WT3 Measurements and Analysis

I. Driver Impedance

So how accurate is this thing? How does it compare to other measurement tools out there, in terms of accuracy? What features does it have? To answer these and other questions, let’s look at how the WT3 stacks up against LinearX’s LMS system. Some back-to-back measurement comparisons are in order.

Impedance plots of a 10” Woofer, 4” midrange and 1” concave-domed tweeter were made using the WT3 and LinearX’s LMS system. The results for each driver are shown below in the left hand column. In each case, the blue plot is the LMS data and the red plot is the WT3 data. To further aid in comparison, the WT3 plots were normalized to the LMS plots. The results are shown in the right hand graphics column.

Figure 1a & b. At left: Tweeter free-air impedance, Mag. & Phase. At right, WT3 data normalized to LMS data.

Figure 2a & b. At left: Midrange free-air impedance, Mag. & Phase. At right, WT3 data normalized to LMS data.

Figure 3a & b. At left: Woofer free-air impedance, Mag. & Phase. At right, WT3 data normalized to LMS data.

II. Thiele/Small Driver Parameter Derivation

Aside from its ability to capture impedance measurement data, the most important feature of the WT3

is its ability to derive a driver’s Thiele/Small parameters from the impedance plots.

Basically, the process runs as follows: (1) the free-air impedance of the driver under test is measured; and (2) the delta-mass, delta-compliance impedance plots are generated or specified db spl method is used. The raw measurement data is then processed by WT3, mathematically deriving the required Thiele/Small parameters. With those parameters at hand, a suitable cabinet design can then be worked up.

To further put the WT3 to the test, two enclosures were designed and the amplitude response & impedance plots were modeled for comparison purposes. One design is based on the Thiele/Small parameters as generated by LMS data and the other based on the same parameters generated by WT3 data. Below is a table of the Woofer’s Thiele/Small parameters as derived from the two systems measurements.

Modeling a sealed, high-pass enclosure (Butterworth alignment) with a Ql = 2, produced a cabinet with a net internal of 1.66 ft^3 (46.95 ltrs) for the WT3 data and a cabinet with a net internal volume of 1.70 ft^3 (48.2 ltrs) for the LMS data, an enclosed volume difference of just under 3.0%. Showing above are the amplitude response plots (left) and impedance plots (right) of the 2 systems modeled. Purple plots are for the system based on LMS data and green for the WT3 data.

Though the parameters are generally in good agreement, there are differences. Taking into account differing drive conditions and different post-processing algorithms in each of the products helps to keep these differences in proper perspective. All in all, not bad for a product that costs less than 1/10^th the price of the LMS system!

III. Passive Component Impedance Measurement

Another handy feature is the WT3’s ability to measure the impedance of various electrical components, such as the caps, resistors, and inductors found in the passive crossovers commonly used in loudspeaker systems. It should be noted in passing that the WT3 is not officially specified for measuring capacitors, though it did a pretty good job of it, nonetheless

A good deal of useful information can be derived from a components impedance plot, including the actual resistance, capacitance and inductance values at a specified frequency (especially useful information when working up loudspeaker crossover network designs). Let’s compare the WT3 and LMS performance in measuring the impedance of some component samples.

Above are (from left to right) impedance plots for a resistor, inductor and capacitor. In each case the blue plots are generated by WT3 and the green plots by LMS.

For the resistor, the WT3 has accurately measured the component’s value (as confirmed by an HP 3455A Digital VOM) across the entire measurement frequency spectrum. (The odd jumps seen in the green plot at 100, 1k & 10k are systemic measurement artifacts generated by LMS; the WT3 did not present any such glitches and overall presented a cleaner, more accurate plot of the test resistor’s impedance.

For the inductor, the LMS and WT3 impedance values across the frequency spectrum agreed to within a fraction of an ohm.

The capacitors impedance values were in good agreement across the 1 kHz – 10 kHz portion of the spectrum, with the values diverging outside the band. (The shelving seen below 100 Hz in the LMS plot is a direct result of LMS system impedance measurement limit set to 1 kHz). Not bad for a product not specified for capacitor measurement!