Well over five years ago, I built a class D amp utilizing one of the high power Tripath reference design boards configured as a differential H-bridge. I set up the power suppplies so that each amp channel could deliver 200W @ 8 ohms, 350W @ 4 ohms, and around 650W @ 2 ohms, per the reference design board's datasheet. Today, the sound is super transparent, bass has both natural weight and dynamic punch. Cymbals are reproduced with realistic brass tone and air. Bell and triangle strikes sound sharply percussive followed by natural decay. Vocals have natural presense, clarity, and tone. The soundstage will collapse to a nearly pinpoint mono, such as on some old Louis Armstrong recordings, and then, on the next album, expand to an enveloping 3D field extending far beyong the loudspeakers, both laterally and in depth. This wasn't always so.
Initially, the amp sounded about the same as many other mid-fi (circa, $2,500) power amps I've heard. However, being unable to leave well enough alone, I began to tinker with passive component selection. One particular modification of the original Tripath reference design produced a transformative affect on the sound. So much so, that I eventually sold my ARC Classic 120 tube mono-blocks without regret. So, what was that change? It was to substitute the normal single metalized mylar film capacitors in the LC output filter for double-metallized mylar units sourced from BC components. These capacitors are specifically designed for high slew and switching applications. I sometimes still shake my head in disbelief. Your experience may vary.
Initially, the amp sounded about the same as many other mid-fi (circa, $2,500) power amps I've heard. However, being unable to leave well enough alone, I began to tinker with passive component selection. One particular modification of the original Tripath reference design produced a transformative affect on the sound. So much so, that I eventually sold my ARC Classic 120 tube mono-blocks without regret. So, what was that change? It was to substitute the normal single metalized mylar film capacitors in the LC output filter for double-metallized mylar units sourced from BC components. These capacitors are specifically designed for high slew and switching applications. I sometimes still shake my head in disbelief. Your experience may vary.
Last edited:
yea and class D still sucks.. except for its efficiency..
just replaced # class D amps with a H and a/b no comparison the a/b/h killed it..
oh yes they were crown macro and K series amps..
haven't heard en Ncores yet.. but since the best is not available in anything but DIY i don't have time to fool with it..
and also being a car guy i've had about every class d car amp powering parts of my home system and they all sound like ****..
now from my guy who has done high speed (20ghz) data aquesiton board work and custom rf equipment here in Tech land for the nuclear facilities and is an audio guy too..
"when the learn what they are doing as far a board layout goes they might get close"
inadequate filtering , incorrect filter principles., etc,
they need a true RF guy to do board layout but the the cost goes way over A/B at that point!
Commercial amps are nothing more than to make money(sq might be second) and what sells now is weight and efficiency..
just replaced # class D amps with a H and a/b no comparison the a/b/h killed it..
oh yes they were crown macro and K series amps..
haven't heard en Ncores yet.. but since the best is not available in anything but DIY i don't have time to fool with it..
and also being a car guy i've had about every class d car amp powering parts of my home system and they all sound like ****..
now from my guy who has done high speed (20ghz) data aquesiton board work and custom rf equipment here in Tech land for the nuclear facilities and is an audio guy too..
"when the learn what they are doing as far a board layout goes they might get close"
inadequate filtering , incorrect filter principles., etc,
they need a true RF guy to do board layout but the the cost goes way over A/B at that point!
Commercial amps are nothing more than to make money(sq might be second) and what sells now is weight and efficiency..
From TI directly - pricey at $150 ea but a deal compared to what a class A will cost you for similar performance. I am building it with a custom made pcb and parts from Digikey, I am also building smaller TPA3118D2 using smd dead bug wiring in progress now on another thread. Total BOM about $17 not including $8 ps laptop brick.
I try to keep an open mind and I am not sure the arguments for Class D add up. On one hand not too many commercial Class D gets it right - its a different animal, requires different design skills, etc.,etc. On the other end I can on my own, purchase an eval board with a cheap switching PSU and compete with a Class A design? I suppose my Yamaha Aventage does not sound well because it is 2 years old and does not sport the latest chip or Yamaha does not design Class D very well, yet I can beat it by hooking up a TI eval board?
Try using a more resonant housing, strike it with a metal rod.
"Musical" amp "sound." Please!
Class D is not a whole lot different than the DAC situation. I personally must think that all DACs stink because I still own and use a turntable. DACs are still getting better all the time and the best DAC chips are yet to come. Digital sources must be popular because not many are running turntables these days and yes, I do mostly listen to digital. Absolutely no progress has been made in Class A and Class AB amplification technology in 30 years and in another 15 years, it may even no longer be available because everyone will be listening to Class D. The best chips are yet to come.
I thought we would all be riding around in personal spaceships by now.
Maybe an electric car some day will perform better than a Porsche. I guess it is fun to speculate into the future - but we live and play our music only in the present moment.
I replaced $5k worth of tube monoblocs with the ClassD Audio Spiral Groove Mastering amp. I have been an NPR station manager and classical recording engineer/producer for many years. I think I know what a system should sound like as I use recordings I made with such ensembles as the Chamber Music Society of Lincoln Center. No I do not think class d is perfect, but it is much better than many critics give it credit for. That is why I pay no attention to the "audiophile" rags and blogs. Regards
I am not sure what topology the Yamaha Class D amp uses and if these are better, but given that they are the latest offerings from TI, it wouldn't surprise me if they are indeed better.
The eval board will probably work well but to get the best performance, the main power supply electrolytic caps should be upgraded to a larger size and the input ceramic caps could probably be upgraded with good poly-foil caps. $150 will let you try out the test...
Regarding switch mode PS bricks... Even Nelson Pass says they are fine for supplying class A designs if all you need is 19 volts.
These TI chips are so new that there are very few people who can tell us how they sound. Although there are some diyAudio members who have built one and report very favorable performance. Here is a 50 w/ch TPA3116D2 amp and its bridged mono 100 w/ch sub amp sibling next to some valve amps:
http://www.diyaudio.com/forums/clubs-events/212618-greater-toronto-area-diy-meetup-23.html
I have also heard reports that these amps can sound as good as some of the best tube amps.
Last edited:
Class-D amplifier can be made to sound good. In the bipolar transistor ere, it was not possible to go very far with the carrier frequency. The first class-D amp I saw 30 years ago worked with 100kHz carrier.
Didn't sound very good above 200-500 Hz, So I used it for subwoofer.
Now with 2 Mhz carrier frequency, they sound very good to cca 4000 kHz, very well into the upper midrange.
With 10MHz carrier they'll sound good probably up to 20 kHz, but I still didn't hear any with 10 MHz. carrier...
There is also real physics and maths behind class-d BTW.
Those advocating very high switching frequencies should keep in mind that a given absolute timing error is resulting in a bigger relative timing error when the switching frequency is higher. So there is not much to be gained IMO with exorbitantly high switching frequencies.
The problem with low switching frequencies OTOH is getting enough negative feedback. With higher-order Loops this can still be achieved like the nCore example shows. But one is restricted to self-oscillating topologies. With carrier-based amps (i.e. natural sampling PWM) it is getting trickier because performance doesn't continue to improve above a certain feedback factor due to the so-called ripple aliasing effect which is causing a distortion mode that doesn't exist without NFB. There are ways around that but they are not always simple.
Regards
Charles
Those advocating very high switching frequencies should keep in mind that a given absolute timing error is resulting in a bigger relative timing error when the switching frequency is higher. So there is not much to be gained IMO with exorbitantly high switching frequencies.
The problem with low switching frequencies OTOH is getting enough negative feedback. With higher-order Loops this can still be achieved like the nCore example shows. But one is restricted to self-oscillating topologies. With carrier-based amps (i.e. natural sampling PWM) it is getting trickier because performance doesn't continue to improve above a certain feedback factor due to the so-called ripple aliasing effect which is causing a distortion mode that doesn't exist without NFB. There are ways around that but they are not always simple.
Regards
Charles
...Here is a 50 w/ch TPA3116D2 amp and its bridged mono 100 w/ch sub amp sibling next to some valve amps:
http://www.diyaudio.com/forums/clubs-events/212618-greater-toronto-area-diy-meetup-23.html
I have also heard reports that these amps can sound as good as some of the best tube amps.
Where is the source for those nice little modules?
I am going to experiment with the older brothers of this amp and gathered some documentation here: Dual-Mono TPA3122D2 AMP | H i F i D U I N O
From the older chips to the latest chips, they all have the same "THD signature" even though the latest chips can be run at 1 MHz or higher.
They have better specs if configured in BTL mode with balanced input
And they are incredibly simple to implement: good power caps and bypass, output LC filter and input coupling capacitors.
I agreed with you post absolutely. However the fact that Class D amp is heavily DSP based does not mean it has superior BW then Class A. It is quit opposite. You have explained why BTW.
Obviously there is no way that manufacture can provide a couple of billion consumers with gear that stuffed with audio trafos, chocks and foil capacitors. So Class D is the future for mass prodcution no wonder.
But few audiophiles who can afford a pair of good speakers, turntable or class A decoding computer could benefit from Class A amplification gear.
BTW several watts Claas A would consume less energy then 400 Watts Class D module while playing music. Hi Efficiency speakers is the way for Green route anyway. That is what experienced audiophiles are running for many years already.
There is also a $10 job @ PE:
2x8W @ 4 Ohm TPA3110 Class-D Audio Amplifier Board Only 320-329
Hook it up w a 12v battery and play! I wouldn't so readily dismiss it because it is sooo cheap!
Duc
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.