On Class D Amplifiers Measurements

[pma]

20kHz square sent through
1) A250W4R class AB amplifier
2) NC252MP class D
into tweeter, 1V, acoustic spectrum measured by microphone. No comments.

 

[SIY]

Distance?

 

[Vuki]

Someone should prove that it's audible in blind test.

 

[SIY]

Someone should prove that it's audible in blind test.

Let's make sure there's an "it" first.

edit: Like here, for example.

 

[tmtomh]

20kHz square wave.

Given that this is an acoustic measurement of one transducer by another transducer, the results should be considered identical: the AB's 0.02% edge on THD and the Class D's 0.03% edge on THD+N are both likely within the margin of error.

That 40kHz 2nd harmonic looks to be about 44dB below RMS, which equates to about 0.63% distortion - which means that inaudible harmonic is contributing virtually all of the distortion.

The IMD spikes in the tweeter with the Hypex module appear to be at -90dB or lower relative to the RMS - yes?

It would be interesting to see the same test run with a Purifi-based amp.

And it would also be interesting to know the specs of this tweeter, namely its frequency range, and to see the test run with a tweeter whose frequency extension ended below 40kHz.

At any rate, clearly this is an urgent problem that the world's audio engineers must tackle right away lest humanity forever be deprived of true high-fidelity sound reproduction.

 

[SIY]

20kHz square wave.

Whatever it takes.

 

[dc655321]

That 40kHz 2nd harmonic

Isn’t that an IMD component? Square wave.

20 +/- 60kHz = 40, 80kHz

 

[DualTriode]

1kHz and its harmonics is where the ear is most sensitive, therefore designers prioritize this region for lowest distortion, and reviews show it as well.

Distortion above 10 kHz is irrelevant because:
1. Musical signal is way below full scale, so its harmonics even lower than on mids and bass; and
2. Harmonics are above absolute hearing threshold (20+ kHz) and below low level audibility.

Good old Fletcher-Munson curves say it all.

Some bad assumptions:

Harmonic distortions above 10k can and do intermodulation with other frequencies to produce distortion that is easily within the sensitivity of the ear.

Thanks DT

 

[Matias]

Some bad assumptions:

Harmonic distortions above 10k can and do intermodulation with other frequencies to produce distortion that is easily within the sensitivity of the ear.

Thanks DT

I show this measurement below to make my point, would you share a measurement to prove otherwise?

From the Purifi 1ET400A module datasheet. Note the power level (100W in a tweeter!), and lower than 120 dB artifacts.

 

[SIY]

Some bad assumptions:

Harmonic distortions above 10k can and do intermodulation with other frequencies to produce distortion that is easily within the sensitivity of the ear.

Thanks DT

Maybe you can point to it in the spectrum? Looks like all artifacts in the audible range are way too low to be heard, even with this crude measurement.

Putting aside, of course, the bizarre choice of test signal.

 

[Stinius]

If you look at the open loop gain of the nCore and the Purifi amps it's quite easy to see that the amplifiers are designed to achieve very low IMD 18k + 19k test. Change the frequencies and it it isnt that good.

 

[DualTriode]

Let's make sure there's an "it" first.

........

Maybe you can point to it in the spectrum? Looks like all artifacts in the audible range are way too low to be heard, even with this crude measurement.

Putting aside, of course, the bizarre choice of test signal.

Hello,

You are correct. I do not see any notable IM spikes in the plot.

It is credible that IM beat frequencies could show up in the audible range.

Today I placed an order for an AP AUX 0040 filter for testing Class D (switching) amplifiers. Two weeks delivery time is optimistic.

We will see what the plots look like.

Thanks DT

 

[SIY]

Hello,

You are correct. I do not see any notable IM spikes in the plot.

It is credible that IM beat frequencies could show up in the audible range.

Today I placed an order for an AP AUX 0040 filter for testing Class D (switching) amplifiers. Two weeks delivery time is optimistic.

We will see what the plots look like.

Thanks DT

In my earlier measurements, we saw absolutely nothing untoward. You don’t need the filter to do measurements of the tweeter output, but you do need lab grade mikes and the right interface.

 

[SIY]

If you look at the open loop gain of the nCore and the Purifi amps it's quite easy to see that the amplifiers are designed to achieve very low IMD 18k + 19k test. Change the frequencies and it it isnt that good.View attachment 196817

10/11 kHz looked absolutely clean. Ditto standard IM pairs.

 

[dc655321]

10/11 kHz looked absolutely clean. Ditto standard IM pairs.

Don't you usually perform a multitone measurement as well?
Or am I thinking of someone else...?

 

[SIY]

Don't you usually perform a multitone measurement as well?
Or am I thinking of someone else...?

Yes, I do. In the last couple articles, I also did multitone with a speaker load.

 

[pma]

I/O measurements of Hypex NC252MP step response/rise time and of input/output spectra with 20kHz square signal. Staying technical, showing parameters and behaviour. Not mentioning audibility.

Step response

20kHz square I/O spectra (interesting output noise modulation)

 

[DualTriode]

In my earlier measurements, we saw absolutely nothing untoward. You don’t need the filter to do measurements of the tweeter output, but you do need lab grade mikes and the right interface.

Hello All,

The AUX 0040 filter is on its way for testing amplifiers.

For testing the acoustic output of drivers I do have a APX 1701 interface and GRAS microphones plus piston calibrator.

Thanks DT

 

[pma]

Distance?

Same in both cases, same output level, same SPL. Do not blame it on distance. If it was a distance issue, the class D mess would be seen in class AB acoustical measurement as well. The of source was in acoustical tweeter output driven from class D. You need enough dynamic range in the acoustical measurement to see it.

In my earlier measurements, we saw absolutely nothing untoward. You don’t need the filter to do measurements of the tweeter output, but you do need lab grade mikes and the right interface.

I am sure you have not sent a fast analog generated 20kHz square into Hypex. And low level in case of acoustical measurements, with only 1V amplifier output. Please no excuses on "lab grade" instrumentation, though I know you are a technical expert (BTW - in chemistry?).

 

[pma]

The AUX 0040 filter is on its way for testing amplifiers.

You need it only in case that the input of your measuring instrument is HF level/SR saturated. And it is very easy to be tested.