1MHz the signal generator and coax parasitics can load the components
being measured if capacitance or inductance values are small.
Where only small voltages will result across Zuut (eg inductors at low
frequencies, capacitors at high frequencies, or low value resistors) use
Circuit B as it measures the low voltages accurately (but still not as well as
Circuit C at <1 MHz).
Where large voltages will be across Zuut (eg capacitors at low frequencies,
inductors at high frequencies or large value resistors) use Circuit A.
For high frequency applications, higher impedance, Circuit A yields the best
performance. The 50 ohm termination on the BNC coaxial cable acts as the
Reference resistor. Set R Ref to 50 ohms. For low impedance (such as
ground and power planes) use Circuit C. For best results above 1 MHz when
using probes do a Probe and Fixture Correction (see
Calibration
section).
The scope probes set to 1x give the best dynamic range, but because of the
approx 70pF/1M ohm loading the impedance under test should equate to
less than 1/10 of this - for example capacitances > 700 pF or impedance <
2k at 100 kHz or 200 ohm at 1 MHz. When using smaller capacitors, or
higher impedances, use the 10x probe setting, which loads about 17 pF/10
M ohm on the unit under test. This reduces dynamic range by 20 dB. You
can compensate to some extent by reducing the Max Bandwidth by a 1/10
(eg to 100 Hz). Make sure you do a Probe Calibration before starting out.
In both cases the Probe ground clip inductance ends up limiting the
maximum usable frequency to about 10 MHz. After that you need to use a
50 ohm terminated coaxial cable test jig. When using SMT parts we
recommend a 3 port BNC test fixture (see Impedance Measurement with
FRA video on www.cleverscope.com/videos) which you can build yourself.
Cleverscope also have a fixture available on our website.
A signal generator level of 2-4V works well, dependant on component
rating. For electrolytic capacitors apply a voltage bias by setting the Sig Gen
offset to a positive value (eg 0.5V), although data sheet values are
measured without bias.
If the system is capturing signals, but the spectrum result is 0 for that
frequency, the system has determined that the signal to noise ratio (SNR) is
poor, and the measurement suspect, and therefore not displayed. Improve it
by lowering the Bandwidth, or increasing the signal level.
Keep an eye on the Scope Display, and make sure the signal fills at least
10% of the amplitude range at maximum amplitude, but is not larger than
the graph range (otherwise clipping and errors will result). You can set
Sweep Amplitude to Auto, and higher response goal V, such as 300mV if
unsure, but this will be slower. The wide dynamic range will ensure good
operation if you are using a constant sweep amplitude. You can use Auto
Setup to get a reasonable approximation to the settings you need, and then
turn Auto off, and fine tune if necessary.
Bad SNR will manifest as lots of noise in the measurement curve. If you are
having problems with noise in the plot, decrease the measurement
bandwidth or increase the signal generator amplitude. A good starting point
is 1kHz, but if too much noise, reduce BW to 100 Hz. Note that the ratio
between Stop Frequency and Bandwidth cannot exceed 1M. If it does, the
Bandwidth will be raised, and resolution at low frequencies reduced.
To Next Page
Loading...