AP - The Recognized Standard In Audio Test

By Bruce Hofer, Chairman & Co-Founder, Audio Precision

It’s great to get out on the road and meet fellow engineers who share a passion for audio. I’m in Japan this week, where I’m giving a talk arranged by our friends at TOYO.

I’ll be looking in some detail at the possibilities available with the combination of our new BW52 ultra-high bandwidth option and our AG52 square wave generator for APx. I can’t put my entire slide set into Audio.TST, but I have asked one of AP’s technical writers to pull some of the theoretical content and cover a couple of the more common applications. If there’s interest, we’ll add other parts of my presentation into the knowledge base as time permits.

Bruce

Last month we looked at the new BW52 ultra-high bandwidth option for APx, which extends FFT bandwidth past 1 MHz with 24-bit A/D resolution. Combined with the recently released AG52 option, it turns the APx525 Family into an extremely accurate square wave generation and analysis system. This month we take a closer look at what the AG52/BW52 combination can do.

Square waves are useful because they can reveal many things about the frequency and phase response of a system with just visual analysis. This can be valuable when doing a quick system check, or when doing an A/B comparison while changing settings or components. The following graphs highlight just some of the things that a square wave can show:

What’s a theoretically perfect square wave made of? It contains a sine wave fundamental, and all its odd harmonics. The amplitude of each harmonic is 1/n, so the amplitude of the fifth harmonic, for example, would be 1/5 the amplitude of the fundamental. But making a perfect square wave isn’t easy.

Let’s look at a frequency domain view of a square wave produced by a high quality 15 MHz function generator. This function generator is a good choice for many applications, but not for audio, where the extremely high dynamic range (about 144 dB for 24-bit) requires the purest signal possible.

Most engineers are used to looking at square waves only in the time domain, but the frequency domain view provides additional diagnostic ability about the quality of a square wave, both before and after it has passed through a DUT (device under test).

Square wave created by a function generator, frequency domain view.

Notice how smoothly the odd harmonics diminish. That’s good. But also notice how high the even harmonics are—only about 60dB below the fundamental. A perfect square wave would have no even harmonics. At 1 MHz, the even harmonics are only about 12 dB below the desirable odd harmonics, which means that real information about the DUT may easily be obscured by distortion in the square wave test signal. Notice also that there are intermodulation products 90 dB or so below the fundamental, again potentially hiding problems in the DUT.

Here’s a square wave generated by AP’s AG52 option:

Square wave created by an APx525 analyzer with the AG52 option, frequency domain.

This is the best square wave produced by any audio analyzer in the world.

The even harmonics are now 110 dB to 120 dB below the fundamental. At 1 MHz, the even harmonics are still 50 dB below the desirable odd harmonics. Intermodulation products and noise are now down at least 130 dB below the fundamental. All of which give us much greater clarity to detect and analyze defects caused by the DUT.

The BW52 complements the high performance of the AG52 square wave. Its 1 MHz bandwidth keeps the square wave perfectly square, so that we can be sure any defects seen are in the DUT, and aren’t artifacts of the analyzer itself. The 1 million point FFT and 24-bit A/D conversion allow extremely detailed analysis unobscured by noise.

Now, let’s use the AG52/BW52 combination to look at the time and frequency domain displays of a signal passing through a popular home theatre surround sound amplifier.

Square wave after passing through a home theatre receiver, time domain.

That is not a good looking square wave!

Square wave after passing through a home theatre receiver, frequency domain.

A view of the frequency domain graph reveals what’s going on inside the DUT—there’s a sharp cut-off in the response near the Nyquist frequency of the on-board DSP, and then a sharp rise in noise caused by the sigma-delta type converters.

Let’s look at another audio test where a high purity square wave is required for accurate results. DIM (Dynamic Intermodulation Distortion) uses a combined square wave / sine wave stimulus to reveal slew-induced distortion—distortion caused when an amplifier can not increase or decrease its output voltage fast enough to follow its input. Rarely a problem in modern op amps, DIM can still be an issue in the design of high wattage power amplifiers due to the large voltage swings they must produce.

DIM Level Sweep measurement (ratio vs. measured level).

Many audio analyzers generate a square wave with inadequate bandwidth and visible ringing. Coupled with an FFT analyzer limited in bandwidth and resolution, they can only perform the most rudimentary square wave analysis. In this article we’ve tried to show the powerful analysis ability made possible using a high quality square wave generator with a wide bandwidth, high resolution, and long sample length FFT.

AP’s Director of Technical Support, Joe Begin, has written yet another utility to expand the range of APx measurement capability.

Any Audio Precision analyzer may be used to measure speaker impedance. However, since the analyzers measure and display in voltage, a utility is needed to convert the readings into ohms. In the following article, we cover the basic theory and test setup, and then present the APx Impedance Measurement Utility for APx500 Series instruments. For users of the 2700 Series/System Two family instruments, a tutorial and macro is available for measuring complex impedance. A macro is also available for the ATS-2.

Figure 1 shows a simple schematic of the speaker measurement circuit, which is essentially a voltage divider. In the Figure, V_gen denotes the generator voltage, R_s is the source resistance, and Z_x is the impedance of the device under test.

Figure 1. Impedance measurement circuit.

In general, Z_x will have both resistive and reactive components, and is therefore a complex quantity (i.e., it has both magnitude and phase). However, if R_s is much larger than the magnitude of Z_x (denoted by |Z_x|), then the magnitude of Z_x can be determined using equation (1).

    .................................................. (1)

As |Z_x| becomes large relative to R_s, the denominator term in equation 1 (V_gen - V_m) becomes smaller, and eventually the results become inaccurate. As a rule of thumb, R_s should exceed |Z_x| by at least a factor of 10 for equation (1) to be valid. For measuring a typical 4 or 8 Ω (nominal) loudspeaker, a 600 Ω source impedance is ideal.

The APx525 family has a selectable 600 Ω source impedance built in, and a loopback mode that internally connects the generator outputs to the analog inputs. Cable resistance between the generator and the speaker should be as low as possible so that it does not significantly affect the results. For the APx525 family, you can best accomplish this by connecting heavy gage speaker wire to the analog balanced output banana jacks.

Figure 2. Signal Path for the APx525 family using Loopback mode.

Connections to the APx585 family instruments are a little more complicated. First, there’s no internal loopback mode, so a cable is needed to connect the measurement point to the analog input. Second, the impedance is fixed at 50 Ω unbalanced and 100 Ω balanced. This is too low to meet the requirement that R_s should exceed |Z_x| by at least a factor of 10. To overcome this limitation, an external resistance needs to be added in series with the internal source resistance. We suggest using a Pomona Electronics #1469 adapter (BNC male to binding posts with isolated solder turrets) on the analog unbalanced output. Mount a 549 Ω 1.0% resistor across the turrets on the ungrounded side of the adapter, and short the turrets on the other side. The total of the internal and external resistances will now equal 599 Ω ± 1 %. From the binding post end of the adapter, heavy gage speaker wire is run to the speaker, and a standard BNC cable, stripped at one end, is run back to the analog unbalanced input.

Figure 3. Signal Path for the APx585 family using an external resistor in series with R_s.

The APx Impedance Measurement Utility and associated project file allow you to measure impedance vs. frequency of a speaker and see the results directly in ohms. The utility was developed in LabVIEW 2009 and uses the APx LabVIEW driver to communicate with the APx500 measurement software. The download includes the source code for those who have the LabVIEW development environment, as well as the compiled application for those who don’t.

Figure 4. The APx500 Impedance Measurement Utility.

When the utility is started, it opens the APx500 measurement software if it isn't already running. It first checks to see what model of APx analyzer is attached, and then configures the analyzer and itself accordingly as shown below.

The utility is designed to be used with the accompanying project file, impedance.approj. To load the file, click Load Project File. The file browser will default to the location where impedance.approj has been installed, so just click OK. Settings in the impedance.approj project file are configured to produce accurate results. If you wish to modify the project or create your own, please observe the following guidelines:

The utility checks all the signal paths and measurements in the currently loaded APx project and sets the Selected Signal Path control to the first signal path in the project. It also sets the Selected Measurement to the first frequency response type of measurement in the signal path (if one exists).

Frequency Response is an ideal measurement to use for an impedance versus frequency measurement. It runs very quickly and has very high frequency resolution. To increase the frequency resolution of the measurement, you can increase the length of the sweep using the control in the Advanced Settings panel of the Frequency Response measurement in the APx500 software.

Continuous Sweep and Acoustic Response provide the same results as Frequency Response, except that Acoustic Response offers the option of averaging several measurements, if needed. Note that if the Acoustic Response measurement is used, the time window control should be set to the same value as the sweep time (effectively removing the time window). Otherwise, there will be large errors in the results.

To run an impedance measurement, select the signal path and measurement you want to use from the utility’s front panel and click the Start button. After the measurement is completed in APx, the utility calculates the impedance and displays it on the graph (Figure 5). One cursor is automatically placed at the point of maximum impedance, and another is one at the point of minimum impedance.

Figure 5. Measured impedance curve for a loudspeaker with multiple drivers.

The Print Results button opens a new window that allows you to print the graph image to the PC’s default printer, or to save an image of the graph to a file. This window also has a notes field where you can add details about the test or DUT.

Finally, the Export Data button will export the data as a tab-delimited text file, for import into a spreadsheet program like Microsoft Excel.

The graph provides many controls to customize it as desired. Because this utility is written in LabVIEW, the controls differ from those in the APx500 software.

All graphs and charts, except for the 3D graphs, automatically enable autoscaling, which means they adjust their horizontal and vertical scales to fit the data. By default, autoscaling is enabled for graphs and charts. Right-click the graph or chart and select Autoscale X or Autoscale Y from the context menu to turn autoscaling on or off. To change the upper and lower limits of a graph’s X and Y scale, double-click on the text at the limit of the axis, type the new limit in its place, and then press the Enter key. You should turn autoscaling off when setting the graph axis ranges manually. Otherwise, they will revert to autoscaled values the next time data is displayed on the graph.

You may control cursor appearance and add additional cursors by right clicking inside of the cursor box. If you add cursors, select Visible Items | Vertical Scrollbar so that you can access the entire list.

Figure 6. Graphing palette.

The graphing palette allows you to move the cursors, zoom, and pan. Each button lights in green when you enable it.

• APx Impedance Measurement Utility (102 MB)

• Measuring Sample Rate Error When the Sample Clock is Not Accessible
• Running AP Software on Windows 7

back to top >