|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.115.217.131
I just read somebody claiming Class D is not digital.
Although there may be the odd Class D amp that is all analog, it seems to me that traditional or common class D amps are indeed commonly known as digital with at least one conversion of the signal from analog to do digital and back again
Thx,
Follow Ups:
Class D is PWM (Pulse Width Modulated), but I bet there are various schemes for controlling the Width parameter, including feedback and possibly DIGITAL means as well.
This is definitely not my area of expertise, but I think it could be argued that any analog device which calls for an output transistor to go + and then go - , with the resultant lag (crossover distortion), could be considered "digital", since the output signal is not continuous.
Whether or not it matters in any particular design is another issue.
:)
Except that in a class-B/AB amp, the crossover distortion is presented directly to the output and must be minimized. With class-D, the switched output is averaged over time to give what is (hopefully) a signal without artifacts of the devices that generated it.
You can think of it like a water tank with a steady drain/outflow. To keep it full you could either finely adjust the inflow until it matches the outflow or you could turn the inflow on and off at the right interval to keep the level up, but you won't be able to tell which is happening by just looking at the outflow.
...if the signal is quantized to a finite set.
This doesn't happen in Class-D so it's limits are those of bandwidth, not resolution. It's not really analog in the classic sense either, since devices are being used as binary switches in the signal path.
So when asked, I say Class-D is a weird hybrid.
Yes, it is a weird hybrid.
I cannot fathom any amp sounding better than Class-A...
"I cannot fathom any amp sounding better than Class-A..."
Ear of the beholder.
Three wonderful amps here in the past year...Class A monos, Class D monos, and Class AB stereo. All are great with slightly different strengths but all are musical and any of the three is a "keeper." Weird ending, so far, I sold the Class A monos and purchased the AB and D amps. Still can't decide which of those I like better so will give it a longer audition before deciding which to sell.
All that said, I am not disagreeing with you, Class A done well is magical.
Just look at their square wave response for one reason why. :)
C'mon over so that you may fathom.
Abe and others, thanks much for the info. I've always assumed class D was digital. Anyway, I email my most trusted source and here was the reply:
Hypex is absolutely correct! Switching amps, Class D. They are about as 'digital' as a light switch tuned on and off real fast. In as much as the output devices are either full on, or full off, some would say that is digital, but it is NOT related to any format of making or decoding 'digital audio' signals...That on and off is done very fast, hundreds of thousands of times a second, and the time duration of on, and off during each cycle, is them smoothed out with a filter circuit to end up with an approximation of a variable waveform that "tries" to follow, at large voltage and current, the shape of the variable [analog] input signal which is small. A tiny number of class D amps have an input receiver that can accept some formats of 'digital audio' signals, and process and translate that information, usually be a digital to analog conversion, that can then drive the to run the class D amp input node. Most class D amps are nothing more than a Pulse Width Modulated switching set.The usual means of achieving the variable pulse width is accomplished with 100% analog means, driving a differential comparator with a sawtooth ramp waveform on one input node, and the variable music signal on the other node. No digits ever! I have even done this is the lab with a few analog signal generators and a comparator input on an analog oscilloscope.
Analog,,,actually the whole friggen universe is analog! ...a lot more that a needle being drug around in a spiral groove with lots of woggles. Even the reading with laser beams of a CD or DVD or BluRay is an analog process...The "digital" part does not start until after a large number of signal processing events are done in the analog domain, including light intensity to voltage conversion and a comparitor is driven to produce an ouput that looks sort of like a stream of squared off pulses...THEN an even larger number of processes are done in digital processing to clump certain groups of pulses together, align then up in time, and in specific sequences, so that finally it can be covered into electric waveforms that approximate the music waveform.
Hi,
You must take into account both the dead time and the limited switching frequency and the comparators switching time. This severely limits the possible states.
And I would like to add that using a Ramp Waveform and Comparator is one of the ways to make an AD converter.
Calling this "Analogue" is incorrect, sorry. Unless we call everything under the sun "Analogue".
It is a single Bit modulator (AD) of low order with a power switching stage attached.
A simple way to see what is going on is to extract the error and compare the waveform and spectrum to that of an actual analogue amplifier. If the Class D Amp was "analogue" the error signals would be similar.
I wonder what the error signal of something like this will look like?
You must take into account both the dead time and the limited switching frequency and the comparators switching time. This severely limits the possible states.
Sure there are bounds, but he number of possible states is still infinite. PWM imposes no limitation on the number of signal values which can be represented by the pulse width. PWM is fundamentally an analog waveform. There are Class D amps such as Spectron where the signal stays analog all the way through.
PWM dimmers and motor controllers aren't digital devices either.
And I would like to add that using a Ramp Waveform and Comparator is one of the ways to make an AD converter.
Calling this "Analogue" is incorrect, sorry. Unless we call everything under the sun "Analogue".
It is a single Bit modulator (AD) of low order with a power switching stage attached.
A pulse width modulator is not the same as a delta-sigma modulator. The output of a delta-sigma modulator is a pulse train representing a bitstream where the time averaged frequency of positive values (pulse density) is proportional to the time averaged input signal. Whereas in a pulse width modulator, the duty cycle is determined by the instantaneous value of the input signal. Also, the on-off transitions at the output of a delta-sigma modulator are constrained to occur at regular intervals by a clock, whereas the on-off transitions at the output of a pulse width modulator are determined by the duty cycle. The fundamental switching frequency of a pulse width modulator is determined by the clock, but the output values are not clocked in the digital sense. So on one hand, you have a clocked stream of bits and the number of distinct signal values it can represent is limited by the integration time. It's digital. On the other hand, you have a modulated periodic waveform but it is not a clocked signal and the number of signal values it can represent is infinite. It's analog.
I've recently been thinking about that question as well. I'm leaning to the notion that binary voltage amplitude class-D (PWM, PDM, etc.) is actually a form of analog. I see them as analog in so much as the output waveform they produce is proportional to the input signal, just not in the expected amplitude sense. Their outputs are essentially a clipped modulated carrier, looking much like a modulated FM waveform thats been clipped. The proportional nature of class-D modulation, prior to low-pass filtering, is usually readily apparent when viewed in graphical form. While a class-D waveform's amplitude obviously is not proportional, other aspects of the output waveform (depending on the particular form of class-D utilized) very much appear to be.My feeling is that the real advantage of class-D, aside from it's high efficiency, is that it utilizes power transistors in the mode in which they naturally want to be used, non-linearly. So, rather than taking great pains to force the transistors to operate in a highly linear manner, class-D directs them to operate in a highly non-linearly manner. Either being on or off, with no in-between states of conduction.
Hi,
> I just read somebody claiming Class D is not digital.
They do that to pretend it is something it is not.
> Although there may be the odd Class D amp that is all analog,
That it is not Class D.
> it seems to me that traditional or common class D amps
> are indeed commonly known as digital with at least one
> conversion of the signal from analog to do digital and
> back again
Yup.
Strictly speaking any Analogue input Class D Amplifier is a Delta Sigma A2D converter followed by a power stage and a low-pass filter. With a feedback loop around the A2D.
Anyone who says different is trying to talk away the Class D Stigma.
It's not digital. It is a coding process where the input is translated into square waves but the length of the square wave can vary all over the place. To be digital the coding would be only into two values, 0 or 1(on or off). But like digital a simple low pass filter decodes things, the only similarity.
Thanks for your responses.
I've owned Class D amps from 2006 - 2011 and again from 2014 - preenst day.
I've not researched them lately, nor have I felt the need until I was recently bombarded with folks saying Class D is not digital and referenced Bruno Putzeys at http://www.hypex.nl claiming class D is not digital.
Shoot, I assumed the entire industry considered class D as digital for the last decade or so.
Have things changed or has the industry become more mature with class D with recent discoveries?
I'm at a complete loss of words regarding this assumption as well as with what I assumed were digital distortions that digital amps induce and I experience.
...Class D is 'non-traditional' in the sense that these amps do not operate their output devices in a linear fashion as do Class A, Class AB, Class B. The output devices are either ON or OFF but modulated by the input signal.
In my experience, there are some pretty good sounding Class D amps out there BUT IMHO, if they cost as much as good sounding Class A or Class AB then what's the point? ...unless the point is simply to have efficiency gains in terms of power consumption and space savings.
When Class D sounds as good or better at a LOWER price, then it starts making sense.... to me.
NT
....depending on the particular design.
patented as the Class T amp to differentiate it from "conventional" class D.
I guess there is a rule to every exception.
dee
;-D
True terror is to wake up one morning and discover that your high school class is running the country.
quote by Kurt Vonnegut
...the cost! I have the Wyred 4 Sound mAmps and absolutely love them. They replaced my old heavily modified McCormack amp (which is a great amp) and are better sounding to me in just about every respect and cost only $1800 bucks! Those mega-buck class-D amps may sound great but the cost seems really unreasonable.
David, I own W4S' SX-1000's and based on some things I'm able to do at my rack with them, I've been able to take them to extraordinary heights.I used to own McCormack DNA-2 Rev A that I thought was the cat's meow compared to the DNA-2 Anniversary Edition I owned before that.
But one day in 2006, a very knowledgeable friend and distributor insisted I receive a 14 lbs. package from FedEx containing 2 nuforce mono-block amps to simpy compare to the DNA-2 Rev A. 3 weeks later I sold the Rev A.
The SX-1000's replaced my quite musical $8k BMC C1 int. amp which in comparison to the SX's sounds like a $200 BestBuy receiver.
nope
I agree. Many of the very expensive Class D amps still use mostly off-the-shelf modules so unless they sprinkle very costly magic pixy dust on the modules, I don't see the value.
When Class D amps cost several thousand dollars, there's no incentive to buy them.... just buy a good Class A amp.
Just an personal anecdote regarding the sometimes reality of pixie dust. I own a Tripath chip based class-D power amp which I built myself from scratch. The Tripath reference design specified a certain metalized polypropylene capacitor for use in the output LC filter. I originally built my amp using the suggested cap. Later, out of curiosity, I replaced those with double-metalized polypropylene caps. (which have metalization attached to both sides of the dielectric film) intended for high dV/dT pulse handling duty. The clarity of the sound improved, and not by a little bit. As always, the devil (or the angel, it would seem) is found in the details.
Got any pics of your home-grown amps, Ken?
BTW, my W4S amps use the ICE-Power technology or boards.
Hi, stehno,Actually there isn't much that's interesting to see. The chassis, main power transformer and filter caps. are all reclaimed from a Forte' 200W stereo power amp. The orginal circuits boards of the amp were fried and the chassis scratched, so I decided to have the chassis sandblasted and re-anodized (still black), which also removed the original markings of course. I also changed the original red power LED to a subdued deep blue LED, giving a very nice overall stealth amp effect.
Tripath supplied PCB Gerber files from one of their reference designs, which I then sent to a board house for manufacture. The reference design PCBs are typically configured for 2-channel stereo use, but the documentations also showed how the stereo channels of a single PCB could be configured as a balanced H-bridge mono channel, which is what I did. I hand stuffed two such PCBs, including winding the output inductors on powdered iron toroid cores with heavy gauge solid core wire by hand (which I would not choose to do again). The finished PCBs are mounted directly to the exisitng side-panel style heat-sinks, although being class-D, the now 650W@2 ohms capable amp remains eminently cool when powering my 93dB sensitive speakers.
Call me tone deaf, but the sound quality after some seemingly minor passive component changes, such as I've mentioned, is stunningly clear, dynamic, and natural. When I think of all the processing involved, this result seems quite surprising to me. This is, by far, the best sound I've yet obtained from a power amp, including the original Forte' (Insulated-Gate-Bipolar-Transistor) model I cannibalized, and a pair of Audio Research Classic 120 tube monoblocks I owned for several years.
Whose IGBTs did you use? I worked for IR for a decade plus and worked in-fab for the technical people for all that time.
Just curious.
Too much is never enough
Pretty cool, Knewton. I can't even solder well.
If you're up for experiments would you consider 2 suggestions?
1) Assuming it's stock, remove the Monster power thing from the mix and tell me what you hear.
2) It looks like your gear's on a concrete slap. If so, and after testing without the Monster power, remove the Monster and the bottom plate. Get some little steel pucks, even 3 or 4 car lug nuts to use as footers between the amp chassis and the floor. Try to put a couple of the lug nuts underneath near the mounting studs of the PCB boards, and put about 50 lbs. of some heft on top of your amp. In the end, it should be metal to metal to concrete.
You should notice some gains right away, but give it a week or so and I'll bet dollars to donuts it will probably change everything you thought you knew of audio.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: