Class D Design Issues

@Turbowatch2

Thanks for the advice. I have found a thread about the module I want to use with lots of information. I think I want to make a ‘Gremiln’ with my DAC inside the case so it is nice and tidy. I am interested in DIY audio and want to replace a Onkyo receiver which takes up too much room.
 
Class D...ease of built!

After reading post on Class D or digital amp, my point is that people are building them because of the ease of it... what I don't like about it is that there's always remaining noise. Just like digital picture, the very black or the absence of noise is very hard to obtain. A brief look at Douglas Self's amp notes and his very clear description of stage amp explanation is quite amazing. I builted his load invariant amp and it worked so well that I kept away from Class D amplifier. Self gives nice trials of Class A, B and his Class XD that I might built very soon!
 
Last edited:
Class D...noise left!

Now a days, audio issues are not first concern of every body. Majority of people feel satisfied with MP3 compress audio that may fit well with Class D amplifier. The quality audio level over what you're sensible to or can afford will guide the way you'll see, or build audio. My mother always told me what you don't know won't hurt you... but I know about hi-fi listening and thank God! Unfortunately I never passed a Class D amp under a Spectrum Analyzer but when I have one I'll do it. To resume let's say that the environment surrounding you will lead you to a choice on audio level.
 
If there is noise pcb design is bad or the design itself.

Feedback network can trow noise.

I have not yet build one but I will do and see what happens.

Always decouple driver ic, comparator, mosfet supply with caps, as close to the pins as possible, use separate ground plains for analog and the pwm/digital part.

use good comparators, fast ones and mosfets with low nC or go multilevel.

have a nice evening.
 
Disabled Account
Joined 2005
After reading post on Class D or digital amp, my point is that people are building them because of the ease of it... what I don't like about it is that there's always remaining noise. Just like digital picture, the very black or the absence of noise is very hard to obtain. A brief look at Douglas Self's amp notes and his very clear description of stage amp explanation is quite amazing. I builted his load invariant amp and it worked so well that I kept away from Class D amplifier. Self gives nice trials of Class A, B and his Class XD that I might built very soon!

Make a class D is not easy and that is not true class D has noise. The truth all amp has noise even very small. So does class D. Douglas Self precise at that time he wrote. Now a lot of improvement in class D and the most contribution from better mosfets available also chips and discrete parts.
 
Using multiplevel do help, I did test a three level then the switching frequency is multiplyed x 2, and noise is very low..

Not yet build in real live, I am now to busy with the analog amp.

The example pic is only prepost feedback but have also post feedback with both.

regards
 

Attachments

  • ScreenHunter_2497 Aug. 10 12.46.jpg
    ScreenHunter_2497 Aug. 10 12.46.jpg
    382.6 KB · Views: 527
Last edited:
irs2092s bad sound

hello,i have build a class d amplifier based on irs2092 buffered with totem pole,mosfet pair of irfb4227.i took the schematic in iraudamp9.please,is anyone can tell what make the amplifier sound like BREK BREK.in general,what may cause that.i build it with exactly like in the schematic,part value and so on.thank for all:(
 
Hello,

Regarding the worry about high frequency noise coming out of a class D amp.

I only have experience with the NCOR-400, but the RF coming out of it is so limited there is no noise whatsoever. What little there is is in the millivolt range and the inductance of the cable, crossover network and speaker coils will not let it ever be an issue.

A well done class D design has no issue with RF. I am told RF problems are not uncommon in some of these amps, but a quality one does not have an issue with RF. FWIW IMHO.

Regards,
Greg
 
Can someone explain me why the output filters of most class D boards consist of smd inductors and capacitors while if the amp is followed up by an analog crossover, the coils and caps used there are absolutely enormous in comparison.
Are the output filters just under spec'd with cheap class D boards, or is the average DIY'ers analog crossover network just overkill?
 
Or use this a multilevel class D. It goes in this case to 1.4 Mhz.

Only small lowpass then or even not, just use the coil speaker or filters.

Or using a aircoil because it is small, never trouble with non lineair behavior
and can use prepost feedback.
 

Attachments

  • ScreenHunter_2027 Feb. 08 22.17.jpg
    ScreenHunter_2027 Feb. 08 22.17.jpg
    307.8 KB · Views: 535
Last edited:
The frequency the class D filters have to operate on is much, much higher than the frequencies an inductor in a speaker xover operates on.

The higher the frequency, the lower the required inductance for the same filtering effect. Lower inductance = smaller physical size, all other things being equal.

Jan

That makes a lot of sense. But does this also hold up with wire gauge? Those smd inductors seem to have hair thin wire.
 
pulse transformers

What do you have in mind? Pulse transformers have been used to drive FETs (etc) in PWM applications, especially where high voltages are present. There is a technique where a series FET is used to store the gate pulses because, of course, transformers do not couple DC. Pulse transformers do not couple audio frequencies either. And, also of course, there is no point in using a high frequency transformer on the output. By design, the FETs see a very high impedance at the switching frequency that the output filter input choke provides, and passes, ideally, only audio frequencies. As the FET totem pole switches +Vcc and -Vdd, the output choke charges and discharges slightly about the audio signal current value. The FET conducting less than 50% duty cycle is actually dumping current back into the supply as the output choke discharges.
 
Last edited: