Now Shipping: 48K/192Ksps 24-bit Audio Analyzer
For too long now, serious hobbyists and professionals on a budget have been making due with sound cards and audio analysis programs in an attempt to bring high-quality analysis to the PC at a reasonable price. And it is understandable why. Even the most basic projects today can deliver performance far beyond what a scope can display. And forget about trying to characterize the non-linearity of your audio circuit with a scope and desktop sig gen. The scope front-end alone is likely far worse than the $0.50 op-amp your project uses.
And we’ve been there ourselves. We’ve tried to peek into problems on various projects using just a scope. When that didn’t work, we studied what others were doing and started using ARTA and SpectraPlus combined with a high quality sound card. While the solution was workable, it was expensive. SpectraPlus has a free demo with amazing options, but once the demo is done, the price can reach north of $500 easily. And that just for the software.
ARTA is amazing piece of software, but it is shareware. And use on a commercial project is north of USD$200. And that’s just the software. There’s still a sound card to pick, and that will likely run you another $125. Combined, your software and hardware combination can be tweaked to give very good results. But the tweaking will be ongoing. Windows can make changes to your setup. Your input level knobs will need the faintest and repeated touch of your finger to reach a precise level of adjustment. And then there’s the output. While soundcards are offering amazing input abilities today, their outputs are still on the noisy side.
And what if you want automation, or the ability to control the instruments from a program so that complex measurements can be made quickly?
The QA400 Design Philosophy
The QA400 was designed from the start with a focus on cost. Looking across the landscape of high-end audio CODECs, there was a flurry of activity from all the major silicon vendors up until 2008. And at that point, a surprising thing happened: Development of new codecs basically stopped. Sure, there are still some esoteric parts being worked on by smaller boutique vendors. But largely, the various codecs that were out there were delivering solid performance, and the performance was quite a ways beyond what Windows was able to take advantage of (in terms of sample rate and bit depth). To work around the Windows limitation, new drivers and driver models were needed. The pro audio folks also needed miniscule round trip delay times, further distancing them from the plethora of Windows driver models.
And slowly but surely, 192Ksps 24-bit audio made its way on to PCs. But it wasn’t cheap. The big vendors could do a licensing deal on the drivers, but the smaller players, no way. The driver costs alone could sink the project economics.
And so, we made a decision up front to de-couple from the standard Windows audio directions. And this has proved a good decision, as it has allowed us to sidestep the normal issues that prevent you from trusting your PC audio measurements and it’s saved quite a bit of money in the process. After all, why on earth should Skype be able to adjust the various gain settings of your audio setup? It shouldn’t. And yet any app you install can tweak your core audio settings. How can you ever purpose a testing device for unattended use in a factory if a 3rd party game a technician installed from a USB key is able to change your painstakingly arrived at audio settings?
With the cost side of the equation within our grasp, we needed to pick a performance level. ADC performance is largely a matter of how much money you want to throw at the problem. There are some very, very good parts out there. The PCM4222 was a leading candidate for us early on, but it would cause break out of the $200 target selling price. Not only that, the performance of the PCM4222 is so good that there are few audiophile and consumer products that will need that level of performance. But from our early study, we felt we could get within about 10 dB of the PCM4222 with a much cheaper part: The CS4272. And that 10 dB would still give us plenty of margin for measuring an enormous range of pro-audio and consumer products. Not only that, we felt we could achieve the design target using just USB power. Hitting the limits of what the PCM4222 can offer likely won’t happen without some extra cost spent on the power supplies, and it would be a shame to compromise the performance of the part due to USB power issues.
And with the codec picked, we set off to develop what you see below.
At its core, the QA400 can generate and analyze very accurate tones. It doesn’t do this by playing constant tones. Instead, it generates a series of bursts at the frequency of interest. The reason for this is so that we can treat each measurement event as a discrete happening. In other words, the event has a clear beginning and a clear end. With the start and finish clearly defined, the hardware can then detect if any bits are lost, or if there are any gaps in the transmission or reception with the PC. This means that you won’t be surprised by a measurement that was distorted because your audio was skipping.
The basic measurements engineers are interested in while characterizing a device-under-test are:
Power Measurements. This tells you the noise floor of the device are measuring. Because the QA400 has an absolute reference point, it can tell you precisely the amplitude within a given bandwidth. Below, you see the noise floor of a well regarded portable player Sansa Clip+ at max volume playing a silent MP3 file. You can see the noise floor measures –86.5 dBV. But more importantly, notice the QA400 has almost 25 dB of measurement margin at 1 KHz.
Harmonic Distortion Measurements. The QA400 makes two types of distortion measurements, THD and THD+N. Both measurements can auto-pick the fundamental (in case you are using an external circuit), track the internal generator, or allow you specify a precise frequency. The harmonic measurements will continue up to the specified measurement stop point. Below, you can see the QA400 in loopback mode at roughly its sweet spot for THD. The sweet spot for THD+N happens at about the –3 dBV output level.
With the limitations of the QA400 shown above, we can look at a very well regarded FiiO E17 Headphone Amp playing a –10 dBFS 1 KHz tone. This time, we look out across the full 96K ADC bandwidth. Notice the bump out beyond 40KHz is inherent to the CS4272 ADC used in the QA400. We can see the peak on this –10 dBFS tone is –3.5V. This implies the 0 dBFS output from the E17 is nearly 7 dBV. Note that in the plot below, you can see the E17 output with no load (green trace) and 16 ohm load (yellow trace). There is very little difference in the output under this different loads.
Referring again to the Sansa Clip+, a THD and THD+N measurement on the Clip+ is shown below. Others have measured this too, with different equipment. For example, NWAvGuy measured his Sansa Clip+ on his dScope analyzer (186mV into 15 ohms) and arrived at a THD figure of 0.04812% and a THD+N figure of 0.05516%. Both figures are extremely close to what the QA400 measured, and the spectrum resemblance is clear too.
The final major measurement made by the QA400 is frequency response. The QA400 uses an impulse response as the stimulus. The FFT of that captured impulse response is the frequency response of the DUT. In the plot below, you can see the lower frequency response extends down to a few Hz, while the upper 3dB corner is beyond 40 KHz. Across the 20 to 20 KHz bandwidth, we spec +/- 0.07 dB.
The UI & Automation
Manual measurements on the QA400 are controlled via large buttons, like on the QA100.This makes the device usable on emerging touchscreen devices. Sharing data is easy, and the parameters related to each measurement are there for you to readily see on each copied graph. Markers can be readily applied to your data, helping you to interpret the results for others. And rather than a cluttered, main menu, the levels and levels of options, each button has its own submenu.
The submenu for the THD menu is shown below. And it’s here you can readily set the parameters that are relative to the THD measurement you are making.
Automation refers to the ability to make the instrument automatically make a series of repetitive measurements. The QA400 offers 3 ways to help you accomplish this. The first is via canned swept tests. The 1.0 release of the software ships with the ability to measure THD over a range of frequencies and output settings. Just set the values you are interesting in measuring across, start the test, and the data is automatically recorded to a text file for deeper analysis.
Not only that, but the swept test subsystem allows you to plug in your own solutions. If you want to sweep THD across a range of output levels, frequency AND amplifier voltage, then no problem. You can write your own plug-in in C#, C++ or Visual Basic that can accomplish just this. And in your own code you can call the GPIB or similar code needed to set your supply to the correct voltage. We provide the source code for our own THD versus Output Level Configuration plug-in.
The final level of automation permits you to write an application that can drive the QA400 application remotely. And by remotely, we mean from your desktop, or across the factory floor, or across the world. The remote connection happens over TCP. In the code below, you can see just how little code it takes to capture a THD measurement. Automatically. This type of code (and the plug-ins described above) can be written using free tools from Microsoft. And Intellisense and strong typing are all part of the offering. If you’ve ever wondered why the string your sent to the power supply over GPIB wasn’t doing anything, you don’t have to worry about that with the QA400. The Intellisense in Visual Studio prompts you to get the exact call terminology and ordering correct. And the strong typing ensure that you aren’t stuck sending long abstract strings of text over a USB connection masquerading as a serial port.
Click here for full specifications as well as come comparisons to the popular ARTA + EMU0204.
Purchasing and Shipping
The price of the QA400 is $199. We offer a 15-day money-back guarantee and a 180 day warranty against defects. You can use your credit card (no PayPal account needed) or PayPal account credentials for the actual purchase. Shipping will be calculated at checkout based on your location.
We can ship to the US for $14 S&H via USPS Priority. This will usually be a 2-3 day shipping time and is very economical.
We can ship to Canada for a $32 S&H via USPS Priority. Total transit time is usually 1 week to Canada. The purchaser is responsible for additional tariffs, taxes and/or duties. For import tax purposes, the analyzer HARDWARE is noted as having a value of $70, and the tariff code is 903040. This code might be used to determine any tariffs that will be due. For example, in the US, importing a product with this tariff code results in a 0% duty on the product. The analyzer SOFTWARE represents the balance of the purchase price. As the software is downloaded, you will need to check if local governments require a tariff or duty on software downloads.
We can ship internationally to other countries served by USPS Priority for $48 S&H. Total transit time is usually 2 weeks. The purchaser is responsible for additional tariffs, taxes and/or duties. For import tax purposes, the analyzer HARDWARE is noted as having a value of $70, and the tariff code is 903040. This code might be used to determine any tariffs that will be due. For example, in the US, importing a product with this tariff code results in a 0% duty on the product. The analyzer SOFTWARE represents the balance of the purchase price. As the software is downloaded, you will need to check if local governments require a tariff or duty on software downloads.
Feel free to contact us with any question using the email alias "sales", or use the support forum.