New Emersa amps vs XPA gen2
Dec 27, 2015 14:45:12 GMT -5
mdanderson, knucklehead, and 5 more like this
Post by KeithL on Dec 27, 2015 14:45:12 GMT -5
I'm sorry, but, while opinions are all well and good, some things just call for facts.... I would agree with you that it's a bit more difficult to design a good SMPS than to design a good linear power supply (although I've also seen plenty of badly designed linear power supplies over the years), but I am a little less clear on where it "appears" that a SMPS "doesn't have the dynamic power capability of a good linear". (I'm not picking on you in particular, but there are an awful lot of "facts" floating around that are actually based on things like outdated information, generalizations that aren't valid, and just plain old wives tales.)
While both Class D amplifiers and switch mode power supplies (SMPS) have actually been around for quite a while, both are relative newcomers to the world of mainstream audio equipment. As with any relatively new technology, there is a lot of outdated information floating around, and many people who are otherwise well versed in electronics don't understand them very well; this leads to both bad information and bad designs. I would advise everyone to be at least a little careful about what you believe, and what you repeat, so as to avoid spreading bad information around any further.
I'm going to address just a few common "facts" here:
Power supplies.
There are all different sorts of power supplies. The main power supply in a power amp, which supplies large amounts of power to the output stage (filtered by not regulated), is in fact usually quite efficient. While a SMPS may be slightly more efficient in this application, because a linear power supply is usually pretty efficient to begin with, that isn't usually an important benefit. The main benefits are that a SMPS weighs a lot less, and, because big transformers are expensive to buy, and expensive to ship, costs less. The smaller power supplies used in the other parts of an amplifier, and in other equipment, operate at lower current levels - and usually are regulated. In these applications, a SMPS usually is more efficient, sometimes a lot more efficient, but since the amount of power involved is small, this isn't a big deal either.
The reason that SMPS operate at very high frequencies is efficiency and component size. Because of the much higher frequencies used, the small transformers and filter capacitors used in a typical SMPS deliver BETTER performance than the much larger components used in a power supply that must operate at 50 Hz or 60 Hz. While it is true that a typical SMPS will generate a lot more high frequency noise internally than a linear power supply, which makes design and layout more critical, that high frequency noise is simpler to filter out than the 50 Hz and 60 Hz hum that is present in all linear supplies. (However, because the requirements are different, if an engineer who isn't familiar with the new technology tries to design an SMPS, the result is likely to be poor performance.)
Your statement about SMPS lacking in "dynamic power" is somewhat true - at least for some designs. However, it isn't a limitation of the technology, but is merely simply the design target for many early SMPS designs. In many applications, including computers, and many others which were typical applications for SMPS until recently, the ideal power supply would deliver perfectly clean and regulated power up to a set limit, at which point it would limit or shut down to protect itself and the associated circuitry... or, to say that another way, it would supply lots of perfectly clean power up to its rated limit, but have no dynamic power reserves whatsoever. (Which would you rather have, a 1000 watt power supply that can deliver peaks of 2000 watts, or a power supply that has no dynamic reserves at all, but can simply deliver exactly 2000 watts all the time?) Well, the reality is that you CAN design an SMPS that behaves more like a linear power supply, with a specified average power ratings, AND dynamic power reserves... which is what we used in the new X Gen3 series amps.
When you read reviews, or comments from armchair engineers, about how "this or that SMPS doesn't have enough dynamic power reserves" - they're probably right. But the reason is because the particular power supply they're talking about wasn't powerful enough to do the job continuously AND didn't have any dynamic power reserves that would allow it to exceed its continuous ratings short-term. (In other words, whoever chose or designed the power supply made a bad choice for use in that application.) We've been very careful NOT to make that mistake with the X Series Gen3 amps.
Class D amplifiers
Just to set the record straight.... There is no inherent limitation in Class D that limits frequency response, distortion, damping factor, output power, efficiency, or sound quality. Because they operate at high frequencies, and by using switching output stages, most Class D amplifiers are very efficient. However, while this is a major benefit, it is also a huge liability - because many Class D audio amplifiers were expressly designed with efficiency as their primary design goal. Yes, it really is that simple..... As with SMPS, designing Class D amplifiers is quite complex, there are a whole slew of possible design trade-offs, and many engineers who are familiar with older technologies are a lot less familiar with designing Class D equipment. As a result, a lot of early Class D audio amps met their primary design goal (efficiency), but failed miserably to sound good.
Combinations and permutations
In case there was any doubt, you can use either type of power supply with either type of amplifier. And, in the recent past, there have been Class D amplifiers that used linear power supplies... As with most things, this wasn't because of some deep mystical knowledge, but simply because, at that point in time, a linear power supply could deliver the power that was needed, at the required quality level, cheaper than using an SMPS.
The reason we chose to use a SMPS, but Class A/B output modules with Optimized Class H rail switching, in our new X Series Gen3 amps is simply that this combination gave us the best combination of performance specifications, sound quality, and cost. (As you probably know by now, we at Emotiva do our best to avoid following fads and trends, and buying this or that technology because "it's popular this year". We prefer to use the technology that offers the best value. Likewise, at the price point we chose for the Emersa line, the all-switching iCE modules provided the best overall value.)
Exploding myths
I love exploding things.... but, sadly, I've found the explosions caused by myths to be far less than aesthetically satisfactory.
Every year last year's myths get exploded, only to be replaced by new myths, which themselves get exploded next year. It's getting so you can;t walk down the street without tripping over, or hearing about, how this or that myth is now known to be untrue, and how "now we know the truth" (or was it "the truth is out there"?)
It is in fact quite possible to get very high efficiency from Class D amplifiers (I've seen designs that claim 95% at full power), and many current designs do in fact deliver higher efficiency than Class A/B designs. It is also quite possible to design Class D amplifiers with low efficiency. The real claim is that Class D amps CAN be very efficient - whereas a Class A/B amp with a linear fixed-rail power supply can never exceed about 70% efficiency, and only that at one specific load impedance . Incidentally, I'm pretty sure that Hypex is now recommending the use of SMPS with their amplifiers (I believe they used to recommend linear supplies).
(But, if you want to get into myths.... then you should consider all the myths about how, rather than using the circuit topology recommended by the manufacturer, who just might know what they're talking about, so many people seem determined to alter things to fit the requirements of this or that audiophile myth about some specific component or circuit topology sounding better than another. I wonder how many of those myths will be exploded spectacularly next year, or the year after.)
I STRONGLY urge you to avoid trying to armchair engineer how you THINK an amplifier with an SMPS or a linear power supply will or should sound.
After all, all that really counts is how it DOES sound, right?
The performance and sound quality of our new X Series Gen3 amps will speak for itself - literally.
Linear VS SMPS is one of those 'third rail' topics. However, one thing IS certain. It would appear that SMPS does not have the dynamic power capability of a good linear. I also would guarantee that the design of a GOOD SMPS is more difficult with problems possible with HF hash distributed thru the audio band being one of several 'possibles'.
People ALSO simply parrot the 'high efficiency' claim. This is TRUE, but at full power where Plug to Speaker efficiencies can easily exceed 80%. (W/SMPS) The output section ALONE can be in the 90s since it is either ON or OFF and the On-Resistance of the devices used is very small. This is one way to get (besides NFB) a good damping factor. At 'realistic' powers, the efficiency of the entire amp from plug to speaker is really not much better than a well executed A/B design.
www.ecnmag.com/article/2010/01/exploding-efficiency-myth-class-d-amplifiers
More complex solutions CAN yield higher efficiency at realistic power levels, but cost goes up fairly quickly and possible problems multiply. It is simply a guess on my part, but bringing in an EMOTIVA amp with a sophisticated SMPS AND using non-PS included modules from B&O would be quite a challenge at prices Emotiva Customers would be willing to pay. Jeff Roland, who makes some wacky expensive 'd' amps would have NO such problem. And I believe he uses N-Core from Hypex.
If I was an HT kind of guy, I'd for sure look at the XPA-5 which for the 800$ 'ask' is verging on 'bargain'. Even using 4 of the 5 channels to BiAmp a stereo pair of speakers would be a good use of such an amp.
While both Class D amplifiers and switch mode power supplies (SMPS) have actually been around for quite a while, both are relative newcomers to the world of mainstream audio equipment. As with any relatively new technology, there is a lot of outdated information floating around, and many people who are otherwise well versed in electronics don't understand them very well; this leads to both bad information and bad designs. I would advise everyone to be at least a little careful about what you believe, and what you repeat, so as to avoid spreading bad information around any further.
I'm going to address just a few common "facts" here:
Power supplies.
There are all different sorts of power supplies. The main power supply in a power amp, which supplies large amounts of power to the output stage (filtered by not regulated), is in fact usually quite efficient. While a SMPS may be slightly more efficient in this application, because a linear power supply is usually pretty efficient to begin with, that isn't usually an important benefit. The main benefits are that a SMPS weighs a lot less, and, because big transformers are expensive to buy, and expensive to ship, costs less. The smaller power supplies used in the other parts of an amplifier, and in other equipment, operate at lower current levels - and usually are regulated. In these applications, a SMPS usually is more efficient, sometimes a lot more efficient, but since the amount of power involved is small, this isn't a big deal either.
The reason that SMPS operate at very high frequencies is efficiency and component size. Because of the much higher frequencies used, the small transformers and filter capacitors used in a typical SMPS deliver BETTER performance than the much larger components used in a power supply that must operate at 50 Hz or 60 Hz. While it is true that a typical SMPS will generate a lot more high frequency noise internally than a linear power supply, which makes design and layout more critical, that high frequency noise is simpler to filter out than the 50 Hz and 60 Hz hum that is present in all linear supplies. (However, because the requirements are different, if an engineer who isn't familiar with the new technology tries to design an SMPS, the result is likely to be poor performance.)
Your statement about SMPS lacking in "dynamic power" is somewhat true - at least for some designs. However, it isn't a limitation of the technology, but is merely simply the design target for many early SMPS designs. In many applications, including computers, and many others which were typical applications for SMPS until recently, the ideal power supply would deliver perfectly clean and regulated power up to a set limit, at which point it would limit or shut down to protect itself and the associated circuitry... or, to say that another way, it would supply lots of perfectly clean power up to its rated limit, but have no dynamic power reserves whatsoever. (Which would you rather have, a 1000 watt power supply that can deliver peaks of 2000 watts, or a power supply that has no dynamic reserves at all, but can simply deliver exactly 2000 watts all the time?) Well, the reality is that you CAN design an SMPS that behaves more like a linear power supply, with a specified average power ratings, AND dynamic power reserves... which is what we used in the new X Gen3 series amps.
When you read reviews, or comments from armchair engineers, about how "this or that SMPS doesn't have enough dynamic power reserves" - they're probably right. But the reason is because the particular power supply they're talking about wasn't powerful enough to do the job continuously AND didn't have any dynamic power reserves that would allow it to exceed its continuous ratings short-term. (In other words, whoever chose or designed the power supply made a bad choice for use in that application.) We've been very careful NOT to make that mistake with the X Series Gen3 amps.
Class D amplifiers
Just to set the record straight.... There is no inherent limitation in Class D that limits frequency response, distortion, damping factor, output power, efficiency, or sound quality. Because they operate at high frequencies, and by using switching output stages, most Class D amplifiers are very efficient. However, while this is a major benefit, it is also a huge liability - because many Class D audio amplifiers were expressly designed with efficiency as their primary design goal. Yes, it really is that simple..... As with SMPS, designing Class D amplifiers is quite complex, there are a whole slew of possible design trade-offs, and many engineers who are familiar with older technologies are a lot less familiar with designing Class D equipment. As a result, a lot of early Class D audio amps met their primary design goal (efficiency), but failed miserably to sound good.
Combinations and permutations
In case there was any doubt, you can use either type of power supply with either type of amplifier. And, in the recent past, there have been Class D amplifiers that used linear power supplies... As with most things, this wasn't because of some deep mystical knowledge, but simply because, at that point in time, a linear power supply could deliver the power that was needed, at the required quality level, cheaper than using an SMPS.
The reason we chose to use a SMPS, but Class A/B output modules with Optimized Class H rail switching, in our new X Series Gen3 amps is simply that this combination gave us the best combination of performance specifications, sound quality, and cost. (As you probably know by now, we at Emotiva do our best to avoid following fads and trends, and buying this or that technology because "it's popular this year". We prefer to use the technology that offers the best value. Likewise, at the price point we chose for the Emersa line, the all-switching iCE modules provided the best overall value.)
Exploding myths
I love exploding things.... but, sadly, I've found the explosions caused by myths to be far less than aesthetically satisfactory.
Every year last year's myths get exploded, only to be replaced by new myths, which themselves get exploded next year. It's getting so you can;t walk down the street without tripping over, or hearing about, how this or that myth is now known to be untrue, and how "now we know the truth" (or was it "the truth is out there"?)
It is in fact quite possible to get very high efficiency from Class D amplifiers (I've seen designs that claim 95% at full power), and many current designs do in fact deliver higher efficiency than Class A/B designs. It is also quite possible to design Class D amplifiers with low efficiency. The real claim is that Class D amps CAN be very efficient - whereas a Class A/B amp with a linear fixed-rail power supply can never exceed about 70% efficiency, and only that at one specific load impedance . Incidentally, I'm pretty sure that Hypex is now recommending the use of SMPS with their amplifiers (I believe they used to recommend linear supplies).
(But, if you want to get into myths.... then you should consider all the myths about how, rather than using the circuit topology recommended by the manufacturer, who just might know what they're talking about, so many people seem determined to alter things to fit the requirements of this or that audiophile myth about some specific component or circuit topology sounding better than another. I wonder how many of those myths will be exploded spectacularly next year, or the year after.)
I STRONGLY urge you to avoid trying to armchair engineer how you THINK an amplifier with an SMPS or a linear power supply will or should sound.
After all, all that really counts is how it DOES sound, right?
The performance and sound quality of our new X Series Gen3 amps will speak for itself - literally.
People ALSO simply parrot the 'high efficiency' claim. This is TRUE, but at full power where Plug to Speaker efficiencies can easily exceed 80%. (W/SMPS) The output section ALONE can be in the 90s since it is either ON or OFF and the On-Resistance of the devices used is very small. This is one way to get (besides NFB) a good damping factor. At 'realistic' powers, the efficiency of the entire amp from plug to speaker is really not much better than a well executed A/B design.
www.ecnmag.com/article/2010/01/exploding-efficiency-myth-class-d-amplifiers
More complex solutions CAN yield higher efficiency at realistic power levels, but cost goes up fairly quickly and possible problems multiply. It is simply a guess on my part, but bringing in an EMOTIVA amp with a sophisticated SMPS AND using non-PS included modules from B&O would be quite a challenge at prices Emotiva Customers would be willing to pay. Jeff Roland, who makes some wacky expensive 'd' amps would have NO such problem. And I believe he uses N-Core from Hypex.
If I was an HT kind of guy, I'd for sure look at the XPA-5 which for the 800$ 'ask' is verging on 'bargain'. Even using 4 of the 5 channels to BiAmp a stereo pair of speakers would be a good use of such an amp.
