QA40x: PWR THD versus Frequency

A new plug-in is getting added to the 1.211 release that should be out in the next day or two. This will create plots at a fixed power level such as this:

The plug-in is called PWR THD versus Frequency. The setup appears as follows. Note you can specify the target power level you expect. A “probe” will be made when the test first runs to learn the system gain, and this will be done at a level 40 dB below the level required for the specified power with the full scale input set to 42 dBV.

After the probe is complete, you’ll see a dialog allowing you to sanity check the test parameters. Here, you can cancel or select OK to have the tests started.

3 Likes

That’s truly excellent. Can’t wait to try that one out, it will save all the back calculations I’ve been doing for target power output sweeps. QA40x software is getting better and better!

I see the FFT size/time being reduced for the high power tests- is that a manual thing we need to remember (so we don’t cook loads) or does the plugin ‘remind’ us?

Hi @restorer-john, the plug-in doesn’t have any smarts about the FFT size needed. You’ll need to pick an FFT that will ensure the harmonics are above the noise. Otherwise, the THD algo will just use the level for the noise at each harmonic which is overly pessimistic.

The Hypex and Purifi are delivering really impressive numbers at all power outputs, but especially at the 1-10W regions. So, you need the larger FFT to really scrub down and find the harmonics in the noise. As the power level comes up, the harmonics come up and they needn’t be as large.

Of course, if you are opting for THD+N, you don’t need to worry about digging out the harmonics as everything is treated as noise. For THD+N you just need an FFT large enough to give the noise performance you need around 20 Hz (too small FFT lets the DC offsets bleed into >20 Hz region).

I will do a writeup on measuring the Hypex soon, and comparing to the published spec sheets.

1 Like

Hi @matt. I also find the new plugin extremely convenient, as it allows specifying the power, and the plugin itself takes care of the tedious calculations for the required voltage levels after determining the gain of the DUT. Great job, Matt.

I would love for the same principle to be applied to the PwrThd Versus Power plugin. In other words, I would like to have the ability to specify either voltage levels (in dBV) myself or to define minimum and maximum power levels and let the plugin handle the necessary calculations, just as in the new plugin.

I would love for the same principle to be applied to the PwrThd Versus Power plugin.

Hi @Claudio, yes, this does make sense as I think anything in the PWR realm of automated tests should probably let the user think purely about power. The question is how we might specify. We could do start and end power, and points per decade?? The current dBV range doesn’t make much sense and/or isn’t very convenient.

Hi @matt, yes, I think using the initial and final power as parameters and specifying the number of points per decade is the best solution. After all, when using this type of plugin, you always consider at which initial power to start the measurement and at which final power to stop it. Then, everything is converted into the voltage to be applied to the DUT in dBV based on the gain through calculations. The step in dBV is then specified, but this operation can be effectively replaced by indicating the number of points per decade (of power) and leaving everything to the plugin’s preliminary calculation. I think the solution you propose would work very well and would be a great step forward in simplifying the use of the plugin.

1 Like

I think this will be a great addition to the QA40x toolbox! Definitely will save some time with how I do things. My question is, will the “probe” gain level that is measured at 40dB down be taken at just one frequency, or several frequencies and the required level be adjusted for that frequency? May not be a problem for a lot of newer amps, but for older gear the variation could be larger (say 1dB)

1 Like

I would expect the “probe” would be at 1kHz and yes @VAR some of the old amps roll-off pretty severely at each end of the spectrum. Maybe the ‘probe’ frequency can be adjusted or is based on the sys gen setting?

I just hope Matt posts it up soon- I want to try it out. :slight_smile:

I would think that doing a quick chirp Frequency Response on the device and using some of those data points for a gain look up table would not be too difficult, but I am not writing the code :sweat_smile: It is a measurement that I do often so am really looking forward to it.

Hi @matt. I downloaded version 1.211 and did a preliminary test of the PWR THD versus Power Watts plugin right away: it’s great to no longer have to do any calculations to be able to have THD as a function of power! Great job Matt. Only one small problem I noticed by testing a buffer (G=1) and specifying the upper power limit at 10W. The preliminary test reports that for that limit the generator will be set to 19.1dBV. Obviously the QA403’s generator will not be able to supply that voltage and so giving “Ok” to run the test will inevitably cause it to crash generating an exception that is reported. In my opinion, it would be a good thing to prevent the execution of the test by signaling that the parameters to be set in the generator exceed its possibilities (and then inviting the user to change the parameters). Of course, this is just my opinion

1 Like

I ran the PWR THD vs Freq and am having a high max power error, which I don’t think should be the case. The power level value it calculates at 1kHz is right on. Here is the amp’s frequency response with it putting out about 5w/8ohms:


Here is what I getting running the new automated test:

The max power error shows 75.9%, which I think is incorrect. I tried both the RECT & HANN windows, as well changed the FFT from 128 to 64.The Sample rate is 96k. If I manually go to a frequency at the band ends the power reads what you would expect. The one thing I noticed is that when manually going to a frequency it shows up very quickly but at 1/2 the power level for about a second or maybe two before it is reads the correct power…

Hi @Var, the first thing I notice is the generator is at 20k and yet the last point plotted is around 12k. The THD didn’t plot, probably because the 2H of 12K is > Nyquist. But I think the power error was still being considered.

Can you set your Stop Frequency to 10k and see if the power error is more reasonable?

Probably the test should look at user’s choice for Stop Frequency, Nyquist and 2H and warn accordingly. And also not track power error for data points that weren’t displayed due to them not have a 2H below nyquist.

Very good feedback, thanks very much!

Hi @Claudio, yes, this is a very good find. It think more error checking on all parameters before proceeding. Will fix very soon.

Hi @Var and @Claudio, I think your issues have been fixed for 1.212. Please see if this release HERE solves your issues when time permits.

Thanks again for reporting these!

Hi @matt. I downloaded version 1.212 and retried the test that generated the exception (exceeding the generator capacity of QA403 in PWR THD versus Power Watts plugin) for me. Unfortunately, everything remained as before: I attach 2 screenshots of what happens:

In my opinion, after this message, you should inform that the 10 Watt option cannot be executed as requested and then ask to change the options or alternatively warn that the generator will be set to its maximum voltage capacity and consequently the maximum power will be stopped at… Watts . Obviously this is just my thinking

P.S. I realize now that in the version notes (1.212) it only refers to the plugin “ PWR THD versus Freq” while I tested and detected the problem on the plugin “ PWR THD versus Power Watts”!

My apologies, yes, the change was only made for Pwr THD versus Frequency but you are right it needs to be made for PWR THD versus Power Watts too.

Version 1.213 is up, and in your case of 8 ohms and 10W with 0 dB gain (basically single-ended loopback) on PWR THD versus Power Watts you should now see this error:

Release HERE

Hi @matt. I downloaded version 1.213 and tried it: now, as far as the reported bug is concerned, everything is fine! Thanks again Matt

@matt - I tried the new release and it is still showing a large 76% power error.:

Here are the settings for the plugin:

Here is a screen shot I made while the test was running and it shows two frequencies which go across the display, with the power down about to 2.5w instead of 5w- not really sure if that is normal:

Here is what I measured manually at 20kHz. The power is down a little bit from 5w as would be expected based on the amp’s frequency response I showed in an earlier plot. The THDs for both channels agree with the plot, which is the most important thing:

Hello Matt,

Sorry to be a little bit not so excited, to the annoucement. But while I read a lot of interesting posts, there is a subject I have never seen adressed : what I can do with my QA400 ?

Although spec wise it’s sufficient for me testing (old) Analog audio gear, the total lack of continuing support makes me ask a reasonnable question, what should I do with it ? trow it away ?
As in there no futur with it ?

Making some of the simplest measurements (aside spectrum analysis) like bandwith versus phase analysis is much simpler with BODE PLOT on my scope than having to convert a csv to see the same results… and I am not even talking about THD versus frequency…

My Question is: have you got an automated solution (simple) for easely plotting / showing nice graphs of your QA400 csv file when doing a THD vs Frequency run with the dll included in the original product?

Any thoughts ?

(Trowing away your first product is against my way of doing, and the always the latest and greatest or did I wrongly bought too early that product ??)

Best,

Eric (too early adopter ?)

I @Eric, I think the $199 QA400 first started shipping in China in 2014, so it’s an ~11 year old product at this point. Obviously, it’s not possible to continue development forever. I have a drawer full of CODECs from Roland and others that no longer work. I can still use them as Windows audio device, but the low-latency ASIO drivers and apps to unlock their audio routing and DSP effects have all gone stale. My phone is a Pixel 7a, and when I bought it, the promise from Google was 3 years of updates, and then 2 more years of security fixes. I have Windows 10 machines all over the place, they can’t be upgraded to Win 11 (due to lack of TPM) and Microsoft has said “that’s all” after 10 years. Not even security patches. We bought a new $12k UV printer last year because the software evolution on the 2018 printer had stopped. The $8k 2018 printer still works, but it couldn’t keep up with the printing speed from a new printer.

Now, none of the above is ideal. But it’s the reality or product development, unfortunately. At some point, the older platforms will move to bug fixes only, followed by security fixes only. And then they are frozen in time as Google et al decide make sense from a business perspective.

Sorry the news isn’t better…