Self Biasing Class D Amp

Output filter is ringing cause its made for 2 Ohm and not for 4 Ohm

Interesting. I see you used an integrator to self bias it.

Do not forget why this is done. Its a standard methode in bigger amplifier, also in analog amps. Its not importante how the bias is made at input. Importante is to avoid DC at the output!!! So the problem is DC at output and the solution is to average the output signal with a huge Tau, much below 1/20Hz, and feedback it to the input, and this is best done with an integrator/lowPassFilter. Here we have a capacitor at output and can only expect DC in case the cap is damaged! In case we have NO cap its getting interessting for such a DC offset elimination! E.g. for a class d in bridge mode. Again, the problem is not the DC at input, its the DC at output.

No, the filter is for 4ohms but the cutoff frequency is 30kHz, not 20kHz. Also yeah, a realistic approach would make the time constant of the integrator closer to DC, but simulation would take forever to settle. Anyway, a better approach is a split rail version with a 0V offset at the output, but I decided to try out a single rail version - they're always harder to design.

http://everycircuit.com/circuit/6398919209910272 Here's the split rail version. The output to the load can be wired up to a fault detector of some sort to negate any DC going into the load.

Your LC filter is for 2 Ohm and not for 4. If you ever made a class d amp irl you should know. Your filter with 21.2uH,1.36uF and 4 Ohm does cause 17% overshoot in voltage. Test it with 2 Ohm, which is correct and the overshoot is gone. Your filter is underdamped! http://everycircuit.com/circuit/5013854466867200 you probably trust to much the formula application you are using for calculations.

17% overshoot doesn't actually sound that bad, I think it's within boundaries. Anyway, a typical loudspeaker is a much more complex load than a simple resistor, so variations throughout the entire bandpass are to be expected anyways. I don't think the filter here is such a critical issue. Class D amps suck anyways, however designed

Statements like “Class D amps suck anyways“ disqualifies you totaly.

Well they do suck. The only thing they're good at is Efficiency. Other than that everything else sucks. I don't like them for numerous reasons

No sureprise if you accept underdamped filters.

The filter is irrelevant for the most part. It is actually the one problem with such amps. An output filter lacks the dynamic characteristics of a linear amp, it completely destroys transient response, and for the most part it's only useful for recreating sinewaves. That's why class D amps looks unrealistically good on paper, but sound flat and metallic in reality. No matter how much you dampen the filter, you're driving a speaker, not a resistor. This is bound to cause problems with any LC filter, as it would cause it to oscillate in a given part of the frequency range, and overdampen in another part of it.

Just take a look at your own circuit. Is this how a square wave should look like at the corner frequency of the amp? http://everycircuit.com/circuit/5041770059792384

For me, you can't extend one property of a circuit, without compromising another.

Plus, I'm an audiophile and I don't like how class D amps sound

A rectangle should look like this, right and dont have overshoots http://everycircuit.com/circuit/6269238376660992 why i have the feeling it doesn't make sense to have a discussion with you? Its like years ago were you hate amplifiers with feedback?

I'm talking about the cornee frequency, not midrange. Furthermore the delay in the filter, makes transient peaks become compressed. Basically class D amps work best with sinewave. Try to reproduce a complex audio signal, and it becomes all jumbled up.

Anythink we humans can hear is made out of sine waves from 20Hz upto 20kHz. Any sine waves frequency above or below is useless. How does this “complex audio signal“ looks like and can not be made with 20 to 20kHz components but can a human hear a difference? Answere, this wave form does only exist in your phantasie!

Try to play classic music on a class D amp, and see if the dynamic response is satisfying. Some things look good on paper, but don't sound good in reality.

Statements as usual and nothing you can proof. Again, how does your “complex audio signal“ looks like? Bring us an example and stop this blahh blahh.

@thebugger, I looked up online whether class-Ds sound bad, and one place seemed to say that no matter what every amp from class-A to class-D will always sound different compared to another amp, even if it’s the same model provided by the same company. And it also said that some amplifiers can have an external preamp added on, with the internal one disabled, because the preamp is the source of the differences because the power amp is always transparent. One guy said that he has gone through several amplifiers in some studio or something including at least one class-D and the class-D was no worse than any of the others.

No, power amps do have some noise, but they usually have very low voltage gain, so a very deep negative feedback usually makes them transparent. An open loop power stage would have as much distortion as any other. I'm talking about something else. Every technological advancement so far had its ups and downs. Tubes were huge and inefficient, but were easy to work with and very robust. Some military applications and space stations still use tube due to their radiation resistance. Transistors came in and 2 viable options were possible - efficient switching and more-than-tubes efficient linear mode. The two options are so different you cannot compare the sonic qualities in any way between them. For instance in class D amp, you cannot talk about classic noise, because as with any switching system there is no noise - there is jitter and similar.

Blahh blahh blahh blahhh. This sounds like a horoscop, but nothing what a scientist would say. Were is this “COMPLEX audio signal“ ????????

Look, you're acting like a little whining bitch. It's a matter of personal choice. I for instance like class B transistor and tube sound if well designed. Some people like class A sound, because it has specific harmonic richness. I see you like class D amps - that's fine. Try listening to very dynamically rich music like classical music on class D amp and notice the flatness, then we'll speak

You do not know what i like. But what i dont like are statement frauds like you. Just progf one of your statements you made here. Why i already know at beginning of this thread it ends in your foolmouth arguments and is waste of time. Its buggzy and he wont change. Instead of being rude and nauty bring us just ONE little proof. Use numbers and facts and circuits and try to avoid this esoteric audiophil blahh blahh

How does a “complex audio signal“ looks like and can not be amplified by class D amps and only from class A B AB amps??????

No, this stuff has already been explained in length in research documents online. I'm not your personal google to explain it to you.

And as you already know a complex audio signal is not comprised entirely out of sinewave. It has different harmonics and such, which for a class D amp might come in as a compromise. Either a broad audio bandwidth, by choosing a higher cutoff frequency, which means you need a higher switching frequency to not have troubles filtering out the switching noise, or a lower audio upper frequency, and a sharper filter. There's also one more quality a class D amp can't match. Transistor amps tend to have almost negligible output resistance in respect to the load, while class D amps have ab output filter with a considerably high characteristic impedance in respect to the load. You do know what this means do you? I really don't want to go into lenght, considering you well damn know what that means, but buzz me if you want to and I'll explain

Wrong. All audio signals are a mix of sinewaves from 20Hz to max 20kHz. Just sinewaves and NOTHING else.

Yes, but not from 20Hz to 20kHz. Audio signals can contain square wave shapes, which have endless harmonics, so that's why most Hi-Fi amps have a 70kHz upper cutoff frequency. A 70kHz filter would need to compromise much of the filtering strength for a class D amp, so a higher switching frequency needs to be used. This leads to numerous problems, like jitter or even simple driving issues.

And you can hear 70kHz? The higher cut off for class D is, what you do not know, also a common thing todo, but not because humans can hear that. Again, all audio signals are made with frequencies from 20 - 20000 Hz everything else is audiophil esoteric and not scientific. Can you hear 25000 Hz? If so then i will expand the range extra for you to 25kHz. Question, what do you hear if you listen to a 20kHz rectangular signal? Answer, a 20kHz !sinewave! followed with 1/3 60kHz sinewave and 1/5 100kHz ..... maybe you can hear 20kHz i doubt you can, anyway 60k and all above u definitive can not! So the range is 20Hz to 20kHz. What do you hear for a 7kHz rectangular signal? 7kHz sinewave and nothing else, cuz 21kHz is already to high for your ears.

Look up overtones

Look up idiot

Don't need to look it up, I have a good example here

Thanx for this collection of metallic, flat, transparent, deep feedback, dynamic-less, complex audio signals which are not made with sinewaves, audiophil BULLSHIT!

@thebugger, did you even understand a single word @hurz said about us not hearing any frequency components too high for us to hear? Or do you not believe that every repeating signal is composed of sine waves of different frequencies, amplitudes, and phases? If you do not believe this then you clearly haven’t studied up on the subject and have no right to act as if you have. If you put a signal through a sharp 20kHz filter (or lower if you don’t have perfect hearing or higher if you have exceptional hearing) then what comes out on the other end is what our ears hear. Try an experiment: run a 7kHz symmetrical (as seen with an oscilloscope) square wave on a speaker (or higher frequency if you have exceptional hearing and can hear the third harmonic) and then add a 10kHz low-pass filter between the speaker and the audio source and see if you can hear a difference. If you can, then show us your setup, perhaps with a link to imgur.com. If you can’t, then I’m sure both me and @hurz want to hear you try to explain why. Until you’ve done this experiment, you have no proof for any of your claims and you should not share them as if you do have proof like you are doing right now.

No response?

As I said, loom up overtones. Upper harmonics (above 20kHz) are inaudible but play important role in shaping the base frequency signal. This is where destinct timbers come from. Complex musical inatruments require sophisticated systems that can handle upper harmonics as well, or they sound dull. Just play some classic music on a class d amp

Right, you sound dull.

Show us how these inaudible harmonics shaping base frequencies?! How do they do that and how do they change a 7kHz sinewave from my example above?

BTW, all the audiophil idiots who store music in raw format, which means 16bit resolution and 48kHz sampling rate are total idiots because they just capture music in best case upto 24kHz? Damn they are all idiots, thats what you are saying! Buggzy really .... ... .... ..

http://everycircuit.com/circuit/5171953337106432

Again a nice example of buggzys ignorance and ARROGANCE about topics in electronic he knows nothing but must look cool and all known. Being rude is your answere.

No response, here your audiophil phrases. Thanx for this collection of metallic, flat, transparent, deep feedback, dynamic-less, complex audio signals which are not made with sinewaves, inaudible upper harmonics which shaping by magic the base frequency, destinct timbers, complex musical Instruments require sophisticated systems that can handle upper harmonics, or they sound dull, audiophil BULLSHIT!