Audio Analyzers (but not a toy)

[KungFuJosh]

I've been interested in getting a spectrum analyzer, primarily for audio equipment testing, for a while. This seems to be a weird quest for some reason. It seems most suggestions are to lean towards basically any sound card for a computer, but I'd prefer something standalone, unless there's something 'attached' that's really nice. I'd also like the "instant gratification bode plot" feature. 😉

100Hz to 20kHz would be the absolute minimum range, but down to 10Hz or 20Hz would be better. It seems the fancier new SAs all bottom out at 9kHz for some reason.

[2N3055]

I've been interested in getting a spectrum analyzer, primarily for audio equipment testing, for a while. This seems to be a weird quest for some reason. It seems most suggestions are to lean towards basically any sound card for a computer, but I'd prefer something standalone, unless there's something 'attached' that's really nice. I'd also like the "instant gratification bode plot" feature. 😉

100Hz to 20kHz would be the absolute minimum range, but down to 10Hz or 20Hz would be better. It seems the fancier new SAs all bottom out at 9kHz for some reason.

SA are primarily RF instruments.
SA for audio are called audio analysers and there are many of those.
You won't like the price.
Take a look at Quant Asylum.
Or stick with audio interfaces...

[David Aurora]

I've got a 35665A which is built exactly for this purpose, and in all honesty lately I keep thinking about getting rid of it and using a sound card instead.

It's nice having a standalone thing on the bench that you can just plug into and go, but it's also pretty slow/limited compared to computer based solutions.

[themadhippy]

If you can live with 1/3rd octave steps maybe something like the old klark teknic dn60

[dobsonr741]

I’m using an external pro USB audio interface, 24 bit resolution with 192KHz sampling rate. REW is an excellent free app for spectrum analysis, I also have Jupyter notebook to do the same FFT Rew can do. Tons of discussions on how to do it and what interface has the lowest noise floor and distortion - see https://www.diyaudio.com/community/threads/how-to-distortion-measurements-with-rew.338511/

Back to you post - it’s attached, and REW updates the spectrum display real time on screen. Can generate signals too, however it’s not as good as a standalone precision Wien bridge. Much discussion on that too.

[KungFuJosh]

Fiiiiiiiiiiine. Software with sound card  and/or audio box like the QA403 Audio Analyzer is probably the way to go. The QA403 makes me want to make something myself though.

I already have a nice Roland Studio Capture, so I guess I should look at what software is good. I'll check out REW, what else is good?

Edit: For the record, I really would prefer to avoid using my DAW for this. I would like at least the software to be standalone.

[dobsonr741]

Two more software - free and paid:

https://sourceforge.net/projects/audmes/
https://artalabs.hr/

And again, it was not a big deal for me to roll my own 64K window size FFT in Python either.

If you like dedicated hardware, you can make a raspberry pi4 running the FFT with Jupyter notebook serving your browser to view - however you’ll not be refreshing as often as a PC or Mac does.

[4thDoctorWhoFan]

There is the HP 3561A which is nice.  Old but nice.

[St66666]

https://www.virtins.com/

They have some interesting stuff. Not free, but free to try for a while.

[srce]

https://www.ap.com/analyzers-accessories/apx555/

[Floopy]

There is the HP 3561A which is nice.  Old but nice.

Second that, they're nice, but the screens tend to fail easily because of the flyback. They do DC - 100kHz, I've had one for years.

[Grandchuck]

I have an HP 8903B but lately have been using REW software along with a Focusrite Scarlet Solo.  You can see, for the cost, the performance is good: https://www.eevblog.com/forum/testgear/question-cheap-pure-sine-wave-genertor-1khz/msg4737746/#msg4737746

[MathWizard]

What is the problem a wide BW SA has with low frequencies ? And why are AA so expensive then ? Measuring up into the GHz must be way harder and more expensive than audio to AM radio Frequencies.

IDK I just figured a SA could do better, even if they added another PCB for LF

[Sighound36]

I feel that pretty much all the audio analysers are 24bit adc's with large signal to noise ratio's of upt 120dB. Plus many different audio/Blue tooth/I2s/rub & buzz and headphone measurment applications. They all perform both digital & analogue domain measurments The APX555b offers 1.2M FFT's from dc-200Khz.
Does this justify the cost, for hobbists no it's out of the equation. Although these are mainly geared for production line testing with they ability to set up multipule tests in sequence mode

We have four audio analysers R&S UPV66 all options / AP APX555B most options / Prism sound D scope series III / Standford research SR1 (under rated in my book).

IMHO the Picoscope 4262 is also a great device for audio investigation up to 16 bits of resolution 16Mpts buffer, decent analogue genny and really good FFT's and for the cost its decent, a couple of us on here use them regularly for good reasons.

As good as all these audio anaysers are, I still use the Lecroy attached at the same time to the signle ended analogue outputs using both time & frequency domain above the usual audio BW restrictions, you can see a lot more going on further up the rev range if you like 

This peaked my interest although purely an analogue domain unit, its features and size make it ideal for a portable in the field unit.

https://www.virtins.com/RTX6001.shtml


Would also agree with the HP 3561A as well if you have the spa on the bench.

[mawyatt]

We second the PicoScope 4262, although don't use as much as we should since it requires a laptop, and we much prefer a dedicated standalone DSO. Also does quite a respectable job as a low frequency FFT analyzer, or FRA.

Best,

[2N3055]

We second the PicoScope 4262, although don't use as much as we should since it requires a laptop, and we much prefer a dedicated standalone DSO. Also does quite a respectable job as a low frequency FFT analyzer, or FRA.

Best,

1+ for Pico 4262

[danils]

In my view it's unclear if the OP is looking for a spectrum analyser working in the audio band or a so called "audio analyser" with all the audio related measurement function.

If the request is about the second, there are some 90s R&S UPA-UPL-UPV etc. all-in-one boxes on the bay site at interesting price. As usual there are limits like FFT resolution, no more support etc. some are MS-DOS based machines.

With about 2000 USD you get a UPA that is still a good machine for not-so-demanding measures.

I personally own a AP SYS-2300, still impressive but still pricey

[KungFuJosh]

In my view it's unclear if the OP is looking for a spectrum analyser working in the audio band or a so called "audio analyser" with all the audio related measurement function.

What I really wanted was instant gratification bode plots (for things like guitar pickups, and speakers), but I'll stick with my scope for now, and eventually try out the software based stuff.

[nenea dani]

Focusrite 4i4 has only one linear source and about 5 in switching mode, so the sinusoidal signal is not extremely clean in generator mode. I am attaching a photo for compliance, although one of the sources is masked by cables. With REW I noticed that after 25kHz the signal drops and above 30kHz it starts to have phase noise. In loop mode, the oscilloscope seems to provide a synthetic signal because everything seems perfect. I tested with the Fluke196C, which I trust.
I also asked on this forum why most SAs start with 9Khz or a few with 5Khz. One of the answers was that there is a problem with the input capacitor. For the analog situation when the audio signal is mixed with another one, I found out that in the first Hz's nothing conclusive can be measured because the phase noise from the local oscillator extends too much. They are just findings in amateur conditions.

[trobbins]

What I really wanted was instant gratification bode plots (for things like guitar pickups, and speakers), but I'll stick with my scope for now, and eventually try out the software based stuff.

Soundcard and REW provides frequency response of amplitude and phase, from 2Hz to 96kHz.  Compensation is automated for 2Hz to circa 10-20Hz, and from circa 40k to 96kHz where the soundcard and probes and test rig may have a non-flat response.  That is, as I understand it, your instant 'Bode' plot.  For some esoteric audio applications where feedback related stability is a concern, the 96kHz upper limit can be too low, as measurement of gain and phase margins require a wider bandwidth (and hence a need to go to something like a Picoscope 2206B with FRA software to extend Bode plot bandwidth to 1MHz).

  Focusrite 4i4 has only one linear source and about 5 in switching mode, so the sinusoidal signal is not extremely clean in generator mode. I am attaching a photo for compliance, although one of the sources is masked by cables. With REW I noticed that after 25kHz the signal drops and above 30kHz it starts to have phase noise. In loop mode, the oscilloscope seems to provide a synthetic signal because everything seems perfect.

I'm unsure what you are describing.  You seem to infer that internal circuitry of the 4i4 is adding spurious signals to a generated sinewave, but only show a photo rather than a loopback response (!?)  Your test setup may need some refinement, although it would be expected that a raw loopback would show the effects of high frequency roll-off from either the DAC side, or the ADC side, or your interconnection, which is why REW allows you to calibrate out such droop.

[jasonRF]

In my view it's unclear if the OP is looking for a spectrum analyser working in the audio band or a so called "audio analyser" with all the audio related measurement function.

What I really wanted was instant gratification bode plots (for things like guitar pickups, and speakers), but I'll stick with my scope for now, and eventually try out the software based stuff.

If/when you decide to try out the software-based stuff, in addition to REW you might want to look at a software package called ARTA as well.  It has a number of different bode plot options to choose from, including ones that are essentially real-time using swept-sine or noise waveforms.  I haven't used it, but from the documentation it is written by someone who understands signals and systems pretty well. 

jason

[Smoky]

I'm getting the hang of using the 35660A, 35665A, and the 3588A when fixing stereo and radio projects.

Crossovers and filters can be checked to see if channels match by using the built-in source or your frequency generator.

Set the span and the marker's location and your off and running.

The analyzers are self-calibrating and easy to use but I'd recommend that the analyzer's power supplies be rebuilt.

[KungFuJosh]

Soundcard and REW provides frequency response of amplitude and phase, from 2Hz to 96kHz. 

I assume you're referring to hardware limitations depending on the sound card in use? Or does REW have frequency range limits?

[trobbins]

REW has a nominal sine signal generator range from 1Hz to half the sampling rate of the soundcard used (so for nearly all soundcards that means 96kHz upper limit).  The impedance measurement also uses a sinewave frequency sweep, which is from half the lower setting frequency, which afaik is 0.1Hz (I beta tested this a few years ago when I asked about lowering the low frequency limit to get a better understanding of choke impedance below the previous limit of 2Hz).

Hardware limitations mean that the soundcard and probes will likely have increasing roll-off below about 10-20Hz, and above about 40-50kHz, which is compensated for by using a calibration curve to get a net flat response for loopback, although the DUT doesn't experience a flat response signal and the compensation starts to introduce noise/artifacts depending on the level of roll-off being compensated.  The low frequency response of a soundcard may be able to be modified (eg. by increasing value of any coupling caps, or removing the coupling caps and managing for a dc input).  The high frequency response may be modified by adjustment of probe compensation or tweaking the compensation circuit in the probe.

[nenea dani]

What I really wanted was instant gratification bode plots (for things like guitar pickups, and speakers), but I'll stick with my scope for now, and eventually try out the software based stuff.

Soundcard and REW provides frequency response of amplitude and phase, from 2Hz to 96kHz.  Compensation is automated for 2Hz to circa 10-20Hz, and from circa 40k to 96kHz where the soundcard and probes and test rig may have a non-flat response.  That is, as I understand it, your instant 'Bode' plot.  For some esoteric audio applications where feedback related stability is a concern, the 96kHz upper limit can be too low, as measurement of gain and phase margins require a wider bandwidth (and hence a need to go to something like a Picoscope 2206B with FRA software to extend Bode plot bandwidth to 1MHz).

  Focusrite 4i4 has only one linear source and about 5 in switching mode, so the sinusoidal signal is not extremely clean in generator mode. I am attaching a photo for compliance, although one of the sources is masked by cables. With REW I noticed that after 25kHz the signal drops and above 30kHz it starts to have phase noise. In loop mode, the oscilloscope seems to provide a synthetic signal because everything seems perfect.

I'm unsure what you are describing.  You seem to infer that internal circuitry of the 4i4 is adding spurious signals to a generated sinewave, but only show a photo rather than a loopback response (!?)  Your test setup may need some refinement, although it would be expected that a raw loopback would show the effects of high frequency roll-off from either the DAC side, or the ADC side, or your interconnection, which is why REW allows you to calibrate out such droop.

I am attaching a diagram with the output from 4i4 for 20kHz sine somewhere at 1/2 FS with the specification that it is not in the loop. Everything is satisfactory, but higher in frequency we have artifacts. I am attaching two more diagrams, one up to 250kHz and the next up to 2.6MHz. The noise generated by the switching sources can be clearly observed, although it is possible that some artifacts have another origin. The noises in the last diagram do not disappear if the signal is off, but are proportional in any situation to the level of the master potentiometer. These disturbances are best seen with the oscilloscope on a rectangular signal because they have about 12-15% of the useful value. The spectral components used in the composition of the rectangle are stable over time, but the noise of the switching sources wanders continuously. I use an LC filter with fc =136Khz, N=7 at the output of Focusrite to clean all the radio frequency when I need a sine wave 20Hz-100kHz. After all, it's a 200 Euro sound card and for the hobby it's just fine.