Class D Technology

mapman" Hadn’t thought of DSonic in a while but looks like they are still offering a lot of bang for the buck.
Nice website.....lots of useful specs and measurements.
What Class D modules do they use these days?"

Hello mapman,

     On their current M3a-600-M monoblocks, I know they're using the Pascal Pro2 class D modules, which have the smps built into the module and are the same ones used in the Rowland Continuum 2 integrated amp.  
     I have the slightly older M3-600-M monoblocks and they use the Anaview/Abletec modules.

Tim

@timlub What is the 'switching distortion' to which you refer?

Obviously No Switching distortion.

Both the switching distortion and non-linear should be addressed. True, the output devices are switching, however, the output devices are unlike other amplifiers as they are not operated in the linear region, which is not actually linear, but operated in saturation. That is a good thing. There is a lot less variability in the saturation region, which leads to less distortion, not more, from an architecture potential standpoint.

While technically distortion, the filter is not removing distortion in the the sense of a standard linear amplifier which would be distortion in the audible band, but harmonic products above the audible band.

Obviously No Switching distortion.

I also understand the weaknesses somewhat. 1. Non-linear & lots of distortion that needs to be cleaned up with an output filter.

Hello atmasphere,

     As I understand class D technology as a layman, timlub could be referring to 1 of 2 different possible types of distortion:

1. Distortion caused by the output transistors not turning on and off quickly enough- Most current class D amps utilize MOSFET transistors that have a limit to how fast they can switch from completely off to on and vice versa.  Any lag time between the transistor being fully in the on or off state is termed 'dead time' and can cause signal distortion.

2. Distortion caused by the carrier frequency being too low- This is a possible form of distortion that georgehifi claims is audible but has never provided any valid evidence to support his claim.  Basically, he claims the ultra-sonic frequencies (typically in the 300-600 kHz range), that class D amps modulate to carry the audio signal and subsequently completely filter out the frequency leaving just the signal, are too low and somehow affect or distort frequencies in the uppermost audible frequency range.  He claims carrier frequencies need to be above at least 1 mHz in order not to affect or distort frequencies in the audible range.

     I've never subjectively perceived any distortions or sonic artifacts on the many class D amps I've owned or heard.  However,I know some amps are beginning to utilize the newer, ultra-fast switching GAN output transistors and I believe at least 1 company, Technics, offers a class D amp model with a carrier frequency above 1 mHz.

Tim

@noble100 

OK- I just wanted to be clear on what was meant by 'switching distortion' since that's not an engineering term at all!

Artifacts from dead time show up as harmonic and intermodulation distortions, so a class D amplifier can easily be compared to other amps on that basis. IOW, on its own 'switching distortion' has no meaning. 


With regards to the switching frequency, with the current state of affairs with switching devices, its actually easier to make a low distortion amplifier when the switching frequency is less than 1MHz. This is because the amplifier has to meet certain standards for RF radiation- various countries have specs for how much RF noise can be tolerated. The faster you go, the difficulty in getting the amp to behave in this regard goes up on an exponential curve! Its quite a feat that Technics was able to switch so fast, but in reading their ads for the amp, I suspect a bit of marketing hype is also involved.


The problem with a low switching frequency is probably that of bandwidth, but class D amps can have a way of getting around that, since it is possible to add enough feedback to allow the amp to compensate for phase shift that would otherwise plague an amplifier with bandwidth from only 20Hz-20KHz.


In a nutshell its really dangerous to make blanket statements about 'all class D amplifiers' just as it is about 'all tube amplifiers' or the like- there are often many exceptions.