I was wondering if there was more info on multitone as it pertains to the QA40x. Have read the AP note on it and several threads on ASR, but would like just a bit better understanding of what it is telling me. For instance, I got this from a receiver I just tested:
I’m reading this that I have about 55dB or so THD+N ? The normal 1Khz THD +N is ~70dB. I am pretty sure that if I selected either THD or THD+N to be displayed during this multitone test it would not give me accurate results ? Thanks for any input…
Hi @Var, yes, you are right that certain measurements don’t mean much when using multitone as a stimulus. The primary reason to use multitone, IMO, is to see if you can tease out complex IM products that might not otherwise show up because they occur at odd places in the band. In the plot you shared, note the “noise floor” appears to be around -60 dBV. In reality, that probably isn’t noise. It’s a flood of IM components.
On page 21 of the QA403 product brief (linked below) you can see the multitone loopback performance of the QA403. Note that this is a 512 point FFT and that the space between the harmonics extends below -140 dBV. That is what “very good” IMD performance would look like. That is, the level between the discrete tones is very, very small.
On Class D amps, it gets hard to measure linearity with just a tone at higher frequencies, because these amps have output stages that intentionally limit the energy. A multitone test (or even a dual tone test) can reveal things that a normal single tone test in class D. And the multitone takes that to an extreme and does it in one shot.
In short, multitone is a torture test for amps, and it will quickly let you see if the amp has a linearity problem or not. In production, you could hit your amp with a multitone, and know ahead of time that the level between two tones should be below -100 dBV for example. And then spot check a few places across the band without taking another measurement. If the level between tones is where you expect it, the amp linearity is probably good.
Matt-Thanks for taking the time to explain a bit better. I looked at the Q403 multitone loopthru and almost understand it (I think). THe QA403, with IMD products <110dBC across the band, would be pretty good, particularly when compared to my receiver which had IMD products <55dBC or so. What would be a good figure of merit is for multitone for amps (non class D) ? Also, in your Q403 example, you specified a level of 0dBV, but the plot shows the multitones starting at about -18dBV ?
One issue I have with the typical multitone test signals is that the harmonics and IM products all too often are buried under source tones. I wanted to attach a spreadsheet of sets of tones that don’t overlap but its not allowed. I’ll e-mail it to Matt. It turns out this is not a simple problem. Its kind of like prime number sequences.
I run a 1024FFT on it with QA403 set 18dBv and 96Khz. I read the answer above, so because level between two tone is less than -100dBv amp linearity is good ? I’ve difficulties on how interpret these measure…
And If I set multitone generator above -5dBv my amp ‘crash’… It’s specified for 20W rms on 4R load.
I would say that your amp looks pretty good. I use the dBr mode and have it set for 0dBr at the top of the scale. If you ignore just of few of the peaks, your noise is around -110, which, if you divide by 6 gives you about 18bits of distortion free range, which is great for an amp. I typically run the test such that the amp is not putting out more than about 5w/8ohms, and my FFT size is 256k and 96k sample rate. Here is a recent capture for a vintage SS amp I tested, which is rated at 200w/8ohms- this is one of the best multitone test results that I have measured.
Hmm ok, glad to read my design is prety good Actually i’m designing amps starting from copy paste from Pass open design and now with my own idea. I’m not a audio professional, this is my hobby since almost 10 years, nevertheless I’m a professional electronic designer working in several field and company. So… let’s the interesting thing:
the noise floor:
So how is linked QA403 SW setup output level (setting in dBv…howw man I know you are labeled as audio but… why using such metrics ? ) in dBv to output level ? Why when I set -5dBv as QA403 outputl level I just don’t have the same thing than when I did a single frequency measure, example below with QA403output set to -5dBv:
From the testing I have done, the Multitone test results don’t seem to vary much based on the input signal level to the amp or preamp, as long as the output is not overdriving the QA40x…
Hi @Briks, the multitone amplitude settings can be a bit confusing. In the plot below, I have specified the multitone level to be -50 dBV. That level is the RMS level. And as the measurement shows, that is -49.92 dBV, so very close.
Now, you can see that each tone is about -68 dBV. But all the tones sum their power to give the final -50 dBV target. The more tones you specify, the lower each tone will be. But the final sum will always be the target you specify.
Next, I repeat with one tone per octave. Here you can see the RMS is very close to the -50 dBV target. But there are many fewer tones. And each tone is now around -60 dBV, and they all sum together to meet the -50 dBV target.
Hi Matt, thanks for explanation! I didn’t know playing multi tones get more power, but thinking about that I understand, I visualize a low frequency tone and a high frequency tone playing same time. HF tone + LF tone will results in a more powerfull signal in terms of dBv than playing only HV or LF with same level. Right ? I mean for example a LF will be peak peak 10V, this LF + HV (with same amplitude) will results n a 20V pp ?
For looking at IMDistortion I like to use several different two-tone FFT’s. You can identify individual Harmonics of the test tones plus identify the IMD side bands.
The 30 tone tests have a lot of stuff going on and it is difficult if not impossible to see individual harmonics and sidebands. There is just a lot of stuff.
It’s specified for 20W rms on 4R load.
Actually i’m designing amps starting from copy paste from Pass open design and now with my own idea. I’m not a audio professional, this is my hobby since almost 10 years, nevertheless I’m a professional electronic designer working in several field and company. So… let’s the interesting thing:
