Information

4 Measuring methods

4.2 Measuring cable routing

There is the possibility of influence from external noise sources when using long,

unshielded measuring cables. Although they allow uncomplicated measurement, the

problem arises that the cable is located within the range of another mains-powered

electrical device and therefore interference may be coupled into the cable, which

causes measurement errors. It is therefore essential that no mains cable is located in

the immediate vicinity of the probe, because even the power socket of the isolating

transformer can be susceptible to capacitive coupling.

In order to avoid coupling of electromagnetic fields into cable loops, the supply and

return lines should always be twisted. The use of shielded measurement cables is

recommended.

4.3 Compensation of the probe

The measurement results should not depend on the probe used. Therefore, prior to

each measurement, the probe should be matched to the particular oscilloscope input.

This can be achieved by adjusting the probe. Oscilloscopes have a calibration con-

nector for this purpose, which provides a perfectly shaped square wave to which the

proble is attached. The signal displayed on the screen is tuned at the probe until there

is no overshoot or undershoot on the rising and falling slope. This allows fast signal

(e.g. DC/DC converter switch node) to be measured correctly.

4.4 Selection of the duty cycle

Different probes (divider ratios) have a considerable influence on the measurement

result, which is very clearly shown with two measurement curves (Fig. 4.2).

As the layout of standard EVAL boards is optimized and therefore shows no input sig-

nal with steep edges, an unfavorable layout design with a high level of ripple is taken

to explain the issue. The voltage peaks occurring here are strongly attenuated with a

1:1 probe in contrast to a 1:10 probe. This is explained by the fact that the 1:10 probe

has a higher internal resistance (10 MΩ versus 1 MΩ) than the 1:1 probe and there-

fore hardly influences the rise time of the voltage signal. Furthermore, the probe’s high

bandwidth avoids an incorrect visualisation of the measurement voltage. Signals such

as the switch node voltage that show a low rise and fall time can realistically only be

recorded using a probe with a bandwidth of > 100 MHz and an input divider of 1:10.

PAGE 1
4 Measuring methods 4.2 Measuring cable routing 4.2 Measuring cable routing There is the possibility of influence from external noise sources when using long, unshielded measuring cables. Although they allow uncomplicated measurement, the problem arises that the cable is located within the range of another mains-powered electrical device and therefore interference may be coupled into the cable, which causes measurement errors.
PAGE 2
4 Measuring methods 4.5 Ground connection of probes 0,4 0,3 0,2 0.1 V/div 0,1 0 –0,1 –0,2 –0,3 1 μs/div 1:10 probe 1:1 probe Fig. 4.2: Comparison probe attenuation 1:1 and 1:10 4.5 Ground connection of probes Inductive influences of PCB tracks or THT components can cause unwanted coupling into the probe at different locations of the switching regulator. Through the use of the probe’s long ground conductor, additional parasitic effects are coupled into the circuit.
PAGE 3
4 Measuring methods 4.6 Probe compensation Fig. 4.4: Probe connection for correct measurement of the output/input ripple 4.6 Probe compensation Probes are usually adapted to the input characteristics of certain oscilloscope models. It is observed, however, that there may be differences between the individual oscilloscope models and even between individual probe inputs of a device, which distort the measurement results. So the aim should be to eliminate such differences.