0.0004% THD Class D Module

[Lars Clausen]

Using the (rather radical..) new measurement methods as proposed by Per Anders, i hereby post the first measurements of THD on the 700XE.

I guess the serious message in this result is that pre filter NFB can give as good THD measurement as post filter NFB.

Take a look below

 

[peranders]

Did I? Forgot it already

How does it sound then, the obvious question? Pretty certain that you at least are pleased, yes?

Do you have a more revealing picture than the one you have on your website?

 

[KLe]

Hi Lars

Excellent THD results, and what, there for a pre filter NFB design!!!

As peranders asked, "How does it sound then ..."

Lars, the graph only goes to 10K, what's the THD from 10K to 20K?

 

[Lars Clausen]

Hi KL

Well the THD between 10 and 20 kHz is usually eaten up by the 24 kHz 'brick wall' filter.

For example if you take a 3rd harmonic THD of 10 kHz, the harmonic will be at 30 kHz, and simply vanishes in the AES filter.

 

[BWRX]

Just curious - what kind of load was used and what output power was the test done at?

 

[Lars Clausen]

This is 2 W rms in 8 Ohms

 

[IVX]

Lars, can you show us THD/Power plots for 200Hz, 1KHz, 6KHz? Do you know about how much the loop gain in your amp? What is the (rather radical..) new measurement methods, really?

 

[Lars Clausen]

IVX: Kind of a joke between me and Per Anders. From another thread.

And please note, the posted THD data are real, but NOT official data release.. I just posted the best looking results i could get (under the most beneficial op conditions .. but not totally useless anyway).
Of course a more parametric result will be released later on.

 

[TerryG]

This is the problem I have found with measurements, if not measured with a true inductive load then your not getting real world results. If you can get below 0.01 with a real speaker including crossover, then you have something.

Why does the brick wall have to be at 24k Hz, can't that point be higher and still not have a great deal of noise? And if the noise is above 20K why do we worry about it so much, or why is it even audible? These are honest questions here. The best switch mode amp I have ever heard was noisier than most, but during music levels you didn't even notice the noise.

 

[Lars Clausen]

measured with a true inductive load

.. or even a loudspeaker ..

 

[jeroen_d]

Why all this attention to THD? The numbers are nice though!

But what about IMD? Way more important in relation to perceived sound quality.

 

[IVX]

I have even lower THD on the loudspeaker, then on the 4om resistor, at least in highs.

PS: Reducing THD by the huge loop gain, IMO, is not so much intelligent approach. I have said "STOP" to myself, when could not provide flat THD/Freq more, it's occured at 50db loop gain and no more then 500-550KHz idle switching, because i still prefer cold amp. But if i could develop the discrete amp with 30db loop gain and .005% THD at 10W etc, i would be really proud.

 

[Pabo]

Lars Clausen

How do you measure THD and do you include noise? If the noise is not included it is not uncommon to reach those levels but if the noise is included I am impressed.

 

[Pabo]

IVX

What do you really mean by loop gain of 50dB? Do you mean the ratio between the modulating signal and the supply rails? If the loop you are using is hysteresis modulated I assume that you are using an integrator and this will naturally give a loop gain that falls by -1 or -2.

 

[IVX]

Yeah, integrator in use of course, i tend to call it PID even.

 

[Lars Clausen]

IVX: Have you published your circuits somewhere? I am curious to see how you get your results, and how you calculate your loop gain.

 

[IVX]

Have you published your circuits somewhere?

I have not such plans so far, though it's would be simple (TDA8939) and no bad amp to build by DIYers.