Measuring the on PCB 0603 Surface Mount Component Location with PicoVNA
To measure a component, the calibration reference plane needs to right at the component pads, requiring either an on PCB calibration or a compensating characterization of the on PCB feedlines (de embed or ref plane shift and normalization). Connect Port 1 to the Chip Component feedline. This will be a single port reflectometry measurement only, so Port 2 is not connected.
From the recommended default settings, adjust to display instead:
• Display Ch2 S11 Smith Chart
• Display Ch3 S11 TD (Hann), 0.25U/div, Ref 0 U at grat 6, timebase span 1.0 ns to 9.0ns.
• Display Ch4 S11 Phase, 45 °°/div, Ref 0 at grat 6.
• PicoVNA settings file: NMT kit component set.sta
In the example of figure 2, a 3.3 pF 0603 capacitor has been soldered to the open pads provided. Reference to the left hand plots alone might lead to a contradiction. The time domain plot indicates a short presented to high frequencies that transitions to an open in the long term (low frequencies). But the S11 plot indicates a constant magnitude of full reflection. This is not an impedance that passes anywhere close to 50 Ω in its transition between short and open.
The Smith Chart of course reveals what is going on: the component measures close to constant shunt capacitance rotating around the chart, a little lossy with increasing frequency and greater deviation beyond Marker 5. The chart enters the inductive region of the chart and wrapped phase flips on the phase plot. The marker readouts give us a capacitance value that varies a little with frequency, more strongly at higher frequencies. This is because measurement sensitivity and therefore accuracy fall away as the measured impedance deviates from Z0 towards short or open. Best measurement will be at Mkr 2 and particularly if that is moved to slightly lower frequency, closer to the 90 phase point.
The example of figure 3 below represents a 15.0 nH 0603 inductor soldered to the open pads. Here is the opposing contradiction in the left hand plots. The time domain plot indicates an open presented to high frequencies that transitions to a short in the long term (low frequencies). Again the S11 plot indicates a constant reflection and no passing anywhere close to 50 Ω. The Smith Chart again reveals what is going on: the component measures close to constant shunt inductance rotating around the chart, a little lossy with increasing frequency and greater deviation beyond Marker 5. The chart enters the capacitive region of the chart and wrapped phase flips on the phase plot. The marker readouts give us a inductance value that varies a little with frequency, more strongly at higher frequencies. This is because measurement sensitivity and therefore accuracy fall away as the measured impedance deviates from Z0 towards short or open. Best measurement will be at Mkr 2 and particularly if that is moved to slightly higher frequency, closer to the 90 phase point.
In both cases significantly smaller values can be characterized in this way. Significantly larger values might require narrower sweeps to a reduced maximum frequency.
