|
Post by thorcorps on May 6, 2015 20:19:00 GMT -5
Sounds like just the thing I'd like to pick up and try in my Bedroom system. Dual mono Class D's to go with my DC-1.
As long as they can handle 4 ohm loads without wimping out in the bass department, I'm in. Call me when they're ready to ship.
oh, and XLR connections, of course.
|
|
KeithL
Administrator
Posts: 9,941
|
Post by KeithL on May 6, 2015 20:41:10 GMT -5
A perfect square wave is composed of a base frequency and ALL odd order harmonics - in the correct phases and proportions (in practice, in order for it to look pretty good, you need at least the first four or five of them). This means that a 1 kHz square wave is made up of 1 kHz, 3 kHz, 5 kHz, 7 kHz, 9 kHz, 11 kHz, etc. A 20 Hz square wave is made up of 20 Hz, 60 Hz, 100 Hz, 140 Hz, etc. And a 10 kHz square wave is made up of 10 kHz, 30 kHz, 50 kHz, 70 kHz, 90 kHz, 110 kHz, etc. So you really need to specify WHAT FREQUENCY of square wave you're talking about. As you can see, it wouldn't be too difficult for a speaker to reproduce the first six harmonics needed to make a plausible 1 kHz square wave. What speakers usually miss is that they get the proportions or phase a little bit wrong, which sounds about the same, but looks a lot different. Also, as you can see, while many audio amplifiers could do the 10 kHz square wave, there aren't any speakers I know of that could do it. For diagnostic purposes, when looking at amps, you look at the shape of the wave. At high frequencies, incorrect proportions of the upper harmonics will show up as either ringing or rounding off on the edges of the square waves. For very low frequencies, you can look at the top of a low frequency square wave. It should be flat and horizontal but, if the low frequency response isn't flat, it will be tilted. (Essentially, for an amplifier, it's a nice fast way to get an idea about the frequency response without taking a lot of measurements.) However, since speakers often exhibit a huge amount of phase shift, which may often be inaudible, a square wave can be very misleading (a speaker with a phase response that varies smoothly over several hundred degrees may sound very good, but it's square waves will be totally unrecognizable, and a speaker with lots of plain old distortion can still make a decent looking square wave). As with most diagnostic tests, you have to know how to read the results. But, then, you won't see a very nice looking square wave coming from your speakers either; luckily for us we don't listen to square waves anyway. As the article said, square waves can be a handy diagnostic tool (although, lately, they've mostly been replaced with impulse tests). With analog amplifiers, how well the amplifier can reproduce very high and very low frequencies (which is what square waves show) is often an indicator of how well it can reproduce audio frequencies. (More specifically, if you know how to read them, certain flaws in square wave response can point out flaws inside the audio band which may well be audible.) Back when being able to reproduce very high and very low frequencies was difficult from a design point of view, this was very useful information. However, modern designs often have the exact opposite issue - the bandwidth needs to be deliberately limited to prevent the amplifier from amplifying things like the high frequency noise leakage from your cordless phone. Therefore, while being able to reproduce a designed-for frequency band is important, the old idea of "more is better" isn't necessarily true any more. With Class D amplifiers, this is even more true. A Class D amplifier is designed to reproduce a certain range of frequencies (audio), while the output section is carefully designed NOT to pass frequencies outside the desired frequency range. While you certainly could design a Class D amplifier that could produce nice looking 10 kHz square waves, doing so would make the amplifier a lot more expensive to design and produce, but wouldn't necessarily improve its performance in the audio band - or make it sound better. To put it bluntly, while square waves can be a handy and useful diagnostic tool, being able to reproduce clean square waves simply is NOT a design requirement for a good sounding audio amplifier. Apparently...a certain magnepan (or was it quad?) speaker could actually reproduce something of a square wave.
|
|
|
Post by Gary Cook on May 6, 2015 21:41:48 GMT -5
At high frequencies, incorrect proportions of the upper harmonics will show up as either ringing or rounding off on the edges of the square waves. Audible Keith? If so, perhaps this is what annoys me when listening to Class D amps. Cheers Gary
|
|
KeithL
Administrator
Posts: 9,941
|
Post by KeithL on May 6, 2015 22:00:14 GMT -5
Actually, it's a bit of both. Conceptually a Class D amplifier shares a lot with a DAC. The analog input signal is converted into a digital signal, which is boosted in level (very efficiently) by using switching components, then the resulting (bigger) digital signal is converted back into analog (which requires that the artifacts of the conversion process be filtered out). However, unlike a DAC, a Class D power amplifier must work with much larger signals. The switches used to do the amplification must be able to handle large currents and voltages, and be able to switch quickly enough to do it efficiently (if a switch is too slow, any time it spends between full-on and full-off translates to inefficiency and heat). And, once the signal has been boosted, the filters that eliminate the sampling artifacts must also handle large amounts of power, and do so efficiently. To take an example, to make a Class D subwoofer amplifier that can handle frequencies between 20 Hz and 200 Hz, we would probably want to use a switching frequency of 5 kHz or more, and, in order to be able to switch clean square waves at 5 kHz, our switches would have to be able to handle several amps at 100 kHz or more. (And, for full range audio, we would want to switch at square waves at several hundred kHz or more, which would mean being able to switch tens of amps in the mHz frequencies.) Switching transistors (and MOSFETs) that can handle those voltages and currents at the necessary frequencies have indeed gotten much cheaper and more easily available. We've also seen the appearance of complex integrated circuits that handle most of the control functions associated with the process - and, again, they've gotten cheaper and more readily available. (We're talking a drop from several dollars to a few cents in some cases.) In addition to the parts themselves, switching amplifiers are dealing with significant amounts of power at megahertz frequencies. When you design a Class A/B amplifier, you're mostly concerned with analog audio performance. When you design a Class D amplifier, to that you can add concern about things like whether your device will interfere with a neighbor's TV, and whether it will meet FCC noise requirements, and whether moving a wire a half inch one way or the other may cause your circuit to fail. (Remember that, with a Class D amplifier, you're dealing with frequencies that reach up past AM radio.) As the parts are becoming more easily available, and more people are designing and working with Class D equipment, the expertise necessary to deal with these extra concerns is also becoming more available. So, it's easier to get the parts you need, and they're getting cheaper, and there's more information and expertise available to help you use them... and there's enough market to keep the volume up (and the cost down). As with most technological things, there's no magic.... it's simply a matter of progress and money. Forty years ago a 1 watt LASER cost $100k and weighted 200 pounds; today you have one inside the $29 DVD burner in your computer. Class D amplifier technology has been around for a very long time; there were even Class D amplifier designs that used tubes. It's just taken until now for us to be able to make a Class D amplifier with performance (and sound quality) that matches the more common types of analog amplifiers, but with the added benefits of Class D size and efficiency, and still keep the price affordable. (Face it, all that talk about saving ten cents worth of electricity every hour you run it is nice, and being smaller and lighter does make it easier to dust, but you wouldn't actually pay thousands of dollars more up front to get either of those benefits.) Up until now we hadn't been able to find any Class D amplifier designs that met our combined requirement of good sound and reasonable cost (we're not interesting in building an amp that doesn't sound good, or in trying to one that sounds good, but that you need a mortgage to afford). The design options available for Class D amplification have finally reached the point where they constitute a viable alternative. Keith, now you got me curious . If I may ask, what kind of design options have developed more recently that were not available in the past? Are you talking about components availability or rather about the "collective experience" of the design community that has been building over the many years since the first commercial Class-D designs, which allows now to design a good sounding, yet affordable amp in such topology?
|
|
KeithL
Administrator
Posts: 9,941
|
Post by KeithL on May 6, 2015 23:01:06 GMT -5
You're missing the point. The same flaws that show up in an oddly shaped square wave also show up in a plain old frequency response sweep, and they're easier to make sense of when you see them in that context. (Saying that a certain harmonic isn't there in proper proportions is just another way of saying that the frequency response dips at that frequency.) A Class D amplifier has a specified frequency response - just like any other amplifier - and there's no reason for a Class D amplifier to sound specifically different than any other sort of amplifier with similar specs. It is true that most Class D amplifiers have a frequency response limited to the audio band while most Class A/B amps have much wider bandwidth - which could possibly account for some difference in how they sound. (In analog circuitry, being able to handle a wider range of frequencies than necessary is considered to be a sort of performance safety margin; in digital circuitry, the considerations are different. It is an interesting argument - for somewhere else - whether it makes any sense to have an amplifier that can pass 80 kHz when the source you're using is a CD (which, by standard, is absolutely limited to 22 kHz).) In fact, many older Class D amplifiers weren't able to cover the entire audio band. Sub amps often only went to 500 Hz or so, and even many "full range" amps were rated "20-20k +/- 3 dB" (any amplifier that can't handle the audio band without going 3 dB out of flat is pretty sad - and a 3 dB fault in frequency response might well be very audible). However, a bigger problem is that the output filters required and used by most Class D amps are very sensitive to load impedance. This means that the output performance you get in real life can vary quite a bit from what you read on the spec sheet - because your real world speaker isn't anything like the pure resistive load the filter on the amplifier's output was designed for. This almost certainly accounts for the wide differences in how well early Class D amps performed and sounded in different situations. Many of the better quality (and better sounding) models in the current Class D amplifier space have found effective ways of avoiding this issue. An amplifier should be a black box - you shouldn't care whether it's Class A/B or Class D, and you shouldn't be able to hear the difference. As long as there is an audible difference, the one or the other isn't performing up to spec. At high frequencies, incorrect proportions of the upper harmonics will show up as either ringing or rounding off on the edges of the square waves. Audible Keith? If so, perhaps this is what annoys me when listening to Class D amps. Cheers Gary
|
|
|
Post by Gary Cook on May 7, 2015 0:06:22 GMT -5
You're missing the point. The same flaws that show up in an oddly shaped square wave also show up in a plain old frequency response sweep, and they're easier to make sense of when you see them in that context. (Saying that a certain harmonic isn't there in proper proportions is just another way of saying that the frequency response dips at that frequency.) A Class D amplifier has a specified frequency response - just like any other amplifier - and there's no reason for a Class D amplifier to sound specifically different than any other sort of amplifier with similar specs. It is true that most Class D amplifiers have a frequency response limited to the audio band while most Class A/B amps have much wider bandwidth - which could possibly account for some difference in how they sound. (In analog circuitry, being able to handle a wider range of frequencies than necessary is considered to be a sort of performance safety margin; in digital circuitry, the considerations are different. It is an interesting argument - for somewhere else - whether it makes any sense to have an amplifier that can pass 80 kHz when the source you're using is a CD (which, by standard, is absolutely limited to 22 kHz).) In fact, many older Class D amplifiers weren't able to cover the entire audio band. Sub amps often only went to 500 Hz or so, and even many "full range" amps were rated "20-20k +/- 3 dB" (any amplifier that can't handle the audio band without going 3 dB out of flat is pretty sad - and a 3 dB fault in frequency response might well be very audible). However, a bigger problem is that the output filters required and used by most Class D amps are very sensitive to load impedance. This means that the output performance you get in real life can vary quite a bit from what you read on the spec sheet - because your real world speaker isn't anything like the pure resistive load the filter on the amplifier's output was designed for. This almost certainly accounts for the wide differences in how well early Class D amps performed and sounded in different situations. Many of the better quality (and better sounding) models in the current Class D amplifier space have found effective ways of avoiding this issue. An amplifier should be a black box - you shouldn't care whether it's Class A/B or Class D, and you shouldn't be able to hear the difference. As long as there is an audible difference, the one or the other isn't performing up to spec. Nah, I got your point (ie; frequency response dips) I'm just searching for the reasons why I have yet to hear a Class D amp that I really like the sound of. Perhaps they are actually performing to their spec and I'm just sensitive to the frequency response dip/s that is/are in their specs. Cheers Gary
|
|
|
Post by AudioHTIT on May 7, 2015 0:32:51 GMT -5
I would never, ever want to listen to a square wave....... Actually back in the 70's I remember playing with a Buchla Synthesizer, it had the ability to product most any kind of waveform you wanted, including a square wave. On some occasions I liked the sounds I could get from a pure square wave, like KeithL explained, rich in odd harmonics and rather brash. Most of the time though a triangle wave gave a similar feel with less offense (same odd order harmonics, but decreasing logarithmicly). Anyway, not a common thing but I'm sure there are some recordings out there.
|
|
|
Post by wizardofoz on May 7, 2015 1:46:08 GMT -5
Im playing with an old roland Jupiter 6 synth at the moment ...it's got some 'issues' and trying to find the problem is proving to be fun and games
Sorry a bit off topic...and now back to usual content
|
|
|
Post by bolle on May 7, 2015 1:48:35 GMT -5
Don´t forget that the preamp and the DSP also alters the signal. You can visualize this very good with ... a square wave! It is interesting what effect DSP on or off etc. has with most processors regarding rise time, ringing etc.
|
|
klinemj
Emo VIPs
Honorary Emofest Scribe
Posts: 14,747
Member is Online
|
Post by klinemj on May 7, 2015 6:05:09 GMT -5
I would never, ever want to listen to a square wave....... It's hip to be square. Mark
|
|
klinemj
Emo VIPs
Honorary Emofest Scribe
Posts: 14,747
Member is Online
|
Post by klinemj on May 7, 2015 6:11:13 GMT -5
I agree. I would have wished they stuck to a more robust u-series. We went from the beefy UPA-2, UPA-5 to the more petite in build UPA-200 and now we have what I assume are tiny class D amps (DC-1 chassis). Hopefully their price point and performance will be good. There are plenty of cheap-a** class D amps on the market. Crown is a good example. I hope that their offerings will be competitive. They know their arket I guess. But the UPA-2 was THE reason I joined the Emotiva family because they were the only game in town offering so much power at a good price and price was a key factor for me. At that time, if I was shopping and saw a tiny class D amp, I would have glossed over it and gone for more larger looking amps. And I'd have missed all that's great about Emotiva with that one move. But that's just me..... there some type of market for small class D amp that can sit on the bottom of a dc-1. Don't assume tiny to mean not robust or low power. My nCore monoblocks each have an amp circuit that is an 88 mm circle, and the power supply is tiny...they whole thing in a box can be DC-1 sized. My boxes are bigger - only for aesthetic reasons as I hang them on the wall. They look great and the sound is outstanding at any volume I dare to use. I do look forward to what Emotiva has to offer in the Class D range. If the price and size are right, I would not mind swapping out my XPA-5 for a small 5 channel Class D. I have considered building a 5 channel nCore to replace the XPA-5, but that would be fairly pricey and for duty on surrounds and center - I don't want to spend that much. On a related tangent...with various things coming down the pike that drive more and more channels in surround (like Atmos and such), it will become a practical limitation to have smaller yet very powerful amps that use minimal power. Class D is a natural for these applications. Emotiva is making a smart move into Class D and it appears very well timed. That said - not sure I will be adding a lot of channels any time soon, but it's smart for them to make the amps available! Mark
|
|
|
Post by audiobill on May 7, 2015 6:44:10 GMT -5
I would never, ever want to listen to a square wave....... Actually back in the 70's I remember playing with a Buchla Synthesizer, it had the ability to product most any kind of waveform you wanted, including a square wave. On some occasions I liked the sounds I could get from a pure square wave, like KeithL explained, rich in odd harmonics and rather brash. Most of the time though a triangle wave gave a similar feel with less offense (same odd order harmonics, but decreasing logarithmicly). Anyway, not a common thing but I'm sure there are some recordings out there. I have a function generator if I want to listen to test signals, and Tidal for music!
|
|
|
Post by rogersch on May 7, 2015 7:50:56 GMT -5
. I have considered building a 5 channel nCore to replace the XPA-5, but that would be fairly pricey and for duty on surrounds and center - I don't want to spend that much. I can assure you it is worth the investment.....
|
|
KeithL
Administrator
Posts: 9,941
|
Post by KeithL on May 7, 2015 10:07:03 GMT -5
While your facts are correct, your philosophy is not quite.... If you try to record a 10 kHz square wave onto a CD, and then play it back, it won't come out looking like a 10 kHz square wave, and the same is true for a DSP, or for a Class D amplifier. However, this doesn't signify that there's something wrong. In fact, since a CD is specifically supposed to be band-limited to 22.5 kHz, it would be impossible for a CD to properly reproduce a 10 kHz square wave (and, in fact, it would be functioning incorrectly if it appeared to do so). To put it bluntly, it happens to be true that, if a Class A/B amplifier can't produce a good looking square wave, it often suggests that there are other design flaws that may be audible. The reality is that, with this type of design, flaws that occur outside the audio band, and so are themselves inaudible, may indeed be symptoms of flaws inside the audio band - which may be audible. (If I wanted to, I could design a linear amplifier that worked very well inside the audio band, but was unable to reproduce good looking square waves - it usually just doesn't work out that way.) To put that another way: you can sometimes "visualize" how a linear amplifier will sound from its square wave response; you can also tell some things about how a digital device will sound from its square wave response - HOWEVER, the information you can ascertain in each case is not at all similar. DACs and digital amplifiers are designed to operate over a limited bandwidth. It's not a mistake, or a flaw (unless someone makes a bad choice or fails to meet their design goals). If someone wanted to, they could design a Class D audio amplifier that had frequency response up to 200 kHz or so, and so could produce good looking square waves. However, since doing so would increase the cost and complexity considerably, and wouldn't necessarily improve the performance inside the audio band in any measurable or audible way, nobody's going to bother to do it. Class D amplifiers are designed to perform well between 20 Hz and 20 kHz (or maybe a bit wider range for "safety margin"). To put it yet another way, a lot of the information you can try to puzzle out about a digital system from its square wave response is redundant. You can tell from the square wave response of a CD player that it can't reproduce frequencies above 22 kHz... but we already know that... because, by spec, a Red Book CD player is supposed to be band-limited at 22 kHz. So, "figuring this out" by looking at its square wave response is really moot. (The type of ringing you observe with square waves with that CD player is meaningful, and it tells you things about the digital filters it uses, but the information you get isn't at all comparable to the information would would get from the square wave response of an analog amp.) Likewise, a DSP is designed to be able to process a certain range of frequencies (the audio band, generally accepted to be from 20 Hz to 20 kHz). Being able to process a high frequency square wave "properly" is simply not a design goal - or necessary in order for it to sound good. "Alterations of the signal outside the audio band" SHOULD NOT be audible - unless they also indicate alterations INSIDE the audio band - where you CAN hear them. Don´t forget that the preamp and the DSP also alters the signal. You can visualize this very good with ... a square wave! It is interesting what effect DSP on or off etc. has with most processors regarding rise time, ringing etc.
|
|
KeithL
Administrator
Posts: 9,941
|
Post by KeithL on May 7, 2015 10:30:20 GMT -5
It's quite possible that's it.... and you also have other things that aren't always mentioned in the standard measurements (like what the damping factor is at different frequencies, and how the specs change with different types of loads). We agree entirely. Which is why, until now, we haven't offered any Class D amplifiers..... However, finally, there are a few practical designs available that actually sound good enough to join the Emo product line You're missing the point. The same flaws that show up in an oddly shaped square wave also show up in a plain old frequency response sweep, and they're easier to make sense of when you see them in that context. (Saying that a certain harmonic isn't there in proper proportions is just another way of saying that the frequency response dips at that frequency.) A Class D amplifier has a specified frequency response - just like any other amplifier - and there's no reason for a Class D amplifier to sound specifically different than any other sort of amplifier with similar specs. It is true that most Class D amplifiers have a frequency response limited to the audio band while most Class A/B amps have much wider bandwidth - which could possibly account for some difference in how they sound. (In analog circuitry, being able to handle a wider range of frequencies than necessary is considered to be a sort of performance safety margin; in digital circuitry, the considerations are different. It is an interesting argument - for somewhere else - whether it makes any sense to have an amplifier that can pass 80 kHz when the source you're using is a CD (which, by standard, is absolutely limited to 22 kHz).) In fact, many older Class D amplifiers weren't able to cover the entire audio band. Sub amps often only went to 500 Hz or so, and even many "full range" amps were rated "20-20k +/- 3 dB" (any amplifier that can't handle the audio band without going 3 dB out of flat is pretty sad - and a 3 dB fault in frequency response might well be very audible). However, a bigger problem is that the output filters required and used by most Class D amps are very sensitive to load impedance. This means that the output performance you get in real life can vary quite a bit from what you read on the spec sheet - because your real world speaker isn't anything like the pure resistive load the filter on the amplifier's output was designed for. This almost certainly accounts for the wide differences in how well early Class D amps performed and sounded in different situations. Many of the better quality (and better sounding) models in the current Class D amplifier space have found effective ways of avoiding this issue. An amplifier should be a black box - you shouldn't care whether it's Class A/B or Class D, and you shouldn't be able to hear the difference. As long as there is an audible difference, the one or the other isn't performing up to spec. Nah, I got your point (ie; frequency response dips) I'm just searching for the reasons why I have yet to hear a Class D amp that I really like the sound of. Perhaps they are actually performing to their spec and I'm just sensitive to the frequency response dip/s that is/are in their specs. Cheers Gary
|
|
|
Post by MusicHead on May 7, 2015 10:35:48 GMT -5
Keith, thanks for replying to my initial question and for your other follow-up posts. As usual, thorough but accessible and well articulated. Ever thought of writing a book on audio engineering?
|
|
|
Post by Gary Cook on May 7, 2015 13:44:20 GMT -5
Keith, thanks for replying to my initial question and for your other follow-up posts. As usual, thorough but accessible and well articulated. Ever thought of writing a book on audio engineering? The XMC-1 Manual is as big as most books, so he kinda already does. Cheers Gary
|
|
|
Post by Dark Ranger on May 7, 2015 13:45:13 GMT -5
I do look forward to what Emotiva has to offer in the Class D range. If the price and size are right, I would not mind swapping out my XPA-1 for a small 5 channel Class D. It's the new tagline for Emotiva's XPA-1 monoblock: so much power, it'll drive FIVE channels! On topic: although I prefer Class A and A/B designs for my primary systems, I'll be very interested to see how the new Ultra series works out. Maybe we'll even get an integrated unit fit for a small office or bedroom. I'd buy one of those.
|
|
|
Post by garbulky on May 7, 2015 14:14:18 GMT -5
A class D amp would be ideal for office setups especially in the size of the DC-1. So maybe a pro series amp would be a better branding for it?
|
|
klinemj
Emo VIPs
Honorary Emofest Scribe
Posts: 14,747
Member is Online
|
Post by klinemj on May 7, 2015 17:23:15 GMT -5
Dark RangerGood catch! But as noted...don't think class D as a second class citizen anymore. I got rid of my XPA-2 with class D monoblocks. Much smoother/less harsh sound. Same for you garbulkyMark
|
|