2254 noise floor and REW setup

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
ruffrecords said:
Select the RTA, set the displayed band with to 10Hz to 20KHz ...The RTA will also give you a figure in dBu for the total power in the selected bandwidth.
Does the measurement BW actually follows the graph's frequency boundaries?
I  can't check it as my PC dedicated to measurements has suddenly decided to call it quits.
 
abbey road d enfer said:
Does the measurement BW actually follows the graph's frequency boundaries?
I  can't check it as my PC dedicated to measurements has suddenly decided to call it quits.

I am not certain. I am not at a PC that runs REW at the moment. I can check that out this evening.

Cheers

Ian
 
RTA in REW has a Distortion settings setup box where low and high pass filters can be applied.  The Noise level shown in the  'Show Distortion' pop up  is related to the filtered bandwidth, as long as the bandwidth includes the identified fundamental being used for distortion calculation (which by default could be the mains frequency hum for an open probe type input).  So yes the measurement bandwidth follows the graphs frequency boundaries, and the plot outside the boundaries is nicely shown as ghosted to help recognise the applied bandwidth.  If there are hum related signals in the bandwidth then they are excluded from the calculated noise measurement, as they relate to the distortion portion of the calculations.

There is also a set of input related measures in Input RMS Level pop up (top RHS), which includes a 22 to 22kHz bandwidth unweighted noise level - but best to see the Help description for those various levels.  The Help also indicates any calibration file may also be being used in the measurement.
 
In addition to previous post, John M elaborated in a REW forum thread that the RTA distortion window and THD part of the calculation method is configured for:
"If the signal generator is playing a sine wave that frequency is used. It doesn't actually have to be the source of what is being captured, so this provides a way of forcing REW to use a specific frequency."

Now that a test signal level down to -180dB can be set, any input can be assessed for noise in RTA using the filters to provide a bandwidth, and the noise level is calculated in the Show Distortion pop up, by using the trick of generating a test signal within that bandwidth at an inconsequential level (ie. -180dB).
 
I have an Audio Precision System 1 with Duke's (Audio1Man in GroupDIY) USB interface and it works great, but its impractical to move around,  so I finally got this https://quantasylum.com/collections/frontpage/products/qa401-audio-analyzer for a bit more than a 2 channel common  USB audio interface and I couldn't be happier, the thing can read down to 0.0004% THD which is around the same as my AP SYS-1, the noise floor is at -113dBV. It can measure noise floor, frequency response, THD, THD+N, IMD, speaker impedance and more, and it is USB powered. It uses a proprietary software which I think looks and works great, and you don't have to worry about calibrating anything. It is not perfect, bandwidth is limited to 80KHz so if you want to measure THD at 20KHz then you are not going to see many harmonics, but still, for the price it can't be beat, technollogy these days, huh?

BTW I saw the other day in an audio magazine (perhaps its old news) that AP is selling the software alone so you can use whatever hardware you want, but I think the software is over $3000  :eek: thats a bit too much just to get the "AP" logo on your measurements, but for a small audio company who wants to publish their measurements and be backed up by the AP name, then its a good alternative, not so much for DIY thou.
 
user 37518 said:
I have an Audio Precision System 1 with Duke's (Audio1Man in GroupDIY) USB interface and it works great, but its impractical to move around,  so I finally got this https://quantasylum.com/collections/frontpage/products/qa401-audio-analyzer for around the same price of a good  USB audio interface and I couldn't be happier, the thing can read down to 0.0004% THD which is around the same as my AP SYS-1, the noise floor is at -113dBV. It can measure noise floor, frequency response, THD, THD+N, IMD, speaker impedance and more, and it is USB powered. It uses a proprietary software which I think looks and works great, and you don't have to worry about calibrating anything.
Interesting. My QA401 has been sitting on a shelf for about 5 years now. my main concern were:
  • Impossible to overlay graphs, making comparison difficult
  • No sinewave sweeps (freq resp graphs actually derived from FF
  • No amplitude sweeps
I enquired with Quantasylum; their answer was that everything I asked was possible as long as I did the programming myself. I'm not a coder so I asked my assistant at the time, who was very good at all sorts of languages, C++, Python...
She said she didn't understand a thing about their language.
So it looks like the software has improved...?
 
abbey road d enfer said:
Interesting. My QA401 has been sitting on a shelf for about 5 years now. my main concern were:
  • Impossible to overlay graphs, making comparison difficult
  • No sinewave sweeps (freq resp graphs actually derived from FF
  • No amplitude sweeps
I enquired with Quantasylum; their answer was that everything I asked was possible as long as I did the programming myself. I'm not a coder so I asked my assistant at the time, who was very good at all sorts of languages, C++, Python...
She said she didn't understand a thing about their language.
So it looks like the software has improved...?

To be honest I've had it for a short time and I can't say if the software is now capable of doing the things you mention.  I'm also not a coder but I've had to program a lot in MATLAB, before I had my AP SYS-1 I used an HP 8903a connected via GPIB to the computer, I coded some programs to control it and read data from the unit, I loved it because I could basically graph everything exactly the way I wanted in MATLAB, the HP unit doesn't have a residual THD as low as the QuantAsylum but I used a Cordell Distortion Magnifier and it dramatically increased its resolution, the only thing I didn't liked about the 8903a it is that the output impedance was fixed at 600 ohms which is fine in most cases but quite high when measuring THD in audio transformers, the "b" version had selectable output impedance between 50 and 600.

I have yet to fully try REW, I know that some people swear by it, I always felt discouraged by the name Room EQ Wizard...
 
I have a QA400 (not 401) and find it to be quite handy for many things but it is a little wonky sometimes. I have taking a SNR measurement and then restarted the software and it yields a different result. Sometimes I'm not super-confident that it's working perfectly. But again, I'm using the QA400. Maybe the QA401 has worked out some kinks.

However hardware doesn't really matter much anymore. A vaguely good USB audio interface is more than enough for just about anything you would ever need. All that matters now is the software. Even something like measuring speaker impedance is just using a second channel to measure the voltage across a 0.1 ohm resistor. From there it's a fairly simple matter of plotting something like Vch1 / (Vch2 / 0.1). There's a lot to be said for anything that is actively developed and has a vibrant community like REW. Personally I never got into REW. The QA400/QA401 is great for quickly getting a measurement. I have very limited space for E. I usually keep everything tucked away in a closet. It's very nice to be able to break out that little box, plug it in and immediately get a sanity measurement of something. Incidentally, under Test Plugins > THD Versus Output Level and Frequency, that is a sort of "amplitude sweep".  The QA401 code is written in C# and the control API is listed at the end of the manual. But I don't see it has being terribly useful because ...

If you're going to write code, it's hard to beat Matlab / Octave. I've never been the math wiz but I know enough to know that Matlab / Octave is clearly the way to go here. Just record a 32bit float .wav file using one track for a ref-tone and that's pretty much all you need to get hardcore with Matlab / Octave. A lot of conventional purpose-built FFT software insists on doing FFT in real-time (eg QA401) which means the LF resolution is going to be pretty chunky. You really need something like Octave to sit there and average together more data to get good LF resolution. Someone should really create a little lib or maybe even something with a GUI (newer versions of Octave have a simple GUI API to make buttons and fields and such). That could be very accessible to just about all users. Octave runs very well on Linux and Mac. Never tried on Windows but I'm sure it runs there as well.
 
I agree, something like a Focusrite Saphire USB interface is more than you need to make audio measurements, the thing with the QA401 is that you get the hardware plus the software and you don't have to worry about calibration, all that has been taken care of, and for the price I think its a really good product.

I love MATLAB, its so versatile, the instrument control interface is not perfect, it will work sometimes and others wont, but having the flexibility of making your own graphs the way you want them and process the data in many ways is just something that is not easy to do with commercial software. For the HP8903a I tried using Pete Millet's apps, but it didn't work for me so I ended up doing my own code in MATLAB.
 
I just checked out the Q401 at their web site. It looks to be a vast improvement over the 400 and also over the earlier 410 version. Galvanic isolation from the PC, balanced ins and outs, accepts inputs up to +26dBV, frequency response -3dB at 70KHz, measurement bandwidth can be specified, frequency response and distortion sweeps plus THD & N and it can be programmed in C#. It cost less than £400. I am seriously tempted.

Cheers

Ian
 
I have revisited their website, and I think whatever improvements thay have made are not enough to convince me to fork another $450.
They don't have yet a programming language that comes close to that that Audio Precision delivered in 1985.
Writing procedures for AP is similar to writing html with a graphic editor. You don't have to write a line of code, but you can edit the code produced by the editor.
QA are introducing a pre-Beta version of something named Traction, that promises to do many things, as long as you're ready to write code.
That's what makes the difference between kids and big boys IMO.
Their newest product, a programmable 300W loadbox, for me is a sign that they are  considering the HiFi market.
Serious manufacturers produce amps in power that range from half a kW to 20kW.
ATM I'm content using REW and RMAA with the NI souncard and an interface that gives me all the facilities I need, input and output attenautors, +26dBu drive, source impedance selection, mute and CMRR modes.
I believe the ergonomics are as important as the performance metrics. IMO QA does not deliver.
 
Considering the QA401 is an order of magnitude cheaper than an AP I think it does a pretty good job. Covers just about everything a DIYer wants out of the box. I am sure a community cold soon generate some useful programmed additions if required. YMMV

Cheers

Ian
 
abbey road d enfer said:
Interesting. My QA401 has been sitting on a shelf for about 5 years now. my main concern were:
  • Impossible to overlay graphs, making comparison difficult
  • No sinewave sweeps (freq resp graphs actually derived from FF
  • No amplitude sweeps

Apparently all of that is now is possible, see the attached screenshots from the manual I took and saved to a pdf. Have you updated to the last software revision?
 

Attachments

  • Screenshots.pdf
    204.6 KB · Views: 16
user 37518 said:
Apparently all of that is now is possible, see the attached screenshots from the manual I took and saved to a pdf. Have you updated to the last software revision?
Thanks for that! I haven't updated my QA; as I wrote, it's sitting on a shelf since I don't know when...
 
Back
Top