Impedance Measurements with Small Amplitude Test Signals

June 3, 2021 by Tim Ashworth

When taking an impedance measurement, the test signal amplitude can often be a factor that needs to be minimized to avoid unnecessarily disturbing the sample under test. This blog post presents measurements showing how even small test signals of just 500 nV can be used to take a reliable impedance sweep with the MFIA Impedance Analyzer or the MFLI Lock-in Amplifier with the MF-IA option.

Experimental Procedure

For the measurements taken in this blog post, we chose a simple piezo buzzer (TDK PS1240P02CT3) which nicely exhibits a non-linear impedance over a frequency sweep from 10 kHz to 1 MHz. It exhibits resonances at both 56.8 kHz and 301.5 kHz. These features allow us to check the fidelity of the measurement as we reduce the test signal amplitude (also known as the drive signal). We manually set the current input range to 10 uA due to the relatively high background noise of this low-cost component.

Figure 1: GIF capture of the LabOne® Sweeper module. The upper chart shows the absolute imepdance (absZ) and the current. The lower chart shows the phase(Z) as a function of frequency (10 kHz to 1 MHz), with twelve different test signal levels from 1 mV to 200 nV, plus output disabled. The history tab on the right hand side shows the actual test signal amplitude, labelled as "Drive". The impedance sweep from the sample is consistent as the test signal amplitude is decrease until 2 uV. (Click to zoom)

Although the MFIA is capable of test signals as high as 10 V, it has a four-stage output allowing for the 16-bit output signal to be spread over the full range. In this post, we are using the lowest output range, 10 mV. We start the series of sweeps at 1 mV, which is already below the lower limit of many other impedance analyzers. We then reduce the test signal amplitude after each sweep and repeat until the impedance spectrum is no longer distinguishable.

Figure 1 shows an animated GIF of a series of impedance sweeps from 10 kHz to 1 MHz, each sweep consisting of 800 points. The features in the sweep remain unchanged while the test signal amplitude is decreased, down to 2 uV when the resonance at 301.5 kHz is not longer reproduced. Below this, the features in the sweep become noisy down to a lower limit of 200 nV, where the general features can be seen in the chart albeit with higher noise. Finally, we disable the test signal by switching an internal relay to ensure zero voltage. The final sweep shows only a noisy phase sweep with no impedance information from the sample under test.

Further Considerations When Using Small Test Signals

Figure 2: Screenshot of the LabOne Scope module and Impedance module. The test signal is set to 500 nV, and the resulting output signal is displayed on the Scope in the lower part of the figure. The bit resolution can be clearly seen, which goes some way to explain the reduced fidelity and increased noise when measuring with small signals. (Click to zoom)

As we saw reduced fidelity and increased noise as the test signal dropped below 2 uV, let's have a look at the signal output using the LabOne Scope module. Figure 2 shows the Scope module displaying "signal output 1" which is the test signal used for the impedance measurement.

The Impedance module above the Scope shows a test signal of 500 nV (amplitude) which almost matches what we see in the scope (circa 900 nV peak to peak). The test signal is not exactly 1000 nV as we are very close to the digitisation limit of the MFIA, and we see the signal is made up of three bits, each 300 nV in height. Clearly, reducing any lower will result in a deterioration of the measurement signal, as we see in the trace for 200 nV (we can still measure impedance with just one bit!). Please also note that there is always a small offset in the voltage output, which is worth measuring when using small signals.

Summary

This blog post goes some way to answering the question "What is the smallest test signal I can use to measure impedance?" The answer for the MFIA with this sample is 500 nV. Naturally, for other samples this will depend on the impedance characteristics and also on the noise background of the fixture and cables.

Take care to consider the bit resolution of the output and any offset when using small signals, and also take data slowly to allow for sufficient noise suppression.

If you have any questions or comments, get in touch with us.