I disagree, most people get headaches from listening to music through a class D amp!
Seriously, I still disagree to the categoric statement above and BTW it´s noting wrong to use a tube amp or vinyl records if you like it this way.
If you are getting a headache from listening to a class D amp, try a different amp or different headphones.
MrkTrout.
You are totally missing my ironic twist, please do not qoute in an inappropriate way....this should be clear if one goes back to the first few pages!
Last edited:
Well... very interesting discussion indeed.
I am in the process of building a small scale PA system for community cultural programs (small auditorium or large party room). I tried a 300W+300W TAS5630 based Class-D. I bought this from sure-hifi on ebay and powered with 50v dc from a toroid and I must say that the quality of sound is OK (read good) but not as good as an old Pioneer VSX815 which is a medium consumer grade home-theater receiver. A few years ago I had a Panasonic SA-XR55 (again a common consumer grade and inexpensive Class-D home-theater receiver) and it was very tiring to listen to music on that. We tend to say, "Can you please keep that volume down?". The speakers in both these cases were 8 Ohm Natalie-P (Parts Express/Dayton). You don't need golden ears to distinguish the difference as my hearing stops after about 15 kHz
So, for my own purpose I am going to use the Class-D for low frequency (below 1200Hz) on an active cross-over along with LM3886 chip-amp for high frequency.
In my experience Class-D is adequate for most purposes but not quite as good as my old 2N3055 based DIY amplifier (Class AB) built as published in an old Elektor. It is hard to convince that to an audience who says MP3 songs played on a tablet or ipod sound good enough!
I am in the process of building a small scale PA system for community cultural programs (small auditorium or large party room). I tried a 300W+300W TAS5630 based Class-D. I bought this from sure-hifi on ebay and powered with 50v dc from a toroid and I must say that the quality of sound is OK (read good) but not as good as an old Pioneer VSX815 which is a medium consumer grade home-theater receiver. A few years ago I had a Panasonic SA-XR55 (again a common consumer grade and inexpensive Class-D home-theater receiver) and it was very tiring to listen to music on that. We tend to say, "Can you please keep that volume down?". The speakers in both these cases were 8 Ohm Natalie-P (Parts Express/Dayton). You don't need golden ears to distinguish the difference as my hearing stops after about 15 kHz
So, for my own purpose I am going to use the Class-D for low frequency (below 1200Hz) on an active cross-over along with LM3886 chip-amp for high frequency.
In my experience Class-D is adequate for most purposes but not quite as good as my old 2N3055 based DIY amplifier (Class AB) built as published in an old Elektor. It is hard to convince that to an audience who says MP3 songs played on a tablet or ipod sound good enough!
Last edited:
Why would I care? I couldn't hear 20kHz. even if it was present on any of my source material, which it isn't, and I certainly couldn't hear any harmonics.
What Class D offers is what Class A offers . . . freedom from crossover distortion (which becomes increasingly noticeable and obnoxious at the lower power levels where amplifiers are operating most of the time). Some Class AB amps avoid it too . . . a notable example being the LM3886 . . . but too many don't, especially as bias "drifts" with time and temperature.
AS a Lifelong Audiophile I find I am unable to listen to any Class D amplifier they make me feel ill and I have to turn them off even if hearing them in another room. Well designed Class AB amplifiers are excellent with no listening fatigue. Anyone not into High Fidelity could probably use a Class D amp but not me.
I had a Skoda with a 12-channel class-D amp rated at 360W. However, when you listen to the music you feel the output to be around 50W overall!
I have my serious doubts on power delivering capabilities on the Class-D amps. Similarly goes for the Sony home theater systems (the compact ones) vs AVR!
I have my serious doubts on power delivering capabilities on the Class-D amps. Similarly goes for the Sony home theater systems (the compact ones) vs AVR!
Any ... and especially a large class AB , has a considerable class A
"first watt" ( some- first 5 W ) !
NO X-over distortion.
After this , some faster AB designs with advanced thermal biasing and
really advanced compensation can even cancel out the class B switching
distortion down to PPM levels.
A 200 watt class AB amp @ <50ppm with a 250khz bandwidth is common
on the forum now.
The LM3886 sucks (tested one) - no headroom , push it and it shuts down.
It does not "avoid" any of the issues in the larger amps - in fact , compared to them - It sucks !!!
Bias "drifts" .... WHAT ??? .... just opened my amp after 3 months , still 60ma
bias , 5C to 35C ... still 60ma.
LM3886's and (most) class D is to benefit cheap chinese walmart junk , NOT
for SQ.
PS- a phillips 5 channel HT system with 6- 3886's was most listenable as compared to a sony walmart class D chinese "junkbox".
OS
My experience with different amp classes is that they can all sound good if they are put together right. Many manufactures have looked to class D to cut cost. When cost is their biggest consern they cut down on component quality and psu capability aswell. In my diy class D experience it delivers a lot more grunt than AB amps with the same psu capability. Also class D is relatively new to most people which reflect extremely poor layout in many class D pcbs for diy. It looks that many designers don't fully grasp the high frequency effect in class D. Crusial components are placed to far away from were they should be,caps high frequency esr is more important, bootstraps filters are placed too far away, oscillating problems cause of skipped snubbers, mismatch between speakers and class D output filter etc. There is a lot of pitfalls in class D.
A good example of a optimum class D card is the wiener 3118 made by Gmarsh. If one reads Ti's design guide you could see that all the design guides points have been achived, no wonder it sounds really good. Looking at many of the el'cheapo cards out there makes me think that they haven't even bother to look into design guide for the Ti's chips. I think this is true for most of the cheap class D cards with their overdone heatsinks and crappy layout.
Perhaps Class A gives a tiny bit better sound quality but my experience with class D is that it delivers good sound and power cheaply. The new Pioneer surround receivers is class D and has way better bass than the previous models.
In car use I now only use class D. In my living room is mostly use the tpa3116/8 and sometimes my diy AB tube amp which is quite costly to use since the tubes now cost a fortune. The tube amp has its own sound which I like but I can't say thats its a true reproduction of the source material.I think all classes have their usefulness but I think class D is going to be the "go to" class for most applications in the future.
A good example of a optimum class D card is the wiener 3118 made by Gmarsh. If one reads Ti's design guide you could see that all the design guides points have been achived, no wonder it sounds really good. Looking at many of the el'cheapo cards out there makes me think that they haven't even bother to look into design guide for the Ti's chips. I think this is true for most of the cheap class D cards with their overdone heatsinks and crappy layout.
Perhaps Class A gives a tiny bit better sound quality but my experience with class D is that it delivers good sound and power cheaply. The new Pioneer surround receivers is class D and has way better bass than the previous models.
In car use I now only use class D. In my living room is mostly use the tpa3116/8 and sometimes my diy AB tube amp which is quite costly to use since the tubes now cost a fortune. The tube amp has its own sound which I like but I can't say thats its a true reproduction of the source material.I think all classes have their usefulness but I think class D is going to be the "go to" class for most applications in the future.
Class-D for your low frequency driver, Class-AB for your HF driver.
use a DSP to split the Xover point, and enjoy the ultimate sound quality.
No evidence that class-d performs better in the HF area, unless some one can show same output signal from a Class-D compared to a Class-AB amplifier, for example a 15K square wave, or 15K sine wave, once both waves LOOKS exactly the same, no distortion, no noise, same slew rate, then OK
Else, god knows!
use a DSP to split the Xover point, and enjoy the ultimate sound quality.
No evidence that class-d performs better in the HF area, unless some one can show same output signal from a Class-D compared to a Class-AB amplifier, for example a 15K square wave, or 15K sine wave, once both waves LOOKS exactly the same, no distortion, no noise, same slew rate, then OK
Else, god knows!
Im about to tackle this one.. very rare.. very vintage.. then the test. Probably the highest quality Class D I will ever listen to.
www.davidsaudio.com/Infinity_DSP.pdf
If I can get it working again.
www.davidsaudio.com/Infinity_DSP.pdf
If I can get it working again.
Attachments
Last edited:
Just out of curiosity, what makes you think this particular class D design will be the best?
Regards,
Owen
The key to getting the most out of the National chips (currently I'm building with LM4766 which is a dual channel 30W-ish) is to optimize the load impedance through using an output transformer. They love higher impedance loads - the LM4766 for example has a 3A min current limit so it can't be guaranteed to swing more than 24V into 8ohms. But its supply limit is 37V - all that capability is unavailable with normal speaker loads. It gets even worse if you like (as I do) to operate your amps bridged.
Use a 2.5:1 step down trafo (so a 4ohm load looks like 25ohms) and the capabilities of the chip can be fully realized in bridged mode. Sounds much better too due to lower supply noise and better OPS PSRR.
Class D is digital in the sense that it creates audio frequency waveshapes with samples, happening at a rate that needs to be at least twice the highest frequency it will ever be asked to handle, OR there will be what's called "aliasing distortion", which causes a folding over of the input frequency in a totally non-musical way. Even if the sampling rate and anti-aliasing filter is adequate (and it is often borderline), the final output is likely to retain some of the Rf energy from the switching on and off of the individual samples. Those high current passive output filters are usually just 2 pole. So there may well be significant supersonic energy delivered to the speaker.
Here's where it gets interesting. I was taught in college that energy just above the audio frequency range, centered at about 27kHZ will cause psychological stress in humans. You can't hear it, but it makes you feel uncomfortable over time. I haven't verified this, but if it's true, this leakage of low level Rf combined with the aliasing distortion products due to the not perfect anti-aliasing or recontruction filter and sample rate, could generate a bit of this psychological stress energy, while the music itself seems to sound perfectly good.
So if I used a class D amp for just the woofer, I'd put a passive low pass filter between it and the woofer (probably don't need to, but some woofs put out significant energy in the upper midrange, so... For mids and tweeters I love using the LM3886 chips, because their bias drift over temperature is tracked very well, since the power devices and the bias sensing circuit is on the same chip, thermally tightly coupled, so crossover distortion is kept very minimal. With a bipolar 32VDC power supply I'm measuring about 50watts rms into a 10 ohm load resistor. Linkwitz drove his pluto woofer with two 3886's wired in bridge mode, which gives over 100 watts to the speaker, if you need that much. A very picky tube guru friend of mine tried to get a 3886 to blow up with various loads, short circuits, higher supply voltages, etc. and came away saying it is a very difficult chip to blow up, when that's what you're trying to do. So hey, there it is. I built a 4 channel 3886 poweramp, and I like it so much I'm in the process of building another... My speakers are all triamp'd so I need more channels, and these chips make it much easier than some discrete circuit.
Here's where it gets interesting. I was taught in college that energy just above the audio frequency range, centered at about 27kHZ will cause psychological stress in humans. You can't hear it, but it makes you feel uncomfortable over time. I haven't verified this, but if it's true, this leakage of low level Rf combined with the aliasing distortion products due to the not perfect anti-aliasing or recontruction filter and sample rate, could generate a bit of this psychological stress energy, while the music itself seems to sound perfectly good.
So if I used a class D amp for just the woofer, I'd put a passive low pass filter between it and the woofer (probably don't need to, but some woofs put out significant energy in the upper midrange, so... For mids and tweeters I love using the LM3886 chips, because their bias drift over temperature is tracked very well, since the power devices and the bias sensing circuit is on the same chip, thermally tightly coupled, so crossover distortion is kept very minimal. With a bipolar 32VDC power supply I'm measuring about 50watts rms into a 10 ohm load resistor. Linkwitz drove his pluto woofer with two 3886's wired in bridge mode, which gives over 100 watts to the speaker, if you need that much. A very picky tube guru friend of mine tried to get a 3886 to blow up with various loads, short circuits, higher supply voltages, etc. and came away saying it is a very difficult chip to blow up, when that's what you're trying to do. So hey, there it is. I built a 4 channel 3886 poweramp, and I like it so much I'm in the process of building another... My speakers are all triamp'd so I need more channels, and these chips make it much easier than some discrete circuit.
I used to think that Class D amps were only good for subwoofers. Then, I came across the Primare A34.2 amp, and lo and behold, this is a class D amp that sounds excellent. They use a propriety configuration ( Ultra Fast Power Device) that exhibits none of the normal limitations of class D. I'm running a pair of them in mono, which provides 1100 watts @ 4 ohms into a pair of Dynaudio 360 speakers. They sound every bit as good or better than most any class A/B amp (and I've owned and re-furbished a LOT of them).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.