I recently compared an old AB power amp (so old it had a single 75V rail supply and all NPN Outputs. I compared it with a tpa3118 (60w) and although not a fair comparison in power rating (the AB amp is about 120w), the old AB sounded a lot more polished and alive. I didn't mind the class D until I "AB-d" them with the same pres (one tube pre and one discrete SS). Maybe class D just can't sound as good? Maybe it was the extra watts in the AB amp?
A good class D amplifier can sound as good as a class AB amplifier. This statement is based on comparison of THD levels. Much more important for the comparison is your personal preference. It was hardly the higher power level of the class AB amplifier that made the difference.
The TPA3118 amplifier may not be the best reference choice for Class D compared to Class AB.
Review and Measurements of Purifi 1ET400A Amplifier: Review and Measurements of Purifi 1ET400A Amplifier | Audio Science Review (ASR) Forum
The Benchmark ABH2 is a top of the line class AB amplifier, the Purifi 1ET400A Class D amplifier does pretty well.
Review and Measurements of Purifi 1ET400A Amplifier: Review and Measurements of Purifi 1ET400A Amplifier | Audio Science Review (ASR) Forum
The Benchmark ABH2 is a top of the line class AB amplifier, the Purifi 1ET400A Class D amplifier does pretty well.
Attachments
In my experience, a class-D chipamp (TDA8932) sounds better than any classAB chipamp I've heard. I wasn't too impressed with the TPA3116 board I listened to.
Hello,
Thinking all class D amps sound the same is the same mistake as thinking all class X (IE: AB) amps sound the same. Just not reality.
There are good and bad designs in every class of amps.
Thinking all class D amps sound the same is the same mistake as thinking all class X (IE: AB) amps sound the same. Just not reality.
There are good and bad designs in every class of amps.
Absolutely right! The Class D situation is a little bit different however in that many are chip amps built closely to the data sheet example. Class AB tended to be more to taste and possibly even incorrectly designed (for the good or for the bad) in some cases and with regards to specs.
One thing I know makes a difference (even using the data sheet example circuit) for the TPA3118 series is capacitor, inductor, values and quality and resistor values for gain etc. The data sheet does leave some things as variables. So as they say... YMMV with a chip amp design. Maybe I need to experiment more.
While theoretically plausible, not quite sure if it's borne out in practice. The various TPA3116 boards I have used do have different sound quality. Some have noise issues, muffled sound, rolled off highs/lows etc. Some are noiseless and lively.
So it's still possible you prefer class AB amps to class D amps, but it's also possible you didn't get a good class D amp.
I've had a number of the china 3118 boards and prefer some to others. I hear a difference based on caps and output inductor size. But the best still doesn't sound quite as good as an old AB amp that I have (same pres used). I haven't experimented with the input gain on the boards. I need to do that next. I've compared two discrete AB power amps also (made by the same designers) and they also sound different. Maybe it's certain OP transistors that I think sound better. Of course it's all perception but my son also agreed. The china class D sounded a little lifeless compared.
The PCB layout is extremely critical to class D amp performance. The feedback circuit implementation is what makes or breaks it. If you dont keep switching noise and HF out of the feedback loop, you end up with too much IMD and odd order HD. The ground plane has alot to do with this. The other problem is phase shift and early response rolloff at the upper end of the audio band, mainly from output stage HF filtering. One of the most effective ways to deal with HF noise is using a transformer coupled input stage, but that isn't cheap. I heard an ICE module based amp that actually sounded very good and to date it was te only class D design I really liked.
In general,, you can't beat a high bias DC coupled AB amp, but when comparing chip amps, there are just as many bad AB designs as class D chip amp designs. I've built a gain clone with LM3886s and ended up loving it. I didnt have any luck with any of the class D chip amps so far.
In general,, you can't beat a high bias DC coupled AB amp, but when comparing chip amps, there are just as many bad AB designs as class D chip amp designs. I've built a gain clone with LM3886s and ended up loving it. I didnt have any luck with any of the class D chip amps so far.
Class D and FDA are relatively new developments and it is going quite fast. When the Tripath hype started things were a bit different. I think it is safe to say that the most recent chips/designs are quite well designed and can sound equal or even better than older analog amplifiers. As conservative the DIY world can be (tubes, LM317, R2R DACs) as sturdy it can remember first hearing an early development Class D device and condemning the whole technology. There is no way class D and FDA will go away, the cheapest technology always wins. FDA will likely be the next standard as all modern sources are digital.
Also it is unfair to compare cheap china PCBs with ready made amplifiers that were way more expensive as these likely have an overdimensioned PSU as was common in the time these were built whereas the Class D board often is fed with an SMPS/adapter and with very low cost and high margin as design goals.
Also it is unfair to compare cheap china PCBs with ready made amplifiers that were way more expensive as these likely have an overdimensioned PSU as was common in the time these were built whereas the Class D board often is fed with an SMPS/adapter and with very low cost and high margin as design goals.
Last edited:
A friend who design high power class D amplifier, compared his amplifier with Hypex class D amplifier. It is not blind test, just casual listening. All sound good, some part his amplifier better and other part Hypex is better. They sound better than many kit class AB amplifier in my country.
Then we compared with several amplifier that I designed, of course, it is class AB. He said all my amplifiers are better. My amplifiers have high slew rate, and relative low distortion at 20kHz.
After this, he modified his amplifier and said it improved. But we do not meet again because of pandemic.
Then we compared with several amplifier that I designed, of course, it is class AB. He said all my amplifiers are better. My amplifiers have high slew rate, and relative low distortion at 20kHz.
After this, he modified his amplifier and said it improved. But we do not meet again because of pandemic.
These type of friendly competitions are fun and worth while in my opinion. I do it all the time but just with my own equipment vs new things I bring in. I like to find out why I like what I like and to try to improve upon it (or at least find if it's possible).
If you really want the best possible sounding Class D amplifier, then maybe consider a self oscillating architecture with properly implemented global feedback. The Class D ICs TPA3118/3116/3255/3255 are driven PWM with local feedback. The optional minimal post filter feedback (such as in the TI app note "TPA324x and TPA325x Post-Filter Feedback") is inadequate to fix the inductor distortion completely.
The Class D ICs TPA3118/3116/3255/3255 are fine for general purpose audio applications, but if one is very critical (obsessive) about sound quality of the amplifier, then those IC amplifiers may not be the best choice.
...........Here are a couple of Class D examples showing good Class D amplifier perfomance....
Fresh From the Bench: A Tale of Two Class-D Amplifiers Orchard Audio BOSC and Purifi Audio Eigentakt EVAL1 | audioXpress
Hypex also has some nice Class D examples: DIYclassd.com
..........Here is a paper on making a low distortion PWM modulator, but I think the self oscillating UcD variant is more elegant approach......
Simple, Ultralow Distortion Digital Pulse Width Modulator, May 2006, Bruno Putzeys: (PDF) Simple, Ultralow Distortion Digital Pulse Width Modulator
..........Alternative UcD approach...
Simple Self-Oscillating Class D Amplifier with Full Output Filter Control, May 2005, Bruno Putzeys: https://www.researchgate.net/public...s_D_Amplifier_with_Full_Output_Filter_Control