What Class-D amp give best sound?

Tpa 3251 or tpa 3255 from Aiyima seems to be very good.
Better than many class A/B amplifiers.
I can recommend the capacitance multiplier from Allo to clean up the power from SMPS . Its a stand-alone card. I use two of those in my active loudspeaker system, with my two Aiyima a04 tpa3251.
With the card, the sound gets slightly better with better bass and a slightly bigger soundstage.
 

Attachments

  • 0DF0531D-DF35-453D-A4EC-5965CA6262BA.jpeg
    0DF0531D-DF35-453D-A4EC-5965CA6262BA.jpeg
    91.2 KB · Views: 876
Last edited:
Resonance smps are quite more cleaner even without card.

Impedance of supply is the important thing, special for bass tigtness and
briljant highs.

But I am not active for a while now because of programming order, but I am just ready so I can go back to designing amp.

regards
 
Impedance of supply is the important thing, special for bass tigtness and
briljant highs.

Indeed so in my experience also. Important to note that classD chips in general have more than one set of supply pins - those most sensitive to supply impedance are the lower power ones feeding the signal stages.

I'll just give a shout-out to Infineon's MA12070 which gives a very good sound when its signal supply is given a low enough impedance.
 
The chips does draw not much so clean these up is not difficult, but important,the output switching part does has impact, for Class D it is quite inportant that there is a good supply rejection otherwise it get bad, more bad then with analog amps.

A class D open loop does sound bad because there is no rejection, a smps supply on a 100 Khz switching frequency does better then a analog supply who has much more ripple and at 100 Hz.

I do now still the circlotron analog stuff but has in past did a version three level who is free to build for everyone here, no time myself but curious about what it does.

For the chip to get clean supply, also the opamp part you can use a regulator, a good one who can de adjusted with resistors like LM317 has deep rejection, good caps close to pins to prevent 100Khz signal is all over the place, separate the analog from digital is very important.

But I have not to say this because people here and you already now.

regards
 

Attachments

  • class d amp distortion.jpg
    class d amp distortion.jpg
    295.7 KB · Views: 738
Hello,

Posted this in another thread, trying my luck also here (since it has to do with the topic):

Anyone any ideea why my Wondom Sure board AB23251V120 (former AA-AB32313/ AA-AB32512) stopped working.
When powered up, no matter the voltage the clipping red led starts blinking (D3) and also another green led (D14) right near one rca input.

It does so no matter the setup, with input signal or no, with load (speakers) conneceted or no.

I have to mention it happened random. I had it around, not connected, and, one day when I decided to play it again it just started blinking, just like that.


Thank you
 
Last edited:
Hi

Did you test mosfets?, if it is a self oscillating type maybe a defective chip, opamp etc.

is the supply oke.

Looks like it has a outputvoltage to the rails because of defective mosfet so the clip detector goes off, cause can be to small deadtime who did blow the fet (s)

regards
 
The chip itself (STA516) doesn't cost a fortune, but repacing it without the means it's next to impossible. There seems to be an opamp, though its markins are unreadable, but i think it's not involved in audio signal. There is also a Tripath driver for STA516 which, to my knowledge is long obsolete, so it's 100% a fake one - maybe this one failed? Replacing it means for sure buying another fake one.

To many variables and no smd soldering/desoldering means...

Anyaway, as a personal opinion about this board (AB23251V120 / AA-AB32313/ AA-AB32512) having it tested, I can say (from my opinion) the statements here abot it are overrated. I didn't know all the details before buying it - why would someone buy stuff they know it's obsolete and fake, even it's reported as "sounding good"?
 
The chip itself (STA516) doesn't cost a fortune, but repacing it without the means it's next to impossible. There seems to be an opamp, though its markins are unreadable, but i think it's not involved in audio signal. There is also a Tripath driver for STA516 which, to my knowledge is long obsolete, so it's 100% a fake one - maybe this one failed? Replacing it means for sure buying another fake one.

To many variables and no smd soldering/desoldering means...

Anyaway, as a personal opinion about this board (AB23251V120 / AA-AB32313/ AA-AB32512) having it tested, I can say (from my opinion) the statements here abot it are overrated. I didn't know all the details before buying it - why would someone buy stuff they know it's obsolete and fake, even it's reported as "sounding good"?

Ahh it is a chip and board chinese? when do stop these people selling fakes to te people.
 
Yes, it's a chinese board it has been reffered at in this topic. It's somehow buyers' fault too for not documenting themselves well enough before buying. At least, I learned my lesson. And, by the way, I bought the board from a german website (some sort of man in the middle because Sure Electronics was not shipping to my country at that point).