Home |
Search |
Today's Posts |
#1
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
ok, had a chance to audition the Crown class I amplifier (CTs2000),
and compare it to an EV CPS2 & my NAD 216thx. Speakers were my oldy, but goody Boston Acoustics A200s. Both the NAD and the EV use conventional toroidal transformer power supplies and linear output stages. The online literature for the CPS2 doesn't state what class amp it is, but my guess is A/B. The Crown is switch-mode coming and going. Ok the Crown has a ridiculous amount of power (1 kilowatt per side), and the low end was tight and visceral, but the overall sound is very odd. Very weird in the mids & highs. It sounded ... hard? Immediately fatiguing. From the next room it seemed to be fine... actually the next room was as close as I wanted to be. It was difficult to be in the same room with it. By comparison, both the EV and NAD sounded fine. I am assuming the bulk of the problem lays in the PWM output stage. But here lies a question.... Scott had mentioned that he didn't like the sound of "trick" power supplies. I've used a QSC with a switchmode pwer supply, but linear output stages in a PA situation, without hearing the horrible mids & highs that the Crown exhibited. Why should the power supply be especially problematic, if the output is a linear design? Secondly, class G or H amps are essentially tiers of linear output stages, are they not? Any good design examples of those? -glenn |
#2
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On May 24, 10:00*pm, "Soundhaspriority" wrote:
The Crown is a PWM amplifier:http://www.rickysaudio.com/docs/Ampl...0Explained.pdf so you're asking in a general sense? yes, in a general sense. What is it, specifically, about a switch mode power supply that adversely affects a linear output stage? (many of the QSC amps, as an example, have a SMPS with a class A/B output stage) Secondly, class G or H amps are essentially tiers of linear output stages, are they not? A related technique is "current dumping." They are not symmetrical tiers. No, the tiers are not symmetrical, but they are linear output stages, yes? Any good design examples of those? NAD did it, but reviews indicated a dark sound. It's almost universal practice in HT receivers, but with mediocre results. The NAD stuff was their early "power envelope" designs. They got away from that in the late 90s. With respect to the power supply, the impedance characteristics of a switchmode supply is usually different from a linear supply. A linear supply MUST have a large capacitor bank, because the current source that charges the bank goes to zero 120 times a second. A switching supply operates, at a minimum, at ultrasonic frequency. Therefore, the capacitor bank can be made tiny. A computer supply that provides 600 watts continuously has capacitors the size of your thumb. However, the dynamic impedance of these caps, as opposed to the very much larger caps in a linear supply, is much larger. If a switching supply were used in combination with a large, low ESR capacitor bank, the only remaining question would be the contamination of fast amplifier circuitry with residual ultrasonic noise that linear supplies do not have. Can't that easily be filtered out? What sort of frequencies are we talking about? Over the years, many people with sensitive ears have been dissatisfied with sound reinforcement amps for critical listening. The priorities of design are rather different. Well, they are sometimes. Personally, I wouldn't worry too much about the PA amp if everyone's up there singing thru Shure sm58s, but I digress. Hey Arny, ever ABX a conventional and a switching design? -glenn |
#3
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On May 24, 10:00 pm, "Soundhaspriority" wrote:
"geezer" wrote in message With respect to the power supply, the impedance characteristics of a switchmode supply is usually different from a linear supply. A linear supply MUST have a large capacitor bank, because the current source that charges the bank goes to zero 120 times a second. A switching supply operates, at a minimum, at ultrasonic frequency. Therefore, the capacitor bank can be made tiny. A computer supply that provides 600 watts continuously has capacitors the size of your thumb. However, the dynamic impedance of these caps, as opposed to the very much larger caps in a linear supply, is much larger. Actually, both switching and linear power supplies need substantial filter capacitors. However, because computer switching supplies provide regulation very efficiently over a wide range of input voltages, the capacitors can be considerably smaller. Ripple that would be intolerable in a linear supply is essentially eliminated in a computer switching supply. The impedance of a switching supply should be vanishingly low as long as it is able to provide regulation and would approach an ideal voltage source. Whether or not voltage is tightly regulated in an audio amplifier with switching power supply is another issue. If a switching supply were used in combination with a large, low ESR capacitor bank, the only remaining question would be the contamination of fast amplifier circuitry with residual ultrasonic noise that linear supplies do not have. Given that some very impressive sound cards have been produced that function well inside of a PC, its not clear that noise should be too much of a problem here with a well-designed power supply. Of course the key phrase is "well designed" and achieving that can be challenging. |
#4
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On Sat, 24 May 2008 20:57:54 -0700 (PDT), geezer
wrote: What is it, specifically, about a switch mode power supply that adversely affects a linear output stage? (many of the QSC amps, as an example, have a SMPS with a class A/B output stage) Hoping to avoid entering a religious discussion, maybe I could add a comment about power supply "refresh". A conventional "linear" (everyone please don't complain; just go with the flow) power supply is refreshed from a very very stiff wall outlet and at 120 refreshes per second. An SMPS power supply (typically; other designs are possible and could give superior results) have a conventional "linear" power supply upstream of an A/D/A-regulated switching device. System output voltage response is spec'd by the A/D/A regulator's response time *with small signals*, and here it's very stiff if within the regulator's slewing ability. With large signals, most practical designs will fall out of small-signal pretty quickly and become limited both by the upstream "linear" supply and by the slewing rate of the A/D/A regulaion. Could an SMPS have as stiff an output voltage as a conventional "linear" supply? Yes; even better, if the money were available. Is this the usual case? Not usually. A related issue is that refresh rates fall within the audio (current-drawing) range. Our audio devices require current both above *and* below refresh rate. Enough babbling. Just glad to be home from a week of serious training. Anybody want to know anything about Crestron programming? And I do mean *anything*! Much thanks, as always, Chris Hornbeck "I'm walking out, between parked cars, with my head full of stars" -Elliott Smith |
#5
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
Chris Hornbeck wrote:
snip Enough babbling. Just glad to be home from a week of serious training. Anybody want to know anything about Crestron programming? And I do mean *anything*! No, but what do you know about AMX? jak Much thanks, as always, Chris Hornbeck "I'm walking out, between parked cars, with my head full of stars" -Elliott Smith |
#6
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On Sun, 25 May 2008 01:09:36 -0500, jakdedert
wrote: Anybody want to know anything about Crestron programming? And I do mean *anything*! No, but what do you know about AMX? AMX is the spawn of Santa, or maybe somebody closely related, or similarly spelled... Actually, I currently know more about Crestron progamming than I really *want* to know. I'm hoping to offload some of the knowledge to some other poor vessel, freeing up room in my head for some more jingles, and memories of pretty women. But seriously, thanks, as always, Chris Hornbeck "I'm walking out, between parked cars, with my head full of stars" -Elliott Smith |
#7
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
"Soundhaspriority" wrote in message ... "jwvm" wrote in message ... On May 24, 10:00 pm, "Soundhaspriority" wrote: "geezer" wrote in message With respect to the power supply, the impedance characteristics of a switchmode supply is usually different from a linear supply. A linear supply MUST have a large capacitor bank, because the current source that charges the bank goes to zero 120 times a second. A switching supply operates, at a minimum, at ultrasonic frequency. Therefore, the capacitor bank can be made tiny. A computer supply that provides 600 watts continuously has capacitors the size of your thumb. However, the dynamic impedance of these caps, as opposed to the very much larger caps in a linear supply, is much larger. Actually, both switching and linear power supplies need substantial filter capacitors. However, because computer switching supplies provide regulation very efficiently over a wide range of input voltages, the capacitors can be considerably smaller. Ripple that would be intolerable in a linear supply is essentially eliminated in a computer switching supply. The impedance of a switching supply should be vanishingly low as long as it is able to provide regulation and would approach an ideal voltage source. Whether or not voltage is tightly regulated in an audio amplifier with switching power supply is another issue. It should be. I don't see a switcher as poison, provided it's done right. Maybe the designers don't listen as much as they used to, or maybe it isn't important to them. If a switching supply were used in combination with a large, low ESR capacitor bank, the only remaining question would be the contamination of fast amplifier circuitry with residual ultrasonic noise that linear supplies do not have. Given that some very impressive sound cards have been produced that function well inside of a PC, its not clear that noise should be too much of a problem here with a well-designed power supply. Of course the key phrase is "well designed" and achieving that can be challenging. It is amazing. Bob Morein (310) 237-6511 The pro amps I've listened to in a hi-fi situation are (in order of sound quality): Crown K1, Crown MT1200, Crown K2, Crown CE4000, Mackie 1400I. The biggest difference was between the K1 and the MT. For hi-fi use, the K1 sounded like crap. The old Hafler 500 is much better, as is anything from Linn or Arcam, Creek, NAD,....... |
#8
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
Chris Hornbeck wrote:
On Sun, 25 May 2008 01:09:36 -0500, jakdedert wrote: Anybody want to know anything about Crestron programming? And I do mean *anything*! No, but what do you know about AMX? AMX is the spawn of Santa, or maybe somebody closely related, or similarly spelled... It's no gift, so I guess it must be... Seriously. AMX programmers make so much money because it's so damn difficult and arcane...plus every installation becomes obsolete after awhile and needs to be reprogrammed to accommodate new equipment. I had to deal regularly with one such installation (100 seat corporate HQ theater/conference room). It was nice when it was new, but as they upgraded the gear, they neglected the programming. By the time I got there, it was a mess of consumer, prosumer, pro and architectural stuff. The touch screen still controlled the room, lectern computer, roll-down screen and overall volume in the PA. The projector, DVD, and several other pieces had been upgraded, but not programmed into the AMX. To switch from powerpoint presentation to DVD, requires seven or eight different actions, on five pieces of gear, using two IR remotes (DVD and projector), direct manipulation of the sound console etc, plus the AMX screen...when all functions could have been controlled by a single virtual button punch on the AMX. If they were made to be programmed by mortals, or someone would simply design a GUI programming decompiler (or 'learning' IR interface) it would help immensely. jak Actually, I currently know more about Crestron progamming than I really *want* to know. I'm hoping to offload some of the knowledge to some other poor vessel, freeing up room in my head for some more jingles, and memories of pretty women. But seriously, thanks, as always, Chris Hornbeck "I'm walking out, between parked cars, with my head full of stars" -Elliott Smith |
#9
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
geezer wrote:
Ok the Crown has a ridiculous amount of power (1 kilowatt per side), and the low end was tight and visceral, but the overall sound is very odd. Very weird in the mids & highs. It sounded ... hard? Immediately fatiguing. From the next room it seemed to be fine... actually the next room was as close as I wanted to be. It was difficult to be in the same room with it. This has basically been my experience with all of the class-D amplifiers that I have tried. I am assuming the bulk of the problem lays in the PWM output stage. But here lies a question.... Scott had mentioned that he didn't like the sound of "trick" power supplies. I've used a QSC with a switchmode pwer supply, but linear output stages in a PA situation, without hearing the horrible mids & highs that the Crown exhibited. Why should the power supply be especially problematic, if the output is a linear design? The thing with most of the trick supplies of various sorts is that they are designed to minimize needing big supply filter caps. Those big filter caps also provide a reservoir to supply a lot of current on brief transients too. At least, that's the only reason I can think of. But try one of the Stewarts with the switching supply and decide for yourself. Secondly, class G or H amps are essentially tiers of linear output stages, are they not? Any good design examples of those? Anything other than Class A, AB, B, or C is something made up by the manufacturers. Everybody pretty much agrees what you mean by calling an amp "Class D" although it's not an official designation. G and H are whatever you want them to be. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#10
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On May 26, 8:49*pm, (Scott Dorsey) wrote:
The thing with most of the trick supplies of various sorts is that they are designed to minimize needing big supply filter caps. *Those big filter caps also provide a reservoir to supply a lot of current on brief transients too. I thought the rationale was to eliminate the big honking transformer, withall the weight & current draw that goes along with it. Since the switching supply is more efficient at supplying power on demand, you no longer need the big storage caps. Do I have this wrong? If it's paired with a linear output stage (as in the case of the QSC amps which have an SMPS coupled to an AB output) where do the perceived sonic problems come from? At least, that's the only reason I can think of. But try one of the Stewarts with the switching supply and decide for yourself. I was under the impression that the Stewarts were one the worst examples. The Stewarts I've seen are incredibly small, aren't they also switch-mode on the output side, too? Anything other than Class A, AB, B, or C is something made up by the manufacturers. *Everybody pretty much agrees what you mean by calling an amp "Class D" although it's not an official designation. *G and H are whatever you want them to be. so class D are really class C, and Class G & H are really A or B or AB? There seems to be covert agreement then among amp manufacturers. Lots of manufacturers tout class H these days. The so-called class G or H seem to be tiered output stages, which shift between lower or higher ouput stages dependent on input signal demand. In either case, aren't those tiers linear output stages? There'd be no need for the tiers if it was a switching output... again, unless I have it all wrong. -glenn |
#11
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
geezer wrote:
On May 26, 8:49=A0pm, (Scott Dorsey) wrote: The thing with most of the trick supplies of various sorts is that they are designed to minimize needing big supply filter caps. =A0Those big filter caps also provide a reservoir to supply a lot of current on brief transients too. I thought the rationale was to eliminate the big honking transformer, withall the weight & current draw that goes along with it. Since the switching supply is more efficient at supplying power on demand, you no longer need the big storage caps. Do I have this wrong? No, what is happening is that because the noise coming out of the switching supply is all very high frequency stuff, instead of at 120 Hz, the filter caps in the power supply only need to deal with very high frequencies, so they can be much smaller values. When you do this, though, you ALSO lose the secondary benefits of having big reservoir caps in the supply. If it's paired with a linear output stage (as in the case of the QSC amps which have an SMPS coupled to an AB output) where do the perceived sonic problems come from? I am thinking it is due to the lack of ability to supply large currents at short peaks, which is a pleasant side effect of having a big 120 Hz filter on the output nof your power supply. At least, that's the only reason I can think of. But try one of the Stewarts with the switching supply and decide for yourself. I was under the impression that the Stewarts were one the worst examples. The Stewarts I've seen are incredibly small, aren't they also switch-mode on the output side, too? The ones I tried were class B amplifiers with a switching supply, and they were pretty awful. But even try something like the Carver amps which switch supply voltages in the middle of the waveform using dual linear supplies and some switching logic. Anything other than Class A, AB, B, or C is something made up by the manufacturers. =A0Everybody pretty much agrees what you mean by calling an amp "Class D" although it's not an official designation. =A0G and H are whatever you want them to be. so class D are really class C, and Class G & H are really A or B or AB? A class D amplifier is a modulator stage, followed by a class C amplifier, followed by an integrator stage. And that's really kind of an informal designation, but everybody pretty much accepts it today. There seems to be covert agreement then among amp manufacturers. Lots of manufacturers tout class H these days. The so-called class G or H seem to be tiered output stages, which shift between lower or higher ouput stages dependent on input signal demand. In either case, aren't those tiers linear output stages? There'd be no need for the tiers if it was a switching output... again, unless I have it all wrong. Dunno, but you have to look at these things on an individual basis. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#12
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On May 27, 10:49*am, (Scott Dorsey) wrote:
I am thinking it is due to the lack of ability to supply large currents at short peaks, which is a pleasant side effect of having a big 120 Hz filter on the output nof your power supply. I thought the switching supplies COULD supply all the current needed. If not, why not just use big capacitors? You'd still be saving the expense & weight of the transformer... |
#13
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
geezer wrote:
On May 27, 10:49=A0am, (Scott Dorsey) wrote: I am thinking it is due to the lack of ability to supply large currents at short peaks, which is a pleasant side effect of having a big 120 Hz filter on the output nof your power supply. I thought the switching supplies COULD supply all the current needed. That's the basic rationale from the designers. But if that were the case, why don't they sound so good? If not, why not just use big capacitors? You'd still be saving the expense & weight of the transformer... Because convincing designers that it's a problem is nontrivial. For all I know, though, that might be why the Chevin amps sound much better than the average switching supply amps. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#14
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
Scott Dorsey wrote:
geezer wrote: Ok the Crown has a ridiculous amount of power (1 kilowatt per side), and the low end was tight and visceral, but the overall sound is very odd. Very weird in the mids & highs. It sounded ... hard? Immediately fatiguing. From the next room it seemed to be fine... actually the next room was as close as I wanted to be. It was difficult to be in the same room with it. This has basically been my experience with all of the class-D amplifiers that I have tried. I am assuming the bulk of the problem lays in the PWM output stage. But here lies a question.... Scott had mentioned that he didn't like the sound of "trick" power supplies. I've used a QSC with a switchmode pwer supply, but linear output stages in a PA situation, without hearing the horrible mids & highs that the Crown exhibited. Why should the power supply be especially problematic, if the output is a linear design? The thing with most of the trick supplies of various sorts is that they are designed to minimize needing big supply filter caps. Those big filter caps also provide a reservoir to supply a lot of current on brief transients too. At least, that's the only reason I can think of. But try one of the Stewarts with the switching supply and decide for yourself. Another way of looking at it is that the audio output current in most amps flows through the power supply (except for a full bridge). Draw the circuit as a bridge, one half is the output stage, the other half is the power supply. The large caps are undoubtably much more linear than the smps with small caps. You're injecting any transients due to overshoot and other non-linearities into the audio. Jerry |
#15
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
On May 27, 4:40*pm, Jerry Peters wrote:
Another way of looking at it is that the audio output current in most amps flows through the power supply (except for a full bridge). Draw the circuit as a bridge, one half is the output stage, the other half is the power supply. The large caps are undoubtably much more linear than the smps with small caps. You're injecting any transients due to overshoot and other non-linearities into the audio. Isn't all that ultrasonic junk? If so, it should be easy enough to filter out, no? If not, it seems the solution is obvious - SMPS with lots of cap storage - it solves the problems weight, current and most of the expense. -glenn |
#16
Posted to rec.audio.pro
|
|||
|
|||
amplifier redux
geezer wrote:
On May 27, 4:40*pm, Jerry Peters wrote: Another way of looking at it is that the audio output current in most amps flows through the power supply (except for a full bridge). Draw the circuit as a bridge, one half is the output stage, the other half is the power supply. The large caps are undoubtably much more linear than the smps with small caps. You're injecting any transients due to overshoot and other non-linearities into the audio. Isn't all that ultrasonic junk? If so, it should be easy enough to filter out, no? If not, it seems the solution is obvious - SMPS with lots of cap storage - it solves the problems weight, current and most of the expense. Yep, that would be my solution. I eliminates the cost, size, & weight of the 60Hz transformer. I'm assuming the designer just looked at the power supply specs without considering the transient performance, and that the power supply is effectively in series with the speaker load. Jerry |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
mic cables redux | Pro Audio | |||
Redux Trident S-20 | Pro Audio | |||
Kwestion for the Krooborg (redux) | Audio Opinions |