

Low Frequency Compensation

Low Frequency Compensation (LFC) involves adjusting the frequency response of the x10 probe in the kHz region.LFC must be performed before High Frequency Compensation (HFC).

Figure 2

Figure 2 shows a model of a typical probe. cp is the stray capacitance of the probe tip itself. r1 is a series 9 MΩ resistor used to isolate the capacitance of the cable and oscilloscope inputs from the device under test. the result is a 10:1 attenuation. The result is a 10:1 attenuator with an input impedance of 1 MΩ to the oscilloscope Rscope.

Ccomp1 is a variable capacitor that forms the LFC tuning portion of the probe. cp is used to adjust the time constants of R1 and Ccomp1 to match the time constants set by the Cscope, Ccable and Rscope. In effect, we have a resistive divider at DC and a capacitive divider at high frequencies (above a few 100 kHz).Ccomp1 represents the trimmer on the top of the probe near the attenuation switch.

Ccomp2 and Rcomp represent the high frequency compensation (HFC) portion of the probe.

The easiest way to compensate for the LFC portion of the probe is to input a square wave with relatively slow edges, but importantly, no overshoot.

Figure 3 shows what the waveform should look like when the LFC is correct. If the value is too high, the probe's high frequency (HF) gain will be higher than its low frequency (LF) gain, and if the LFC is too low, the HF gain will be lower than the LF gain.

Figure 3

High Frequency Compensation

Two variables affect the probe's HF response: cable impedance and oscilloscope input impedance. The oscilloscope input is usually not perfectly capacitive, but also has some series inductance and nonlinearities.

Figure 4 shows the typical characteristics of a ceramic chip capacitor used in an oscilloscope input. The impedance decreases before it starts to increase again with frequency. This is due to the series inductance of the capacitor. The point of minimum impedance is called the resonant frequency and indicates the frequency at which the inductive impedance and capacitive impedance are equal.

Figure 4

This figure provides insight into the fact that at high frequencies (VHF), the input to an oscilloscope is not as simple as a resistor in parallel with a capacitor, and the nonlinear nature of the PCB further complicates the situation. The input impedance of a VHF oscilloscope consists of a 1 MΩ ground resistor and a number of stray capacitors and inductors. Each of these has its own series and parallel inductive and capacitive components, which are often nonlinear in VHF, further complicating matters.

To compensate for nonlinearities, HF probes tend to shunt the oscilloscope input with a very small capacitor and a series resistor on the BNC. This helps to move any nonlinearities into the higher frequency region, beyond the expected range of the probe, without causing severe overshoot.

Rcomp and Ccomp2 represent the HF tuning components of the probe. This circuitry is typically located in a shielded enclosure at the BNC connector on the PCB to minimize the effects of cable and noise pickup. A typical probe has two of these RC networks, each with its own tunable resistor. One controls the midrange frequencies and the other controls the high frequencies. Both should be tuned until the correct response is obtained.

To tune the HFC of the probe, a square wave with very fast edges must be fed into the probe. The waveform must have a fast edge (rise time 3 times shorter than the probe) and very little or no overshoot. We use signal generators with less than 3% overshoot and very fast rise times. The VSWR of the 50 Ω terminator used with the pulse generator should also be taken into account, as low-quality terminators can cause additional overshoot.

When tuning the probe, you should first observe the oscilloscope's impulse response in order to match the probe response to the response of the directly connected oscilloscope inputs.

Slight overshoot and ringing occurs at about 1 GHz. This is mainly due to stray inductance in the PCB traces leading to the first amplifier, as well as some ringing caused by the amplifier itself.

Figure 5 shows the appearance of the overcompensated and undercompensated impulse response. The goal is to make the response as flat as possible. When tuning the probe, attention should be paid to the rise time.

Figure 5

Figure 6 shows a perfectly compensated probe. A slight hump is desirable because it provides more bandwidth for the probe and oscilloscope combination than the oscilloscope alone without much overshoot.