I do? Believe me, I went into this with great enthusiasm and, like MF, pursued it to the point of asking for additional samples in the hope that the initial impressions were based on defective or early production samples. If there was any bias, it would have been in Anthem's favor as I have been quite happy with many of their other products.
I do not know a lot about Class D. I read the 6Moons Review for Hypex nCore and was intrigued. I have a Class D Yamaha Aventage RX-A710 and there was big improvement plugging an F5 into the preamp outs, bypassing the Yamaha amp section.
What is considered the current reference Class D DIY design? A lot of the discussion here is around commercial products - not much DIY.
What is considered the current reference Class D DIY design? A lot of the discussion here is around commercial products - not much DIY.
What's wrong with class-D? Class D is very complicated and it has taken a very long time for good sounding amplifiers to evolve. The best are yet to come. I suppose the Hypex nCore would be considered a high powered reference amplifier. Most Class D amplifiers tend to be low powered. There has been lots of discussion lately of some of the newer Texas Instruments chips for being lower power reference quality. The TPA3110, TPA3100, and TPA3106 are popular chips that have been around awhile. The new TPA3116 and TPA3118 just may be state of the art and are just starting to get evaluated.
Thank you. Not to sound dismissive myself but the spirit of this thread is the dismissive attitude many in the industry has towards Class D yet there is not a lot out there and "the best are yet to come". I would say much of the industry has already moved to Class D due to attractive marketing specs and not necessarily best audio quality. I am interested in the technology but seems to be a be bit speculative in that it has a lot of potential not yet realized. My fairly new Yamaha class D is not even close to my Class A F5. I would say Class D is more marketing friendly in terms high power numbers in a compact footprint. Class D is much more prevalent in consumer grade, big box electronics gear. An average consumer would be more likely to select a 100 watt per channel receiver over a 25 watt receiver, not knowing the differences. I certainly took the bait when I bought the Yamaha before I really got back into audio.
Is there a proven design that you can point to using TPA3110, TPA3100, and TPA3106 that you can point me to that will compete or beat with Class A/B or even A? I am not playing devils advocate, I am interested in building a proven design to hear for myself. A PCB for sale would be great.
So far good sounding Class D amps are implement analog feedback so there is no way they can beat AB analog but Class D delivers excellent efficiency and compactness sacrificing big heat sinks (class D ICs perfect for cell phones, portable media players applications, larger Class D suits subs, consumer home cinemas etc.)
Regarding Class A amps delivers BW up to 250-500KHz while Class D switching frequency is in range of 250-500KHz.
The above statement means there is no way that Class D sounds close to Class A in terms of BW (transparency, fidelity).
Moreover good sounding HiEff speakers requires several watts to sing so I cannot see any practical benefit using AB or D or whatever amp over Class A because wasting few watts from receptacle is not an issue at all.
Regarding Class A amps delivers BW up to 250-500KHz while Class D switching frequency is in range of 250-500KHz.
The above statement means there is no way that Class D sounds close to Class A in terms of BW (transparency, fidelity).
Moreover good sounding HiEff speakers requires several watts to sing so I cannot see any practical benefit using AB or D or whatever amp over Class A because wasting few watts from receptacle is not an issue at all.
The TBI Millenia MG3 amplifier uses the Texas Instruments TPA3100 chip. Positive Feedback just rated this unit as best amplifier of the year. This is hardly something to be taken lightly and the newer TI chips are with little doubt even better.
The new TPA3116D2 can provide 25w/ch at less than 0.1% THD in a total package with all components the size of a credit card. Class A/B will have inherent zero crossing distortion which class D does not suffer from - it is essentially a powerful DAC using PWM technology, and to boot, it uses a single ended supply with zero-level output at Vcc/2. Let's look at the facts and real measurements - someone show me curves from a class A or AB that is superior to this chip. The 31xxD2 series has adjustable modulation freq from 400 kHz to 1.237 MHz selectable by jumper. You are not going to hear any modulation leaking through at these frequencies. Class A has the 'aura' of high end audiophile love, Class D is up and coming and has a battle to fight but I bet a double blind test may prove otherwise. For grins, someone should mount a tiny class D in a humongous aluminum chassis with fins and some resistive heaters added to generate the extra heat that the class D does not. People will connect their speakers, feel the heft, feel the heat, look at the great metalwork, and say this is a great class A amp.
Last edited:
The new TPA3116D2 can provide 25w/ch at less than 0.1% THD in a total package with all components the size of a credit card. Class A/B will have inherent zero crossing distortion which class D does not suffer from - it is essentially a powerful DAC using PWM technology, and to boot, it uses a single ended supply with zero-level output at Vcc/2. Let's look at the facts and real measurements - someone show me curves from a class A or AB that is superior to this chip. The 31xxD2 series has adjustable modulation freq from 400 kHz to 1.237 MHz selectable by jumper. You are not going to hear any modulation leaking through at these frequencies. Class A has the 'aura' of high end audiophile love, Class D is up and coming and has a battle to fight but I bet a double blind test may prove otherwise. For grins, someone should mount a tiny class D in a humongous aluminum chassis with fins and some resistive heaters added to generate the extra heat that the class D does not. People will connect their speakers, feel the heft, feel the heat, look at the great metalwork, and say this is a great class A amp.
And sell it for $18,000!
Cheers,
Rod
Have you built a TPA3116D2 amp? Where can one get a PCB? Is that what you use for your primary amp? If no, what do you use?The new TPA3116D2 can provide 25w/ch at less than 0.1% THD in a total package with all components the size of a credit card. Class A/B will have inherent zero crossing distortion which class D does not suffer from - it is essentially a powerful DAC using PWM technology, and to boot, it uses a single ended supply with zero-level output at Vcc/2. Let's look at the facts and real measurements - someone show me curves from a class A or AB that is superior to this chip. The 31xxD2 series has adjustable modulation freq from 400 kHz to 1.237 MHz selectable by jumper. You are not going to hear any modulation leaking through at these frequencies. Class A has the 'aura' of high end audiophile love, Class D is up and coming and has a battle to fight but I bet a double blind test may prove otherwise. For grins, someone should mount a tiny class D in a humongous aluminum chassis with fins and some resistive heaters added to generate the extra heat that the class D does not. People will connect their speakers, feel the heft, feel the heat, look at the great metalwork, and say this is a great class A amp.
Entirely anecdotal of course, but I finally got to hear a pair of Hypex nCore monos yesterday at the Hi-fi Wigwam show and driving Vivid Audio 1.5s they sounded very good by any reasonable standard. Unfamiliar room, unfamiliar system, but with several other good-sounding set-ups at the show to compare that system to, they certainly weren't a limiting factor, as far as I could tell. Certainly no obvious 'class D' sound. Less expensive class D amps have sounded clear but a bit monochromatic, tonally. YMMV naturally. The nCores, in the context of the system I heard them in, showed none of this. Actually, my initial impression was that they had exactly this weakness, but this turned out to be a quality of the first track I heard. The rest of the music showed significant variations in recorded quality and style, as you would hope and expect. I've no idea exactly how good the nCores are, but they seem very good irrespective of 'class' considerations.
Entirely OT, I also had a brief listen to the KEF LS50s in a system that also included a small Ming Da single-ended triode integrated amp (18W.) and the DSPeaker equaliser DAC. Also excellent sound and the (controversial?) impression that the SIT and the nCore probably sounded more similar than different, allowing for the different associated equipment. That said, both systems were playing computer-based music through high-quality DACs into relatively compact metal-diaphragm driver equipped stand-mount speakers, so the equipment was not that different...
Entirely OT, I also had a brief listen to the KEF LS50s in a system that also included a small Ming Da single-ended triode integrated amp (18W.) and the DSPeaker equaliser DAC. Also excellent sound and the (controversial?) impression that the SIT and the nCore probably sounded more similar than different, allowing for the different associated equipment. That said, both systems were playing computer-based music through high-quality DACs into relatively compact metal-diaphragm driver equipped stand-mount speakers, so the equipment was not that different...
Last edited:
You can purchase Texas Instruments evaluation modules for the TPA3116 and TPA3118 chips. You will only need to hook up a 24 volt/8 amp converter or power supply.
I changed to Hypex UcD 180HG modules to drive Jordan Eikonas from Rod Elloitt's P101s.
The Hypex have better resolution (less distortion) than the P101s - listen to 10 minutes of orchestral tutti from pp to ff on both amps and you hear an easily discenible, but not a huge, improvement; that's why I bought the Hypex after listening to a friend's.
The bass did seem to be lighter after changing from P101 to Hypex. What is difficult to recognise is that there isn't very much really deep bass in most music. Very little goes below 50Hz, and very few speakers reproduce this at the right level. Much of the sound of a bass note is harmonics - in the same region as the fundamental of other instruments. A lack of resolution (greater distortion) will make bass sound heavier by adding to the harmonics - the bit we get from our system, and then interpret as bass because we know the fundamental should be there.
I've added a pair of Monacor SPH275Cs as a sub, driven by the P101s. Now it is very obvious how infrequently the frequency goes very low.
A good example is an organist seeing an fff dynamic (otherwise known as "shake the building".) S/he knows that a 32' Sub-bourdon pedal stop (down to 16Hz, few overtones) won't do much against the full Great, so s/he pulls out all the pedal stops - 16', 8' and 4' stops, then couples to whatever manual s/he isn't playing, giving everything up to 2' stops. This is any number of harmonics; but we know that it is the pedal line (usually moving at half or quarter the speed of the rest of the music) and so we recognise it as bass.
There is a track that demonstrates this rather well (Naxos 8.550582 track 2) in which the organist plays the pedals with couplers (probably to another manual) to achieve fff, then switches out the couplers as the music diminuendos - The 32' fundamental suddenly appears.
Also try http://www.organstops.org/_sounds/CulverAcademy/index.html under headphones for some really low, real world, musical notes then play on your speakers.
At the top end, fiddling about with the miniDSP can make (almost) any amount of top end you wish - bear in mind that I now only hear up to 10kHz relative to a reasonably loud 200Hz. I hope that everyone has got hold of a tone generator (e.g. Tolvan Data) and listened under headphones via a good soundcard to check their hearing before spending money on stuff they can't ever fully appreciate.
Regards,
Andy
The Hypex have better resolution (less distortion) than the P101s - listen to 10 minutes of orchestral tutti from pp to ff on both amps and you hear an easily discenible, but not a huge, improvement; that's why I bought the Hypex after listening to a friend's.
The bass did seem to be lighter after changing from P101 to Hypex. What is difficult to recognise is that there isn't very much really deep bass in most music. Very little goes below 50Hz, and very few speakers reproduce this at the right level. Much of the sound of a bass note is harmonics - in the same region as the fundamental of other instruments. A lack of resolution (greater distortion) will make bass sound heavier by adding to the harmonics - the bit we get from our system, and then interpret as bass because we know the fundamental should be there.
I've added a pair of Monacor SPH275Cs as a sub, driven by the P101s. Now it is very obvious how infrequently the frequency goes very low.
A good example is an organist seeing an fff dynamic (otherwise known as "shake the building".) S/he knows that a 32' Sub-bourdon pedal stop (down to 16Hz, few overtones) won't do much against the full Great, so s/he pulls out all the pedal stops - 16', 8' and 4' stops, then couples to whatever manual s/he isn't playing, giving everything up to 2' stops. This is any number of harmonics; but we know that it is the pedal line (usually moving at half or quarter the speed of the rest of the music) and so we recognise it as bass.
There is a track that demonstrates this rather well (Naxos 8.550582 track 2) in which the organist plays the pedals with couplers (probably to another manual) to achieve fff, then switches out the couplers as the music diminuendos - The 32' fundamental suddenly appears.
Also try http://www.organstops.org/_sounds/CulverAcademy/index.html under headphones for some really low, real world, musical notes then play on your speakers.
At the top end, fiddling about with the miniDSP can make (almost) any amount of top end you wish - bear in mind that I now only hear up to 10kHz relative to a reasonably loud 200Hz. I hope that everyone has got hold of a tone generator (e.g. Tolvan Data) and listened under headphones via a good soundcard to check their hearing before spending money on stuff they can't ever fully appreciate.
Regards,
Andy
It looks like Mouser Electronics has the TPA3116D2EVM in stock @ $159.08. If you search, you might find one a little cheaper but the price looks reasonable.
Hi,
I want to say a few words purely as an amateur, I think I have a good ear after so many years of listening in the range of high class AB (without interest with the company for which I have developed or other competitors who have, with sound measures, the class D perfect ).
In class AB, traditional audio measurements, have a perfect relationship with the result of listening. this is due to the total absence of phenomena that, in class D are perfectly audible (example; poor definition in the high range, not comparable with the class AB. or alteration of the harmonic content in relation to the percentage of modulation. or intermodulation when in the presence of instrumentals and vocals, raise the power, just with reactive load.) these defects are still present on many class D and are not visible to the audio measurements.
these phenomena are the result of the intrinsic concept of class D.
I listened my MC phono set to mono output: 1 channel on AB true high end, the other channel on class D on the market. connected to the two speakers B & W 802 II.
Rotating the balance quickly and repeating some passages of music (to use the short memory of the brain), the difference (to my ears) is huge and is not comparable, especially at 70% power.
then, as always there is the real world, and... marketing!
it is better (in my experience) start from a good design of class D, if possible do not use dedicated chips, and then put up good final result continually working on comparison with the class AB.
Regards
I want to say a few words purely as an amateur, I think I have a good ear after so many years of listening in the range of high class AB (without interest with the company for which I have developed or other competitors who have, with sound measures, the class D perfect ).
In class AB, traditional audio measurements, have a perfect relationship with the result of listening. this is due to the total absence of phenomena that, in class D are perfectly audible (example; poor definition in the high range, not comparable with the class AB. or alteration of the harmonic content in relation to the percentage of modulation. or intermodulation when in the presence of instrumentals and vocals, raise the power, just with reactive load.) these defects are still present on many class D and are not visible to the audio measurements.
these phenomena are the result of the intrinsic concept of class D.
I listened my MC phono set to mono output: 1 channel on AB true high end, the other channel on class D on the market. connected to the two speakers B & W 802 II.
Rotating the balance quickly and repeating some passages of music (to use the short memory of the brain), the difference (to my ears) is huge and is not comparable, especially at 70% power.
then, as always there is the real world, and... marketing!
it is better (in my experience) start from a good design of class D, if possible do not use dedicated chips, and then put up good final result continually working on comparison with the class AB.
Regards
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.