RLC Measurements

The MISC R, L and C Measurement Automated Test allows you to sweep a passive component (resistor, inductor or capacitor) using the QA40x analyzer. The output from this test is an impedance measurement along with a calculated inductance or capacitance. The Automated Test plug-in is accessed from the Automated Test menu:

The QA461 is used to provide the drive. With up to 1.6A of source/sink, the QA461 has plenty of drive to push component tests to extremes in some cases. The typical setup is shown below:

In the diagram above, note that the single-ended output of the QA403 is driving into the QA461 input. The QA461 has 20 dB of gain, so if you set the QA40x to drive at -20 dBV, that will result in a 0 dBV = 1Vrms drive signal to the DUT. The DUT in this case is a resistor, capacitor or inductor.

A DUT test fixture was created to allow easy connections to R, L and C surface mount components. The PCB layout of fixture is show below:

The schematic for the fixture is shown here:

And the actual board is shown here:

What the DUT fixture allows is sliding small surface mount components up into the rails for measurement. The rails are 1.5mm square copper bars, angled such that any size surface mount can be slide into position. The blue arrow in the picture above shows a 0603 capacitor lodged into position.

When the fixture is connected the QA461, what you have is the QA461 driving directly across the DUT. This drive comes from the QA461 output via the INPUT BNC on the DUT test board. And an OUTPUT BNC allows the voltage across the DUT to be directly sensed. The OUTPUT CURRENT SENSE on the QA461 allows the current flowing through the DUT to be sense. That goes into the right channel on the analyzer.

And with the voltage across the DUT and the current through the DUT both known, we can deduce the impedance of the DUT with some fairly simple complex (as in real and imaginary) maths.

In the next section, we're going to start by exploring limits of the hardware. For these tests, we'll use a QA461, but you can readily make the measurements using just the QA40x. You just can drive the DUT with as much current as the QA461 can provide.

10 Ohm Resistor

To start, let's first test a 10 ohm resistor using the setup outlined above. For the automated test plug-in, we use the following settings:

Above we're specifying -40 dBV expo chirp from the QA40x analyzer. With the 20 dB of gain in the QA461, this translates to a drive current of -20 dBV, which is 100 mVrms. We'd expect to see a current of 0.1/10 = 10 mA. When we run the measurement, we can see the main screen on the analyzer as shown:

Note that the blue trace (left) shows we drove across the DUT at the constant level of -20 dBV. And the red trace (right), which is measuring the current, shows -40 dBV. Since the QA461 has an effective output impedance of 1 ohm, this -40 dBV translates to 10mVrms, and the one ohm effective output impedance means that 10 mVrms represents 10 mArms.

In reality, the QA461 output sense resistor is 0.02 ohms, and that is followed by an INA199 current sense amplifier with a gain of 50. The 0.02 * 50 gives us the 1 ohm effective output impedance.

The plot from the Automated Test is shown below. In this plot, we can see the impedance (blue) is very nearly 10 ohms. We can also see the part has 40 uH of inductance or so. In reality, the part probably has some stray inductance measured in the nH range. But the cabling and layout dominates here.

If we run the test fixture open-circuit (nothing installed), we get a main plot as follows. In the plot below, note that the current at 1 kHz is around -75 dBV or so, which is about 177uV or 177uA. But we don't have a DUT connected, so the amount of current flowing to ground should be zero. This non-zero current "floor" is due to two factors. The first is because the output voltage sense goes through a 20 dB resistive attenuator formed by a 4.22k ohm and 470 ohm resistor. This means that with -20 dBV = 10mV of drive voltage, we can expect 10mV/4500 = 2.2uA of "lost" current to the divider. But here our "lost" current is quite a bit larger: 177uA. The second factor is the offset voltage of the INA199. Per the TI spec, the nominal offset (referenced to input) is +/- 5uV, and with a gain of 50, this suggests we should readily expect +/-250 uA of offset when zero current is flowing.

The test plot from the above sweep as follows. Note that the stray reactance is reported as being capacitive in the nF range. But more importantly, note that the impedance is around 500 ohms or so.

Let's run this open circuit test again (remember, we're running open-circuit), this time with -10 dBV of analyzer drive, which means 10 dBV (3.16Vrms) of drive out of the QA461. First, we see the main screen plot. Note here the current level (red) is higher and there's less noise.

And the impedance plot is as follows. Note here we're showing about 15kohms at 1 kHz. And so, the ability to better sense the current using the higher drive voltage has helped a lot. But we're still not where we want to be.

OK, so let's shift gears here a bit and re-visit the speaker impedance test setup from the link HERE. What the link below is showing is how to add in a series R to measure the current, and then measure that differentially using the right channel.

For this test, another board is used, which is shown below. This board enables the insertion of a series R into the BNC signal path, and has a differential tap off the resistor.

Now, when we make an open circuit measurement, we see the resistance is reported as about 43k.

OK, so now we have some understanding in what it takes to measure a higher value resistor. And that means a higher-value sense resistor. And we can see that a 10 ohm sense resistor can probably let us measure up to 1kohm or so with great accuracy. Using the 10 ohm sense resistor, let's re-run the RLC test. First, the main plot. For this test, the output drive was reduced to -20 dBV. With a small sense resistor, we'd expect the voltage sense trace (left = blue) to be at the 0 dBV level. But we have a much larger sense resistor (10 ohms), and so the voltage is showing around -5 dBV. And the current trace, curiously, is at the same level and hidden in this plot.

And the impedance plot is below. Note the impedance is 10 ohms and ruler flat. The inductance is around 100 nH.

Testing L's and C's Using the 10 ohm sense resistor, let's run a 100 uH (10%) inductor rated for 50 mA (5.2 ohm DC resistance). Let's use a -40 dBV drive voltage out of the QA403, which will result in a -20 dBV = 100mVrms into the sense resistor (since the QA461 has 20 dB of gain), and a lesser voltage into the inductor. The main plot appears as follows:

And the impedance trace as follows. The inductance is showing about 120 uH, which is about 20% high. The impedance is where expected.

Now instead of a gentle 100mVrms test signal, let's increase it to -30 dBV out of the QA403:

And next increase it again to -20 dBV out of the QA403:

And next we increase it again to 0 dBV drive out of the QA403. Here we can see the inductance has collapsed. When inductor current exceeds the saturation current, then the inductance drops off quickly. And we've obviously exceeded the saturation current here. The inductor audibly complained and you could hear the swept tone radiating from the inductor via microphonics.

Let's repeat the test with a 2.2uH inductor rated for 640 mA and 429 mOhm of resistance. Since this inductor has much a much higher saturation current, we'll use a -20 dBV drive. Here, the impedance is as expected, but the inductance at 5.5 uH is quite a bit more than the 2.2 uH we expect.

The spec on the part (CBC2012T2R2M...spec HERE) shows the following. Note that the inductance value is specified at 7.96 MHz. And we can see that at 100 kHz, the inductance is about 3 uH. Eyeballing, the plot out another 2 orders of magnitude (from 100 kHz to 1 kHz) it seems plausible. But be aware that inductance isn't flat over frequency.

For completeness, we can measure a 0 ohm 0603 and we get the following plot. The impedance is reported at 40 mOhm, and the inductance is reported around 50 nH or so

Measuring a Capacitor

A 4.7 uF capacitor was measured at a -20 dBV drive level out of the QA403:

Summary There are a lot of possibilities for improvements to this plug-in. Graphing real and imaginary quantities, phase and even providing stepped measurements so that inductor saturation can be recognized. The greatest limitation is the upper frequency bound of the analyzer.

Keep the following in mind when using this plug-in:

1. If using the QA461, there is a lower limit on current you can detect that will be limited by the INA199 offset used inside the QA461, and also the INA199 noise.
2. If using the QA461, there is a 80 kHz bandwidth limit on the current sense output. This is inherent to the INA199.
3. Keep your drive voltage up across the DUT, but not so high you harm the DUT due to excessive power or current
4. If you use a separate series resistor for the current sensing, make sure your polarity is correct AND size the sense resistor so that it makes sense for your DUT. Understand the open circuit and short-circuit readings, and use those as a guide to dial in the sense resistor
5. In all cases, seek settings that will give you low-noise traces on the primary display. The math involved will tend to magnify noise. If your primary display is noisy, your results will be even more noisy.