Silicon ChipMeasuring Audio Gear Without Spending Big Dollars - October 2011 SILICON CHIP
  1. Outer Front Cover
  2. Contents
  3. Publisher's Letter: DAB+ radio broadcasting has a long way to go
  4. Project: Let’s Get Quizzical: A Quiz Game With A Difference by John Clarke
  5. Project: Digital LED Lighting Controller For Christmas Lights by Nicholas Vinen
  6. Project: Build A Loudspeaker Protector by Silicon Chip
  7. Project: Measuring Audio Gear Without Spending Big Dollars by Jim Rowe
  8. Project: The USB MIDI-Mate: A MIDI Interface For PCs by Jim Rowe
  9. Review: CadSoft EAGLE PCB Layout Software by Nicholas Vinen
  10. Project: High-Performance Stereo Headphone Amplifier, Pt.2 by Nicholas Vinen
  11. Book Store
  12. Vintage Radio: A look at the Radiola Model 573-MA Receiver by Maurie Findlay
  13. Market Centre
  14. Advertising Index
  15. Outer Back Cover

This is only a preview of the October 2011 issue of Silicon Chip.

You can view 28 of the 104 pages in the full issue, including the advertisments.

For full access, purchase the issue for $10.00 or subscribe for access to the latest issues.

Items relevant to "Let’s Get Quizzical: A Quiz Game With A Difference":
  • Quizzical Quiz Game PCB [08110111] (AUD $25.00)
  • PIC16F88-I/P programmed for the Quizzical [0811011A.HEX] (Programmed Microcontroller, AUD $15.00)
  • Firmware (ASM and HEX) files for the Quizzical! project [0811011A] (Software, Free)
  • Quizzical Quiz Game PCB pattern (PDF download) [08110112] (Free)
  • Quizzical! Quiz Game front panel artwork (PDF download) (Free)
Items relevant to "Digital LED Lighting Controller For Christmas Lights":
  • Digital Lighting Controller LED Slave PCB [16110111] (AUD $20.00)
  • Digital Lighting Controller LED Slave PCB pattern (PDF download) [16110111] (Free)
Items relevant to "Build A Loudspeaker Protector":
  • Speaker Protection and Muting Module PCB [01207071] (AUD $17.50)
  • Speaker Protector and Muting Module PCB pattern (PDF download) [01207071] (Free)
Items relevant to "The USB MIDI-Mate: A MIDI Interface For PCs":
  • USB MIDI-Mate PCB [23110111] (AUD $15.00)
  • PIC18F14K50-I/P programmed for USB MIDI-Mate [2311011C.HEX] (Programmed Microcontroller, AUD $15.00)
  • Firmware (HEX and C source code) files for the USB MIDI-Mate project [2311011C] (Software, Free)
  • USB MIDI-Mate PCB pattern (PDF download) [23110111] (Free)
  • USB MIDI-Mate panel artwork and drilling templates (PDF download) (Free)
Items relevant to "High-Performance Stereo Headphone Amplifier, Pt.2":
  • Hifi Stereo Headphone Amplifier PCB [01309111] (AUD $17.50)
  • Red & White PCB-mounting RCA sockets (Component, AUD $4.00)
  • Hifi Stereo Headphone Amplifier PCB pattern (PDF download) [01309111] (Free)
  • Hifi Stereo Headphone Amplifier front & rear panel artwork (PDF download) (Free)
Articles in this series:
  • High-Performance Stereo Headphone Amplifier, Pt.1 (September 2011)
  • High-Performance Stereo Headphone Amplifier, Pt.1 (September 2011)
  • High-Performance Stereo Headphone Amplifier, Pt.2 (October 2011)
  • High-Performance Stereo Headphone Amplifier, Pt.2 (October 2011)

Purchase a printed copy of this issue for $10.00.

You can do it with your PC . . . . By JIM ROWE Measuring Audio Gear – without spending $$$$! So you want to measure the performance an amplifier or other piece of audio gear. . . but you don’t have thousands of dollars for the “right” test gear. No problem – all you need is a PC, a decent sound card (or a USB sound interface), plus an appropriate software package. Here’s how it’s done: I analog domain will inevitably depend very much on the n the old days, checking the performance of audio circuitry inside the PC’s sound card or USB sound interface equipment like amplifiers and preamps usually in– in particular, on the ADCs in the audio input circuitry volved a fair bit of test gear: an audio generator, an and on the DACs in the audio output circuitry. audio millivoltmeter (or better still a distortion and noise The analyser software can automatically correct for meter), a scope to keep an eye on clipping and hum and things like sound card frequency response variations and some resistive ‘dummy loads’ to provide the audio equipgain or loss in both the input and output analog circuitry ment with its correct loading (standing in for things like but it can’t really compensate for high noise speakers, with their complex impedances). and/or distortion levels in And even when you had all of this test gear on this circuitry. hand, the actual testing was a rather The same qualifications tedious and time-consuming tend to apply in the case of operation. crosstalk inside the sound Now, thanks to galcard/interface – both interloping digital technolchannel crosstalk between ogy, you do can do your the right and left channels own tests at much lower (within both the input ADCs cost, using a recent-model and the output DACs) and also PC with a decent full-duplex crosstalk directly between the sound card (or a USB ReADCs and DACs. cording/Playback interface) With an ‘el cheapo’ sound plus a low cost digital audio card (such as that included analyser software package. in budget PCs or integrated Before we start telling you with their motherboards) the how it’s done, we don’t want results are likely to be fairly to mislead you about the modest, while with a ‘top of kind of measurement results the range’ sound card or interyou can expect. face they may well approach Although the performance If your PC of the digital analyser softsound card is not up to it, what could be achieved usware tends to be excellent the SILICON CHIP USB Recording/ ing a dedicated digital audio in the digital domain (ie, Replay Interface (published June 2011) analyser system. So to ensure the best posinside the PC itself), the would be an ideal partner for any PC-based performance ‘outside’ in the audio equipment measurement setup. sible results use the highest 62  Silicon Chip siliconchip.com.au performance sound card or interface that you can afford. What you’ll need I’d recommend as a minimum using a Pentium 4 system (or better) running at 1.5GHz or better, with at least 512MB of RAM and a 80GB or larger hard drive. It should also be running either Windows XP, Vista or Windows 7 – which will again increase the demands for RAM and hard drive capacity. Of course if the PC is already fitted with a top-quality sound card, so much the better. If it isn’t, your best bet would be to buy or build a good USB sound interface, such as that described in the June 2011 issue of SILICON CHIP. Software You will also need a digital audio analyser software package, as mentioned earlier. There are a few of these around but the one that seems to have the best reputation is a package called ‘TrueRTA’, written by John L. Murphy, a former space systems software analyst for the US Air Force and an audio design engineer with over 20 years’ experience. Details of this software are summarised in the panel at right. After trying out the free Level 1 version myself for a few days, I decided to upgrade to Level 4 and the next day I was able to use this after entering the registration code. I’ll be using the Level 4 version of TrueRTA to illustrate each aspect of doing audio testing with a PC throughout this article. During the preparation of this article I’ve used TrueRTA Level 4 with a number of PC’s as well as sound cards and USB interfaces. The PCs include a Compaq V2000 laptop running Windows XP, a Compaq D380mx desktop machine also running Windows XP and an Acer Aspire AX1800 machine running Windows 7 (64-bit version). The SILICON CHIP USB Recording/Playback interface was used with the Compaq V2000 and the Acer AX1800, while I used a Creative Extigy USB interface with the Compaq LAPTOP OR DESKTOP PC RUNNING AN AUDIO ANALYSER PACKAGE SUCH AS "TrueRTA" USB CABLE LEFT LINE OUT RIGHT LINE OUT LEFT LINE IN RIGHT LINE IN USB RECORD/REPLAY INTERFACE TO INPUTS OF GEAR BEING TESTED FROM OUTPUTS OF GEAR BEING TESTED Fig.1: The basic setup for PC-based audio testing. The audio line outputs provide the test signals, while the line inputs connect to the meter, scope and analyser. siliconchip.com.au About TrueRTA is a software package available online from TrueAudio (USA). (www.trueaudio.com). A limited capability version (Level 1) is available at no cost (ideal to try out and has no expiry date). The file TrueRTA_se.exe is less than 2MB in size and is a self extracting installer. This can be upgraded to any of the three higher performance levels by buying the appropriate licence (no further download is needed). The Level 2 licence is $US39.95, Level 3 is $US69.96 while Level 4, the highest performance level, is $US99.95 TrueRTA runs on any PC with a Pentium 3 or better, running at 500MHz or faster and with at least 64MB of RAM plus a fullduplex sound card or USB audio interface, running Windows XP, Vista or Win7 (32-bit or 64-bit). It is basically a suite of software-based audio test instruments, including: 1. A high resolution real-time analyser (spectrum analyser). The analyser resolution varies from one octave/band for the free Level 1 version, to 1/24th of an octave/band for the Level 4 version. The Level 4 version also provides selectable resolutions of 1/3, 1/6 and 1/12th of an octave. The analyser can display in either conventional bar mode or line mode, which is better for showing frequency response curves, etc. The maximum upper dBu limit is +20dBu, while the minimum lower dBu limit is -160dBu. These are the measurement limits of the software, of course; the performance of your sound card/interface will determine the actual measurement range. Other features of the analyser include selectable averaging, adjustable frequency range and the ability to store calibration curves for not only the PC’s sound card/interface but also for a microphone so that it can be used for acoustic measurements on speakers etc. Once stored these curves can be automatically used to correct for system errors and improve measurement accuracy. 2. A low distortion digital audio signal generator, the output of which can be varied between 5Hz and 48kHz (when the sound card allows 96kHz sampling). In addition to the low distortion sinewave output the generator can be set to produce square waves (adjustable duty cycle), triangular, sawtooth and impulse waveforms, as well as pink noise and white noise. The higher level versions can also produce a digitally synthesised logarithmic sine sweep from 10Hz to 48kHz (or half the sampling rate), with a response flat within +/-0.05dB over the audible frequency range. 3. A dual trace DSO which covers the full audio spectrum. The amplitude range of each channel can be varied between 5V/ division down to 1mV/division in the usual sequence of ranges, while the timebase ranges vary from 200ms/div to 50s/div. Triggering is selectable in terms of the left or right channel and also adjustable in level. 4. A digital audio voltmeter which in the Level 4 version provides readout of the input level in millivolts RMS and dBu, together with the crest factor in either mV/mV or dB. Each of these four measurements can be selected individually, or all at the same time. Each of the instruments can be started or stopped individually and when stopped the currently displayed measurement screen and all associated data can be saved to memory (up to 20 memories are available) and/or printed out (with or without user-added notes). October 2011  63 Fig.2: clicking on TrueRTA’s Audio I/O menu button gives you this drop-down menu, which is used for most of the initial setup and calibration steps. Fig.3. if you click on Audio Device Selection in the Audio I/O menu, this dialog box appears to let you select which audio input and output devices are to be used. D380mx machine for comparison. Fig.1 shows the basic set-up for using a PC for audio equipment testing. The laptop or desktop PC is running an audio analyser package like TrueRTA, while the analog outputs and inputs of either the internal sound card or the USB sound interface are used as the system’s interfaces to the gear being tested. The line outputs are used to provide the test signals (from the software audio generator), while the line inputs are used to feed the output signals from the equipment you’re testing back into the PC for analysis. You can see from Fig.1 that in order to use the PC and its USB sound interface for audio testing, the operating system (ie, Windows) must be set up not only to use the USB interface as its ‘default’ sound card but also to configure it so that the ‘recording’ signals being fed in via the line inputs are NOT ‘looped back’ internally by the software to the line outputs (this is often done to allow ‘record monitoring’). So as well as going into the Windows Control Panel and making sure that your USB Record/Playback interface is set as the default audio device for both recording and playback, it’s also quite important to go into the Windows Playback Mixer dialog and make sure that the Line Inputs are not selected for playback – only the WAVE signals. This is usually the best way to ensure that the input (‘recording’) and output (‘playback’) functions are kept isolated from each other. Another important step in your initial system set-up is to ensure that the recording and playback volume controls are each set to a known and easily repeatable level. Usually with Windows XP and earlier operating systems this is ‘maximum’ – ie, with the software sliders set at their upper limit. However with Windows 7, you need to go into Control Panel/Sound/Recording and then select the Microphone function of your USB Audio Codec and open its Properties dialog. Fig.4: Then if you click on Input Channel Selection, a submenu lets you select which channel or channels are to be displayed. Fig.5: The Input Sampling Frequency sub-menu lets you to chose the highest input sampling frequency that your sound interface can handle. The basic set-up 64  Silicon Chip siliconchip.com.au Fig.6: the Output Sampling Frequency flyout lets you match the output sampling frequency to the one you selected for the inputs in Fig.5. Then you need to click on the Levels tab and move the Microphone volume slider almost fully to the left, until numeral ‘1’ is being displayed in the box just to the right of the slider itself. This sets the ‘recording’ gain to unity, corresponding to a ‘line level’ input. Initial checkout You should now be ready to check that your audio testing software and hardware system is set up and functioning correctly. Do this by first starting up TrueRTA and then clicking on the Audio I/O menu button (in the top menu bar). This causes a drop-down menu to be displayed, as shown in Fig.2. If you then click on Audio Device Selection, you’ll see a Fig.8 (left): TrueRTA’s signal generator control panel runs down the left-hand side of the screen and provides all of its controls. Fig.9 (right): In scope mode, the DSO control panel is displayed down the right-hand side of the screen. siliconchip.com.au Fig.7: TrueRTA also lets you select the CPU Speed while it’s running. Initially this is best left set to ‘1 (safest)’. column of Input Device Selection choices on the left and a similar column of Output Device choices on the right (Fig.3). USB Audio CODEC should be selected. Then click on [OK] to close the dialog box. Now click on the Audio I/O menu button again and this time click on Input Channel Selection. This will cause a sub-menu to appear to the right (Fig.4) and you should see a small tick next to the top selection ‘Display L Channel’. Since the Audio I/O menu remains displayed, the next step is to click on the Input Sampling Frequency option just below Input Channel Selection. This will cause the first flyout menu to disappear, to be replaced by another giving a list of input sampling frequency options (Fig.5). If you’re using the SILICON CHIP USB Recording/Replay Interface, you can select the 48kHz sampling frequency. If you’re using an interface or sound card which can support 96kHz sampling, this can be selected instead. As a rule of thumb it’s a good idea to select the highest sampling frequency that your audio interface is capable of working at, because this will give the best measurement resolution. Then click on the next menu option, for Output Sampling Frequency, which gives you another sub-menu (Fig.6). This only gives you three options, so select the one which corresponds to the Input Sampling Frequency you chose in the previous step. The next step is to click on the following option in the Audio I/O menu: CPU Speed Setting. This gives you a menu offering a choice of five speed settings (Fig.7) but it’s recommended that during initial set-up you leave it at the default setting of ‘1 (safest)’. Later on when you are happy that everything is working OK, you can try one of the faster settings. Your system should now be ready for its initial functional test, so link the Left Line Output of your USB Interface back to its Left Line Input, using a short screened cable with an RCA plug at each end. Then turn your attention to TrueRTA’s Generator control panel, a narrow box running down the October 2011  65 LAPTOP OR DESKTOP PC RUNNING AN AUDIO ANALYSER PACKAGE SUCH AS "TrueRTA" DIGITAL RMS AC VOLTMETER 244.9 v 399.0mV 3.999V 39.99V 399.0V USB CABLE LEFT LINE OUT RIGHT LINE OUT LEFT LINE IN RIGHT LINE IN USB RECORD/REPLAY INTERFACE Fig.10: Here’s the TruRTA DSO displaying the generator’s output signal, with the generator producing a 300Hz sinewave at around -10dBu. Fig.11: The setup for calibrating the input and output circuitry of your sound card or interface, using an external RMS AC voltmeter. left-hand side of the screen. As you can see from Fig.8, this not only offers a Generator On/Off button at the top but small text boxes lower down which can be used to set the software generator’s output frequency and output amplitude (in dBu). Then there’s a set of buttons providing a choice of waveforms and finally a button labelled ‘Quick Sweep’. We’ll come back to this button later but for the present just make sure that the generator’s frequency is set to 300Hz and its output level to -10dBu. Now click on the oscilloscope tool button at top left on the screen, just below the File menu – the button with the sinewave on it. Then click on the fourth button along in the same toolbar - the one with the label ‘Go’. This will cause the main window to become TrueRTA’s scope display, with its own control panel running down the right-hand side of the screen. As you can see from Fig.9 this gives you two columns of buttons to allow you to set the DSO’s vertical gain and timebase scaling, plus a pair of buttons to select either the left or right channels as the timebase triggering source, a slider to adjust the triggering level and a button to select or deselect auto triggering. For the present, leave the vertical gain and timebase settings at their default values, which should be 0.2V/Div and 0.5ms/Div (as shown in Fig.9). The L channel should also be selected as the triggering source. At this stage your DSO display should only be showing a horizontal line, because the software generator hasn’t been turned on as yet. As soon as you click on the On/Off button at the top of the left-hand Generator panel (Fig.8), you should see the generator’s 300Hz sine waveform appear on the display within a second or two (Fig.10). If everything has gone well so far, you’ll now be ready for Fig.13 (above): The Spectrum Analyser menu allows you to set the analyser’s resolution, its speed/precision tradeoff and measurement limits. Fig.12: Clicking on Line Input Calibration in the Audio I/O menu (Fig.2) displays this dialog, which is self explanatory. 66  Silicon Chip Fig.14 (right): When the analyser is running, its control panel appears at the right-hand side of the screen. siliconchip.com.au Fig.15: After the initial calibration steps, TrueRTA can measure the overall frequency response of your PC and its sound card or interface. the next set-up stage: calibrating your system so its audio measurements will be accurate. By the way, this calibration is necessary because every PC sound card or USB interface tends to have a different sensitivity for its line level inputs and a different line output level. Calibrating your system Fig.16: After getting the overall response, the next step is to calibrate the system via this dialog. TrueRTA then produces and saves a frequency calibration curve. To perform the calibration you’re going to need a calibrated AC voltmeter or millivoltmeter, able to give accurate readings of audio levels around 244.62mV RMS (corresponding to -10dBu, since 0dBu = 0.775V RMS). If you don’t have access to an AC millivoltmeter as such, you could use a DMM set to its lowest AC voltage range (many modern DMMs do measure RMS voltage on the AC volts ranges). The physical set-up for calibration is shown in Fig.11 and again involves the USB Interface’s left line output being looped around and fed into the left line input – but this time with the external AC voltmeter monitoring the actual audio level. The first step is to calibrate the sound card/interface’s Line Input circuitry and this is done by first clicking on the Audio I/O menu button to display that menu again (Fig.2). Then run down that menu and click on the item label ‘Line Input Calibration...’ This will open up the dialog box shown in Fig.12, which as you can see gives you a complete summary of the steps involved in this calibration procedure. The basic idea is that you first set the software generator to produce a 300Hz sinewave at a nominal level of -10dBu. Then you turn the generator on (using the On/Off button in Fig.8) and note carefully the reading on your DMM or AC millivoltmeter. It should be somewhere in the vicinity of 245mV or 0.245V but the exact level will depend of course on the line output circuitry in your PC sound card or interface. Whatever the meter reads, all you have to do is type that voltage value into the text box provided in TrueRTA’s opened Line Input Calibration dialog box. Then simply click on the [OK] button at the bottom of the box and TrueRTA will save the line input calibration value so that it will be used in future. Once you have calibrated the line input circuitry in this way, calibrate the sound card/interface’s line output circuitry as well, so the generator’s output level will accurately reflect the setting shown in the on-screen control panel. Line output calibration is done in a very similar way to the line input calibration and using exactly the same physical set-up (Fig.11). Fig.17: After system calibration, a response plot of the system itself is now virtually ‘flat’, even when the vertical scaling is expanded. Fig.18: Here’s the ‘noise floor’ plot of the left input channel of our USB Recording/Playback interface. It varies from -120dBu at 15Hz to -90dBu at 20kHz. siliconchip.com.au October 2011  67 Fig.19: A spectrum analyser plot of the interface when the generator is producing a 1kHz sinewave at -10dBu. Harmonics are visible up to 7kHz. Fig.20: You can get a clearer picture of the sound system’s distortion products by subtracting the system noise, as shown here. The only difference is that instead of selecting ‘Line Input Calibration...’ in the Audio I/O menu, you select ‘Line Out Calibration...’. This again opens up a very similar dialog box to that shown in Fig.12, giving you a summary of the steps in this procedure. After you have calibrated both the line input and line output circuitry in this way, the next main calibration step is to get TrueRTA to measure the overall frequency response of both the input and output circuitry, so it can save a correction curve for your PC sound system. Once this is done, any deviations from a flat response in the system itself can be automatically compensated by the software, so your testing of audio equipment in the future will be as accurate as if the PC’s sound system response was perfectly flat. There are basically two steps in this last ‘response calibration’ procedure, the first of which is to measure the overall frequency response of the PC’s sound system. This is done using TrueRTA’s spectrum analyser function, in conjunction with its Generator’s Quick Sweep button. The physical set-up remains the same as before, with the Left Line Output looped back to the Left Line Input as shown in Fig.11 (except that the external millivoltmeter is no longer needed). Just before you do this, you need to switch on TrueRTA’s Spectrum Analyser function by clicking on the second tool button from the left-hand end, just below the Edit menu button. This is the button with a little bar-graph symbol on it, alongside the DSO tool’s ‘sinewave’ button. Then move over to the right and click on the Spectrum Analyser menu button, at top centre of the screen. This will cause the Analyser’s drop-down menu to appear, as shown in Fig.13. The purpose of this menu is to allow you to set up the Analyser tool, ready for doing the sound system calibration. You should find there’s a tick alongside the top menu item ‘RTA Mode (Real Time Analyser)’. Then you’ll need to select the maximum Analyser resolution that’s available for your level of TrueRTA – which is 1/24 Octave RTA in the case of level 4, as you can see in Fig.13. Click on the next menu item down, which is ‘Speed Tradeoff’, which will bring up a small flyout box. For this calibration job you should select the ‘20Hz (precise but slowest)’ option. Next check that there is NO tick next to the next menu item down, labelled ‘RTA Bar Mode’. This is to ensure that the Analyser will display its results in line graph mode, rather than in bar graph mode. The Analyser control panel (on the right side) allows you to set the audio level at the top and bottom of the display and also the high frequency and low frequency limits. It also provides another way to select the RTA Resolution and the Speed Tradeoff, plus you can also type in the number of analyser sweeps you want it to average before the results are displayed. For this initial system calibration set all these remaining options as follows: Top limit 0dBu, Bottom limit -40dBu, Hi Freq Limit 50kHz, Lo Freq Limit 10Hz and Averages 10. Make sure that the Generator Ampl level (left side control panel) is set for -10dBu and finally click on the Quick Sweep button at the bottom of the same control panel. After a few seconds you should see a frequency response plot similar to that shown in Fig.15. This is the overall frequency response of your PC’s sound system, at this stage in its ‘naked glory’, ie, without any correction applied. By the way, the curve shown in Fig.15 is that for the SILICON CHIP USB Recording/Replay Interface. Now that you have made sure that the Analyser tool is working correctly, it’s time to use it to perform the actual sound system calibration. This is done by clicking on the Audio I/O menu button at the top of the screen and then clicking on the ‘Sound System Calibration...’ item down near the bottom of this menu. This will cause the PC Sound System Calibration dialog box to appear, as shown in Fig.16. As before this gives you an easy-to-follow summary of the steps involved in the calibration, so once you’ve read this all that remains to be done is to click on the [OK] button. TrueRTA will then generate a calibration file for your PC’s sound system and save it for use in the future. If you now do another Quick Sweep, you’ll get a somewhat different frequency response plot for your PC’s sound system. As you can see from Fig.17, it will now be close to ruler-flat, between the lower frequency limit of 10Hz 68  Silicon Chip siliconchip.com.au Fig.21: This plot shows the crosstalk into the right channel (lower curve) when the left channel was quickly swept at a level of -10dBu (upper curve). Fig.22: The crosstalk plot of Fig.21 with the right channel’s noise floor plot subtracted from it, to show the crosstalk alone (lower curve). and at least 20kHz. This shows that TrueRTA is now using your Sound System Calibration data to correct the overall frequency response and make it effectively flat. Your PC and its sound system are now calibrated, at least to the level where you’ll be able to carry out quite accurate gain and frequency response measurements on amplifiers, filters, mixers, equalisers and the like. But as I mentioned earlier, things aren’t quite so straightforward when it comes to measuring things like noise, distortion and channel crosstalk. So let’s look at these now, to give you a solid grounding of the PC sound system’s full capabilities before you move on to practical ‘real world’ audio equipment testing. from around -120dBu at 15Hz slowly up to about -90dBu at about 20kHz and with a few small ‘blips’ along the way. Clearly this noise performance wouldn’t have a serious effect when you are making noise measurements on equipment with somewhat higher noise levels, say above -70dBu. But it would certainly result in considerable error if you were trying to make measurements in amplifiers or other equipment with noise levels either comparable with the interface itself, or even better. Does TrueRTA provide a way of cancelling out the input noise of your PC sound system, so you can make reasonably accurate measurements on low-noise equipment? Well, it does provide one way to do this, although it doesn’t seem to offer an automatic cancellation in the same way it does to correct for the sound system’s frequency response. Instead it allows you to save a noise plot of the sound system itself, like that shown in Fig.18, in one of its memories. Then when you take a noise plot of your low-noise equipment and save it in a second memory, you can use TrueRTA’s ‘Memory Difference’ utility to subtract one plot from the other. The only complication here is that the resultant plot tends to be displaced vertically, so it can be tricky working out how far to move it up or down using the ‘Shift’ utility, to restore it to the ‘correct’ level. Still, this can give you a reasonable idea of the equipment under test’s own noise performance. TrueRTA’s Analyser also offers a ‘Relative Mode’, whereby once one plot is saved, further plots can be made and displayed in terms of their relative values to the saved plot. This is a bit more like automatic cancellation but it’s more suited to tasks like comparing the gain or frequency response between channels, or the effects of tone controls. Noise measurement When it comes to measuring parameters like noise and distortion, in an ideal world you’d be able to use ‘perfect’ measuring instruments like an AC millivoltmeter or spectrum analyser with no internal noise of its own and an audio signal generator with a ‘pure’ sinewave output having absolutely no distortion or noise. But of course such instruments don’t exist in the real world, any more than perfect amplifiers or any other kind of equipment. Everything in the real world is imperfect, including test instruments. That’s one of the reasons why high-end noise and distortion measuring instruments tend to be so expensive – because the designers and manufacturers have to put so much time and effort into achieving the lowest possible noise and distortion figures. So it’s probably unrealistic to expect this level of performance from our low cost PC-based measuring system. But just what can we expect? Well, let’s use the SILICON CHIP USB Recording and Replay interface as an example. First of all, look at the plot in Fig.18, which shows the ‘noise floor’ of the Left line input of the interface with its input taken to ground via a shielded 50 resistor. This was measured using TrueRTA of course and the Right line input gave a virtually identical plot. As you can see the noise generated within the interface’s line input circuitry is fairly low but quite significant, rising siliconchip.com.au Distortion: a little harder Things do get a little more complicated – read ‘tricky’ – when it comes to distortion. That’s because the sound system’s own circuitry (including the ADCs and the DACs) inevitably introduces some distortion of its own. Some is introduced by the DACs and line output circuitry, October 2011  69 so the output from TrueRTA’s software Generator will not be distortion-free for a start. Similarly, the line input circuitry and ADCs will also introduce some distortion, so our Analyser will not be distortion-free either. So if you take a noise and distortion plot of the PC sound system itself using TrueRTA’s Generator and Analyser, you get a result like that shown in Fig.19. This plot was taken with the Generator set to produce a 1kHz sinewave at a level of -10dBu and feeding directly from the USB Record/Replay interface’s Left line output to its Left line input. As you can see, it shows the Generator’s 1kHz fundamental component in the centre, with a second harmonic ‘spike’ at 2kHz and third, fourth, fifth, sixth and seventh harmonic spikes visible as well, at levels varying from -87dBu down to -94dBu. Of course these are also sitting on the Interface’s noise plot, which tends to make them seem worse than they are. But you can get a clearer picture of the distortion components by subtracting the interface’s own noise plot from it, to produce the plot shown in Fig.20. This was done using TrueRTA’s Subtract utility to subtract the interface’s noise floor plot from the distortion and noise plot of Fig.19 (and then moving the resultant back down into the correct range using its Shift utility). This ‘(D+N) - N’ plot does give a somewhat clearer view of the sound system’s overall distortion performance, as you can see, although TrueRTA doesn’t provide a utility for using this information to calculate the total harmonic distortion (THD) as a percentage. All it gives you is a table in the User’s Guide showing the relative distortion figures for harmonic levels from -5dB to -120dB below the fundamental. So if you want to calculate the THD you have to do this yourself, by finding the percentage levels of the various harmonics and then doing a ‘square root of the sum of the squares’ calculation. I did this myself using the plot of Fig.20 as a starting point and found the THD to be around 0.017% – not too bad but clearly not wonderful either. Of course even when you have done this somewhat tedious calculation, the figure you get is not all that useful when it comes to measuring the distortion performance of external equipment. You could get a ‘rough estimate’ of the equipment’s THD by using TrueTHD to do a plot of its distortion at 1kHz, working out an equivalent THD figure and then subtracting the system’s own THD figure from it but this would not be particularly accurate. In any case it would be for only one frequency. It would be very tedious to repeat this procedure for many different frequencies, which would be needed if you wanted to produce a full distortion plot. I think we can therefore conclude that it’s not really feasible to use a software package like TrueRTA to produce a full THD plot for relatively low distortion audio equipment like amplifiers, preamps and filters. You would be able to use it to produce spectrum plots like that in Fig.20, to give you a rough idea of the equipment’s distortion at different frequencies. Of course using TrueRTA with a calibrated microphone, you would be able to plot the distortion of higher-distortion acoustic equipment such as loudspeaker systems. That’s because with these, the sound system’s own distortion performance would be so much better than that of the gear being measured, you could safely ignore it. Summarising On the whole then, it’s fair to say that a PC-based audio testing system using a software package like TrueRTA together with a good quality sound card or USB interface is capable of making quite accurate measurements of the gain and frequency response of things like amplifiers, preamps, filters, equalisers and mixers. It is also capable of making reasonably accurate noise plots of the same equipment, together with spectrum analysis plots of the distortion at specific frequencies. But it’s not really capable of being used to provide THD or THD+N plots and its utility in plotting inter-channel crosstalk is quite limited. Still, it’s a big step forward being able to measure the frequency response and noise performance of this kind of equipment. SC Presensitized PCB & associated products IN STOCK NOW! •Single Sided Presensitized PCBs •Double Sided Presensitized PCBs •Fibreglass & Phenolic •UV Light Boxes •DP50 Developer •PCB Etch Tanks, Heaters & Aerator Pumps •Thermometers •Ammonium Persulphate Etchant •PCB Drill Bits (HSS & Tungsten) For full range, pricing and to buy now online, visit 36 Years Quality Service 70  Silicon Chip www.wiltronics.com.au Ph: (03) 5334 2513 Email: sales<at>wiltronics.com.au siliconchip.com.au