Looking for Audio Frequency Response and THD analyzer

[davorin]

Evening (o;

Searched this forum but didn't find any post related to my needs...or I just used the wrong words as a non-native english speaker ;-)

I remember using a Neutrik device 40 years ago with a calibrated microphone and audio output, and it plotted than on paper the frequency response from 20Hz to 20kHz on a piece of paper.

But I am more looking for a device with audio input/output and USB/LAN interface to measure audio amplifiers frequency repsonse and THD.
For speakers I use an OpenSource software REW as FuzzMeasure from Røde has been abanded a while ago.

There is the QuantAsylum QA403, but not available here in EU. Also it comes only with Windows software.

Any other suggestions?

[pope]

The QA403 is available in Europe (elektor)

[themadhippy]

from the neutrik stable hows about the  the minlyzer, https://www.nti-audio.com/en/products/minilyzer-ml1 although it gets a bit pricey if you want the pc interface as well

[KungFuJosh]

The Keithley 2016 might have some of what you want, but the QA403 is the way to go. You can also use oscillscopes for that stuff.

[Berni]

The right tool depends on how low THD you want to measure.

For THD down to about 0.5% you can use most test gear that can generate and record waveforms. Some oscilloscopes even have built in features to do bode plots and similar. There are signal analyzers that are basically spectrum analyzers that do down to DC.

When going into THD of <0.01% that's when specialized test equipment is the only way to do it. These low levels of THD are typically only seen in audio. The company Audio Precision is the big brand in these but they don't come cheap.

If you want to do low THD measurements on the cheap then you have to combine both audio and consumer gear, this is a sound card. Any high quality sound card can be used for this. You can easily check how suitable a card is by looping output directly back into input and seeing what its own THD floor is. Problem is that the ports on sound cards are not very robust so it is easy to blow it up when connecting to random things like that. This is where the QuantAsylum QA403 comes in. It is a high performance sound card designed to be robust for uses like this. As said above it is no problem buying one in EU, Elektor sells it.

[rusoaie]

Hello!

For signals up to a few volts rms output, any half-decent audio interface (Focusrite, MOTU, Universal Audio, RME etc) will do the job.

However, for amplifier output measurements you should a some dummy load connected, the amplifier should be powered through a isolation transformer and the amplifier output shouldn't exceed the audio interface's maximum input voltage - otherwise you risk blowing up your interface/laptop. Proceed at your own risk!

I am usually doing only 1W power amplifier THD measurements with the audio interface and full power measurements with the scope (scope THD input is limited to 0.03%, but you can measure hundreds of volts peak-to-peak with a 10x probe).

Attached an example of a MOTU Ultralite mk5 1kHz THD measurement at 14dBV (16dBu) - this unit supports balanced inputs up to 24dBu / 21dBV / 12Vrms / 34Vpp.

As for a proper audio analyzer, besides the QA403, you can consider a chunky, old, vintage, museum-grade HP / Panasonic / R&S audio analyzer from eBay.

[Simmed]

at ASR
have a look

https://www.audiosciencereview.com/forum/index.php?threads/audio-precision-apx516b-review.52335/

i cannot remember the model
there was a cheaper test gear ADC/DAC like AP

ah this 1
ADC DAC etc
https://www.audiosciencereview.com/forum/index.php?threads/e1da-cosmos-apu-teardown-and-personal-thoughts-and-measurements.37739/page-7
https://www.audiosciencereview.com/forum/index.php?threads/e1da-cosmos-adc-teardown.27967/
https://www.audiosciencereview.com/forum/index.php?threads/e1da-cosmos-adciso-review.54020/

some MOTU have quirks
https://www.audiosciencereview.com/forum/index.php?threads/motu-ultralite-mk5-review-audio-interface.24777/

[davorin]

Elektor hasn't the QA403 in stock....and the price is rather high as the US$ is low compared to CHF (o;

Did Quantasylum ever have a black friday sale in the past?

[RoGeorge]

Recent 12bits oscilloscopes (Siglent and Rigol for sane prices) can do Bode plots to see the audio response, and Fourier plots to calculate the THD.  They can be interfaced with a PC, and if you buy one, you are also left with an oscilloscopes with 2-4 channels up to 100-1000MHz.

A 24bit/192kHz sound card can measure a lot of audio parameters, though I would probably go for a 12bits oscilloscope.  An oscilloscope is more than enough to evaluate audio equipment (including the THD if you make a live FFT and square the harmonics), plus that you will be left with an incredibly performant digital oscilloscope for the money.

For Siglent models, there is this demo-like topic, made by somebody from Siglent.  It's kind of mind blowing what a DSO can do these days (see the first 2 link in the first post, they have attached a PDF book with all the demo):  https://www.eevblog.com/forum/testgear/sds800x-hd-review-demonstration-thread/

If you decide for a 12bit oscilloscope, there are a few topics here about either Siglent or Rigol, which can do what and so on.  Buy the best hardware you can afford with the given budget, but don't spend on buying software features for the oscilloscope.  The software features can be unlocked for free.

[hotze]

The task is not just to analyze signals but also to generate them, which requires an appropriate signal source as well (depending on how accurate the measurements need to be).

[RoGeorge]

Some oscilloscope models also have an included AWG, but I would prefer a separate signal generator.  For audio only, a sound card can be an excellent signal generator.

Advocating for an oscilloscope rather than a dedicated audio analyzer, because in practice I never need to make very accurate audio measurements.  It doesn't matter to me if the THD is 0.1% or 0.0001%, I won't hear any difference.  But I do need quite often to measure outside of audio range, or decode digital protocols, or who knows what else.

An oscilloscope can be used at almost anything electronics related, while an audio equipment, no matter how good, it is very limited in amplitude ranges, bandwidth, and limited in what it can measure in general.

For comparable (about the same) money, the oscilloscope can do incomparable more than an audio analyzer, in my opinion.  Of course, each of us may have different needs/priorities/interests/choices/preferences/budgets/etc, not trying here to force others into buying what I would buy. 

[2N3055]

QuantAsylum 403 has fully differential measurements and generation and support large voltage ranges.
It also have many predefined measurements useful in audio.

It is a tool to have if you are doing a lot of audio. It makes many things fast and easy.

Like RoGeorge said, you can go a long way with general purpose AWG + 12 bit scope.
For frequency sweeps it might be even better, for instance Siglent supports 3 different outputs simultaneously in Bode plot. So for instance you can measure a 3-way crossover frequency characteristics in one go and in single graph. And measure automatically.

In ideal world, QA403 + SDS800xHD and SDG1000x+ would give you all the capability needed for pretty much any audio work except audiofooling...

[macboy]

With a very good sound card and some software on the PC, you can actually do THD and IMD measurements quite well. One of the tricky parts is connecting the amplifier speaker output to the line input of the sound card, especially for bridged amps or any type of digital amp. "Real" audio analyzers always have differential inputs for a good reason. The inputs of the sound card are usually single-ended unless you have a pricey pro card. And of course you need to attenuate the levels, and add protection to avoid blowing up the sound card if you get the attenuation range wrong.

Years ago I used RMAA, Rightmark audio analyzer. I know that there are many others out there. I haven't used a PC based system in a while due to lack of need. I'm lucky enough to have an M-Audio sound card. This has some of the best AKM ADC/DACs available plus balanced inputs, so it is about as close to ideal for this task as one can find. I also have an old Panasonic VP-7722A (a.k.a. Matsushita a.k.a. National) audio analyzer on the bench, which I use occasionally. That lovely old boat anchor has a distortion floor around 0.00010% THD (1 ppm or -120 dB). Older instruments like that lack a fancy GUI or USB/LAN connections, but it has GPIB if you are equipped for that. One advantage of a dedicated instrument over a sound card is the bandwidth. Even at 192 kHz sampling rate, the bandwidth is limited to around 90 kHz and is unlikely to be flat in that range, so measuring distortion at say 20 kHz is rather futile. Dedicated instruments may have 500 kHz bandwidth or more, for that reason (that's a general statement of course, so read the specs of whatever you consider buying).

[Fried Chicken]

With a very good sound card and some software on the PC, you can actually do THD and IMD measurements quite well

What constitutes a good soundcard?  Short of buying a purpose built unit, do they typically report specs and/or are spec tests published online by independent testers?  Any recommendations? (just saw the praise for the m-audio sound card)

One of the tricky parts is connecting the amplifier speaker output to the line input of the sound card, especially for bridged amps or any type of digital amp.

A bridged amp has a separate ground (not at the speaker's "negative" terminal), right?  Could connect the amplifier output to a dummy load and then probe over that dummy load?  How would one do so without introducing interference?  What do you mean by digital amp, and why are those tricky?

"Real" audio analyzers always have differential inputs for a good reason. The inputs of the sound card are usually single-ended unless you have a pricey pro card. And of course you need to attenuate the levels, and add protection to avoid blowing up the sound card if you get the attenuation range wrong.

Again, dummy loads connected in parallel, or?

Years ago I used RMAA, Rightmark audio analyzer. I know that there are many others out there. I haven't used a PC based system in a while due to lack of need. I'm lucky enough to have an M-Audio sound card. This has some of the best AKM ADC/DACs available plus balanced inputs, so it is about as close to ideal for this task as one can find. I also have an old Panasonic VP-7722A (a.k.a. Matsushita a.k.a. National) audio analyzer on the bench, which I use occasionally. That lovely old boat anchor has a distortion floor around 0.00010% THD (1 ppm or -120 dB). Older instruments like that lack a fancy GUI or USB/LAN connections, but it has GPIB if you are equipped for that. One advantage of a dedicated instrument over a sound card is the bandwidth. Even at 192 kHz sampling rate, the bandwidth is limited to around 90 kHz and is unlikely to be flat in that range, so measuring distortion at say 20 kHz is rather futile. Dedicated instruments may have 500 kHz bandwidth or more, for that reason (that's a general statement of course, so read the specs of whatever you consider buying).

I was looking at RMAA, and also an analog audio analyzer.  Forgive the simple questions, it's not out of some naïvety.  I'm thinking of getting a soundcard for a PC that will do both audio and THD analysis.  Alternatively I'll scan for an analog audio analyzer.  Reliability is always a question

[GLouie]

If OP wants something less expensive than QA403 and no Windows, there is nothing on the market that I know of. You would have to cobble up something such as a consumer USB audio interface and/or other test gear, use Mac freeware like REW, and do some work to verify the accuracy and limits of measurement.

There is a new product announced from Neurochrome in Canada, which is an interface but unreleased and apparently without software (use REW, I guess), and expected at $5800USD. But compared to the gold standard Audio Precision, cheap.

Re: about measuring digital amps, the output of the class D amps I have seen output a modulated high frequency (typ. 250kHz) signal. To measure, the standard answer is a huge, high powered load & low-pass filter, ala the Audio Precision AUX series digital amp measurement filters. I attach a TI application note;