Primarily on subwoofers, but all in all preference would be interesting to know what everyone thinks.
How would well made SS class A/B fair, apples to apples to one of todays class D on subs do you have a clear preference? Put aside any other advantages or disadvantages, strictly about sound.
How would well made SS class A/B fair, apples to apples to one of todays class D on subs do you have a clear preference? Put aside any other advantages or disadvantages, strictly about sound.
Does that include all the wonderful stand mounts an speakers that dont go low but outshine everywhere else. Id hardly call audio note an-j sub par.
Mains wont fill in problem room nodes in every rooms spots either. No matter how full speakers are subs properly capable of integrating will always sound better. It wasnt really meant to start an off topic debate, one thats over done already I might add.
Mains wont fill in problem room nodes in every rooms spots either. No matter how full speakers are subs properly capable of integrating will always sound better. It wasnt really meant to start an off topic debate, one thats over done already I might add.
Subs are the places you least need a high definition capable Class AB or A. Bass and midbass in a multiway are different though and clearly sound better with a good discrete output stage Class AB. Mids and highs are best sounding with Class A. Subs cannuse Class D and it sounds just fine - better since easy to get 1kW output without turning your house into a sauna.
Arent A?B supposed the better of both worlds? Some of the new pro amps by crown and baringer look tacky but who cares if the SQ and SPL's are there right?
I'm not decided what to use, I have a heavy AB power block thats 150x2 in 4 ohms. will subs shorten life?
Inuke or hypex. I'd have to hide the baringer.
https://www.madisoundspeakerstore.com/AMPS/SpeakerPower-SP1-300-Subwoofer-Amplifier
sale https://www.madisoundspeakerstore.com/AMPS/SpeakerPower-SP1-300-Subwoofer-Amplifier
Hi, some hard facts: Pure Class-D (pure as without Class A/B add-on making a Class-S, say amplifier with feed-forward error correction) has greater problems to output infinitesimally small signals than Class A/B. See "switching range" here. Most Class-D has feed-back error correction but only limited amounts, and this amount is linearily falling with frequency, hence the amplifier can output twice as small signals at half frequency. Cut short, highly efficient loudspeakers outputting 4 KHz listened to at close distance are worst case for Class-D.
Audio folklore may say so, I'm merely reporting my own experience. Audio folklore also says the speakers are the most important and weakest link in the chain, again my experience differs.
Member
Joined 2009
Paid Member
This sounds like the accepted wisdom but I don't buy it for a minute
There are enough reports of excellent results from Class AB on mids and highs that suggests Class A is not universally the best sounding. And there are advances in Class D that have produced results that challenge Class AB on all levels.
Are you saying that a Class-D amplifier has higher distortion with a small signal than with a large one?
That is not my understanding/experience at all. Sure the signal to switching noise is higher for small signals but with good design the switching is purely ultrasonic noise and can be easily filtered out.
I think a lot of people misunderstand Class-D and think that they are 'digital' in their operation like a DAC or ADC and therefore exhibit quantization errors which would make them have a poor noise floor if operated at very low volumes.
This is not the case. Class-D are still inherently analogue in their operation because the times at which the output switches do not have to adhere to a strict master clock like in a Delta-Sigma DAC/ADC - it is free to transition at any time it wishes. It's effectively a 1bit Delta Sigma modulator but the clock frequency is infinite because it can 'clock' whenever it wants therefore it has infinite effective number of bits resolution.
Most Class-D amplifiers actually exhibit a non-linearity that is not unlike the crossover distortion in a Class-AB. This is due to the transition between soft switching and hard switching of the output stage, which in turn affects the timing of the edges of the switching signal. Therefore they are similar in distortion behaviour to Class-B/AB; for really small signals they are very linear but once you reach the 'crossover' region distortion rises. Then for really large signals the crossover region becomes small and insignificant and the distortion falls again as power continues to increase. A Class-D then faces another unique problem and that is when the signal gets near to the rails the duty cycle will be very close to 0% or 100% - this presents a problem because the output stage does not have infinite bandwidth and to produce a 99% duty cycle at a certain frequency is a lot more demanding than 50% duty cycle at the same frequency. You can imagine that with say a 100KHz switching frequency to produce a 50% duty cycle output one transistor is on for 5 microseconds and the other is on for 5 microseconds. To produce a 99% duty cycle signal one transistor is on for 9.9microseconds and the other is on for 0.1microseconds - very demanding.
Therefore most Class-D amps will drop the switching frequency as the signal approaches the supply rails, which due to a variety of reasons usually in turn increases non-linearity/distortion especially for high frequency signals (several KHz)
Last edited:
Yes. In most Class-D data-sheets you see, that THD+N rises with the inverse of level. Some say, this were due to Noise-floor, which dominates THD+N. I say, maybe, and noise is artificially introduced as a dither in order to mask distortion at lo level.
This distortion is like quantization, belo the lowest bit the system ignores everything.
Class-D first transforms level into time, smallest level becomes shortest time, and then time back into level. But minimum time is limited, as a real device cannot switch off/on/off/on for infinitesimally short times. Hence small signals get lost.
This is where your reasoning falls over in a heap. Its true that minimum on-time is limited however minimum on time occurs at maximum output, not at zero crossing. So in a classD minimum on-time limits maximum modulation (typically in the high 90%s - take the TDA8932 which has about 80nS for its shortest possible pulse, running at 320kHz it achieves 98%)
Last edited:
Dear friends,
I think it depends on class D project. Mainly in output filter quality.
I could listen to a small branded class d amplifier with a 805D B&W speaker and it was very impressive. The cost also was impressive
So assmbly or buy application notes amplifier is not the good way to go ...
I think it depends on class D project. Mainly in output filter quality.
I could listen to a small branded class d amplifier with a 805D B&W speaker and it was very impressive. The cost also was impressive
So assmbly or buy application notes amplifier is not the good way to go ...
This sounds like the accepted wisdom but I don't buy it for a minute
There are enough reports of excellent results from Class AB on mids and highs that suggests Class A is not universally the best sounding. And there are advances in Class D that have produced results that challenge Class AB on all levels.
I don't diasagree with you. If I could only have one amp, it would be Class AB. I have yet to hear a Class D that is noise free (no hiss audible) and as transparent as Class A for headphones or high sensitivity compression drivers.
But you see exactly the same in all class A/B - A I have ever seen too. And for exactly the same reason.
So that is no different.
Jan
Most Class-A/B amps produce less noise than Class-D, and (belo around a fifth of a Watt or so for most types) Class-A/B amps produce indeed the less distortion, the lesser level gets. They have no least bit to overcome, while Class-D has.
No. Voltage before output lopass has min on-time at max output, but that voltage does not matter for result, as it gets filtered out. Current thru power actors determines output level and has min on-time at min signal level.abraxalito said:
Yes you said that, you mentioned the rising level with lower signal, but if you look at the THD+N graphs, whether class D, A/B, A, they ALL show rising with lower level. It's is a result from the way it is measured.
That is what I wanted to point out. So your argument against class D that they are 'bad' because of rising THD+N with lower level is an invalid argument.
And if you ever looked at the graphs for the nCore's (did you?) you will see that it puts many class A/B and even A to shame. Also at low level. But you really have to look into it because relying on 'everybody knows' of 20 years ago doesn't work anymore for class D in 2017.
Jan
Member
Joined 2009
Paid Member
I was reading class D amps a lot often lading on pages with 5 and 10 year old opinions, so I thought it would be interesting to read a collaboration of fresh ideas, while also hearing about trickle advancements through time and how class D is perceived, and so on and so forth.
In terms of sound, nothing else, with that thought in mind its interesting to know there are still are a couple praised well known subwoofer manufacturers still putting A/B amplifiers in brand new subs. With the added cost and all. For whatever thats worth. You would think if a company could save while not giving up anything in terms of SQ they would?
I'll possibly try these on, https://www.madisoundspeakerstore.c...cl45-8-8-polypropylene-cone-sub-woofer-8-ohm/
I dont think my A/B will struggle with those in the mean time I'll not beat up on the amp too hard. Then if I were to get a 300w hypex plated DSP. Would also be interesting to try new quality made D tech back to back with A/B on the very same subs and set up.
In terms of sound, nothing else, with that thought in mind its interesting to know there are still are a couple praised well known subwoofer manufacturers still putting A/B amplifiers in brand new subs. With the added cost and all. For whatever thats worth. You would think if a company could save while not giving up anything in terms of SQ they would?
I'll possibly try these on, https://www.madisoundspeakerstore.c...cl45-8-8-polypropylene-cone-sub-woofer-8-ohm/
I dont think my A/B will struggle with those in the mean time I'll not beat up on the amp too hard. Then if I were to get a 300w hypex plated DSP. Would also be interesting to try new quality made D tech back to back with A/B on the very same subs and set up.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.