Notes from the Test Bench
By Bruce Hofer, Chairman & Co-Founder, Audio Precision
As you’ll see below, we’re starting the New Year with a change in our product line up.
Effective March 1, we will discontinue production of the ATS-2.
We introduced this model in 2001, and it has served the market well for the last decade. However, in recent years, sourcing parts has become increasingly challenging. On several previous occasions, we were able to reengineer a solution, but now lack of availability of a particular single-source component means we can no longer keep the ATS-2 in production.
This is all part of the natural product life cycle, and we consciously designed the APx525 and APx515 to be functional replacements for the ATS-2. Indeed, APx offers better analog performance, better digital connectivity, a better UI, and better automation—all at a lower price. From our experience migrating customers from System One to APx, we’ve developed protocols to ease the transition, and AP’s Tech Support department is here to help you design and implement a good migration plan.
I believe 2011 holds a lot of opportunity for the audio industry, and we’re looking forward to helping make it as successful as possible for everyone.
ATS-2 Notice of Discontinuation
On March 1, 2011, Audio Precision will discontinue production of the ATS-2 audio analyzer. As with all AP analyzers, the ATS-2 will remain fully supported by Tech Support and Service for a minimum of five years, starting March 1.
Comparison of APx515, APx525, and ATS-2
Output: Windows Logo Audio Fidelity Testing with APx
The new Technote 112 and its associated APx projects are ideal for research and development, production verification, and confidence testing for compliance with the Microsoft Windows Logo Program. The article below summarizes and introduces the Technote. The full 20 page document goes into much greater detail, with instructions on how to set up the test system, and explanations of all the tests and measurements.
Microsoft requires all PCs and devices running Windows to pass hardware and software tests before they may display the Windows logo. One subset of these requirements is the Audio Fidelity tests on the analog inputs and outputs.
Microsoft’s Windows Logo Kit (WLK) includes their Device Test Manager (DTM) software to test for compliance. The DTM software requires a complex networked system consisting of a DTM server running Windows Server 2003 or later, a DTM client connected to an Audio Precision 2700 Series audio analyzer (using AP2700 control software), and the SUT (System Under Test) configured as a second DTM client. This is the system that must be used to produce the test reports that are submitted to Microsoft to obtain Windows Logo certification. It requires special training to operate, and takes time to set up. Additionally, the results are only output as a complex file for submission.
The new APx projects that go with this Technote run all the same tests (with the same pass/fail limits) as the DTM, but they take advantage of the advances present in the APx platform. Compared to the DTM/AP2700 system, the APx implementation:
A device that passes the APx Windows Logo Audio Fidelity Tests has been tested to the same Windows Logo standards as with the DTM tests. However, at this time, Microsoft does not yet support the APx500 Series as part of the DTM system. Therefore, to produce the final report for submission to Microsoft, the DTM system with an AP 2700 Series analyzer will still need to be used.
This Technote includes a Desktop PC Audio Fidelity Test project, and a Mobile PC Audio Fidelity Test project. The two APx projects perform identical tests on output jacks (called "render devices" in the Microsoft DTM) and input jacks (called "capture devices"), but the acceptance limits differ. A mobile PC could be a notebook, a netbook, or even a mobile phone with wireless networking features. Both project files include signal paths for every possible type of analog output and input.
Desktop PC Audio Fidelity Test project.
Additional projects are also supplied. Two of these are for determining the full scale input levels, which must be done before running the Audio Fidelity Tests.
Find Full Scale Input Voltage project.
Multiple graphs are created by the Find Full Scale Input Voltage projects, to help you in analyzing and determining the input overload level. The figure below illustrates the clipping behavior of a microphone input as measured through a headphone output. The THD+N Ratio of this headphone output at full scale output voltage was -78 dB, well below clipping—therefore, the clipping behavior shown here is occurring in the microphone input and not the headphone output. The full scale input voltage that produced a -40 dB THD+N Ratio was 1.22 Vrms on the right channel.
Microphone input clipping: Output THD+N vs. input voltage.
The last project is the Real-time PC Audio Fidelity Test. The real-time measurements are not part of the Microsoft DTM, but they provide additional diagnostic capabilities. In these tests, measurements of the analog outputs and inputs are completely isolated from each other, since the analog inputs and outputs are not utilized simultaneously.
Signal path for real-time render device tests.
Read the full Technote for detailed instructions on setting up and running the tests, as well as explanations of all
the tests and measurements. The download package includes all the APx projects, as well as the waveforms necessary for
the playback tests.
Sound Advice: Measuring Loudspeaker Impedance
Measuring a loudspeaker’s complex impedance can easily be done directly in APx500 using derived measurements (available in versions 2.6 or later). This article, by AP's Director of Technical Support Joe Begin, describes how to do so using the Constant Voltage method and includes an accompanying APx project.
Measuring loudspeaker impedance by the Constant Voltage method means that the voltage across the loudspeaker terminals will stay relatively constant across the frequency range, regardless of speaker impedance fluctuations. This allows testing to be done at a specific speaker voltage level, and by running the test multiple times at different levels, you can observe if the impedance curve is level dependent.
Figure 1. Schematic of loudspeaker impedance test.
Figure 1 shows a basic schematic of the Constant Voltage method, while Figure 2 shows the actual connections. Two analyzer inputs, configured as analog balanced, are used to make the measurements: One channel measures voltage across the sensing resistor (Vsense), and the other measures voltage across the speaker (Vspkr). The power amplifier provides the ultra-low output impedance and high current necessary to drive low impedance speaker loads. While the analyzer can be used to drive the speaker directly, some combinations of output voltage and speaker impedance may exceed its maximum output current. In addition, the minimum 50 Ω output impedance of the analyzer feeding a speaker creates a current source, defeating the main advantage of the Constant Voltage method.
Figure 2. Connections.
To make the connections easier, we constructed a measurement test jig that contains a 0.1 Ω 50 W current sense resistor (Figure 3). This jig has dual banana binding posts on each end, and an XLR connector on the side, which makes for easy connections to the audio analyzer, the speaker, and the amplifier. The jig shown is one-of-a-kind, but you can easily construct one yourself.
Figure 3. Current sense test jig.
The resistance of the sense resistor is not critical, but it should have reasonable precision (say 1%) and a sufficient power rating. A value of 0.1 Ω is a good choice, because it does not significantly compromise the low output impedance of the constant-voltage source, and a division by 0.1 (or multiplication by 10) is convenient.
Four-terminal current sense resistors are available for use in a Kelvin configuration, in which the current is supplied through two opposing terminals and the sensing voltage is measured across the other two terminals. In addition to offering more convenient connections, the 4-wire Kelvin configuration provides more accurate sensing measurements. Examples include the Ohmite 10 series and the Riedon UAL series.
Based on the circuit in Figure 1, the current (i) is derived from the equation
and the impedance can be derived from
Combining the above two equations yields
In the equations above, the bar above i, V, and Z denotes that they are phasor quantities (i.e., they have both a magnitude and phase).
Figure 4. The Level measurement (primary) result.
In the accompanying project file "LoudspeakerImpedance.approjx", the impedance is derived from the Level result in the Acoustic Response measurement (Figure 4). In practice, any of the frequency response measurements could be used—Acoustic Response was chosen because it is fast and allows multiple averages to be taken. Note that because this is not really an acoustic measurement, the Time Window setting on the Impulse Response and Energy Time Curve measurements is not relevant. It should therefore be set to the same value as the sweep time. In this project file, a sweep time of 1.0 seconds has been set. As shown, the input channels have been re-labeled in Signal Path Setup to Vspkr and Vsense, for convenience.
Figure 5. Intermediate result for impedance magnitude.
To get the impedance magnitude (Figure 6), we first use an intermediate derived result named Z Mag – intermediate (Ch1/Ch2) (Figure 5). This is a Compare derived result, and simply divides Ch1 by Ch2 (Vspkr/Vsense). Next, we use an Offset derived result to multiply the intermediate derived result by Rsense. In this case, we need to multiply the result by 0.1 Ω. A multiplication by 0.1 is equivalent to an offset of -20 dB. Hence a gain of -20 dB was applied to the offset operation as shown in Figure 6. If other values of Rsense are used, the gain for this offset operation should be set according to the equation
Figure 6. Impedance magnitude (derived).
Finally, to get the impedance phase (Figure 7), we used an Invert derived result on the primary Phase result. This is to correct for the fact that the primary phase result is the phase of channel 2 relative to channel 1, and in this case we want the phase of the speaker voltage (channel 1) relative to the current (channel 2). Alternatively, we could have just swapped the two measurement channels and this invert operation would not have been required.
Figure 7. Phase magnitude (derived).
Test Results: AP News & Events
©2011 Audio Precision Inc.