I just read an eBay ad for an X350 that said that it was 50W Pure A, and 350W total. I'm wondering how this is accomplished?
Is this a 400W Class AB amp with the input level shifted / biased to allow 50W of swing to occur on one set of transistors (ie the set that handles the "+" portion of the Class AB wave)? Then when the signal goes above 50W (or whatever voltage this corresponds to), the "-" set of transistors starts to conduct during portions of the cycle? If so, could this be biased even higher to get a higher power 175-200W Class A wave at the expense of a lower or nonexistent higher power class AB output and an astronomical electric bill?
The search tool returned 8712 results. None of the entries on the first few pages seemed to cover this topic. Sorry for my English - I graduated from a public school somewhere in the Midwest
Thanks everyone,
Bryan A. Thompson
bryan@batee.com
Is this a 400W Class AB amp with the input level shifted / biased to allow 50W of swing to occur on one set of transistors (ie the set that handles the "+" portion of the Class AB wave)? Then when the signal goes above 50W (or whatever voltage this corresponds to), the "-" set of transistors starts to conduct during portions of the cycle? If so, could this be biased even higher to get a higher power 175-200W Class A wave at the expense of a lower or nonexistent higher power class AB output and an astronomical electric bill?
The search tool returned 8712 results. None of the entries on the first few pages seemed to cover this topic. Sorry for my English - I graduated from a public school somewhere in the Midwest
Thanks everyone,
Bryan A. Thompson
bryan@batee.com
50W is the audio output power & 350W is the total power consumption of the system.
check this link for more about amplifier efficiency
Amplifier efficiency
check this link for more about amplifier efficiency
Amplifier efficiency
Hi Bryan,
it probably could be biased a lot higher to get more class-a power.
The problem is that the case / power supply / heatsinks are dimensioned for 600W idle.
So if you up the bias X% the power dissipation at idle would also go up X% and the temperature would become higher.
To change the output power (class A ) from 50 watts to 175 watts you would need to up the bias by a factor of 2. This would also change the temperature from around 50°C at ambient 20° to 75-80° at ambient 20°.
I suppose the 50 watts are chosen because most of the time you will need a lot less than that.
William
it probably could be biased a lot higher to get more class-a power.
The problem is that the case / power supply / heatsinks are dimensioned for 600W idle.
So if you up the bias X% the power dissipation at idle would also go up X% and the temperature would become higher.
To change the output power (class A ) from 50 watts to 175 watts you would need to up the bias by a factor of 2. This would also change the temperature from around 50°C at ambient 20° to 75-80° at ambient 20°.
I suppose the 50 watts are chosen because most of the time you will need a lot less than that.
William
I understand that raising the bias would produce more heat that has to be dealt with.
I asked the question because I'm curious to know what technology is used to allow the switch from Class A to Class AB at higher power levels, not because I want to hotrod one of Nelson's amps...
I assume if it's Class AB with bias, then it still has crossover distortion at the point where one set stops conducting and another set starts. Does the output wave reflect the crossover distortion at a lower/higher portion of the wave rather than at the center? Or does the bias change as the load (or output power) changes, to keep the crossover at the center of the wave?
Thanks!
I asked the question because I'm curious to know what technology is used to allow the switch from Class A to Class AB at higher power levels, not because I want to hotrod one of Nelson's amps...
I assume if it's Class AB with bias, then it still has crossover distortion at the point where one set stops conducting and another set starts. Does the output wave reflect the crossover distortion at a lower/higher portion of the wave rather than at the center? Or does the bias change as the load (or output power) changes, to keep the crossover at the center of the wave?
Thanks!
)
Bryan Hi,
To further knock on the nail...
Pure Class A of 200 watts see the monobloc XA200:
http://www.passlabs.com/amplifiers/xa200.htm
This is THE pure Class A monobloc of 200 Watts in 8 Ohms load...
Idle at 650 watts for each monobloc. {it's also the max power consumption}
2 * 175 lb weight. {2 case of X1000}
Nelson did it !!
- - - - - - - - - - - - - - -
And for Class AB look at the X250
http://www.passlabs.com/amplifiers/x250_5.htm
Idle at 350 watts ; max 1000 watts for this stereo amplifier
110/150 lb weight total.
But you cannot bias it higher ; you are limited by the heatsink dissipation !!
And No, there is no "crossover distortion" when going from Class A to AB
in these design. You only ask the Power Supply for more current!
{If I remember, Nelson or Grey did answer about this point some time ago}
Take a look at the manual of each model...
I had a chance to listen to a pair of XA160 a few years ago at Festival du Son
in Montreal ; and WOW, that's nothing to compare with a class AB,
Pure delight for your ears!!! "fasten you seat belts" ; it's pure Magic !!
Not saying that the X350 is not a great amplifier, but the XA 160/200 are
on the "upper class" without any doubt!!!
Regards.
PS: On that day, back home I was still delighted by my little A75 {highly biased!}
pushing the Martin Logan Aerius' i {but missing the XA 160 WOW!!!}
Alain.
To further knock on the nail...
Pure Class A of 200 watts see the monobloc XA200:
http://www.passlabs.com/amplifiers/xa200.htm
This is THE pure Class A monobloc of 200 Watts in 8 Ohms load...
Idle at 650 watts for each monobloc. {it's also the max power consumption}
2 * 175 lb weight. {2 case of X1000}
Nelson did it !!
- - - - - - - - - - - - - - -
And for Class AB look at the X250
http://www.passlabs.com/amplifiers/x250_5.htm
Idle at 350 watts ; max 1000 watts for this stereo amplifier
110/150 lb weight total.
But you cannot bias it higher ; you are limited by the heatsink dissipation !!
And No, there is no "crossover distortion" when going from Class A to AB
in these design. You only ask the Power Supply for more current!
{If I remember, Nelson or Grey did answer about this point some time ago}
Take a look at the manual of each model...
I had a chance to listen to a pair of XA160 a few years ago at Festival du Son
in Montreal ; and WOW, that's nothing to compare with a class AB,
Pure delight for your ears!!!
"fasten you seat belts" ; it's pure Magic !!Not saying that the X350 is not a great amplifier, but the XA 160/200 are
on the "upper class" without any doubt!!!
Regards.
PS: On that day, back home I was still delighted by my little A75 {highly biased!}
pushing the Martin Logan Aerius' i {but missing the XA 160 WOW!!!}
Alain.
Wasn't the Aleph 1.2 a 200W amp? That would make two, although the 1.2 is not in production at this time.
As far as the A vs. AB discussion...it looks distressingly like the cook throwing spaghetti at the wall to see if it sticks. Indeed, some did, but an awful lot fell to the floor and it's in danger of turning into a terrible mess.
The X amplifiers don't use any 'technology' to 'switch' the bias from class A to class AB. There's no dynamic biasing, no sliding bias, no funky sales-spawned bafflegab. The output stage is a conventional complementary follower. Nothing weird. The "X-ness" is all in the front end. Nothing in the output. It just happens to be biased somewhat hotter than some of the other class AB designs out there.
The idea of running class A up to some arbitrary power level before shifting to class B is a nifty one. Almost like having your cake and eating it too. The question is where to set the bias, which in turn determines where the A-to-B transition will occur. Some people set it as low as possible; their intention being to keep heat dissipation to a minimum. As a corollary--something you never see pointed out in the sales literature--they get to use cheaper transformers and heatsinks. Hmmm. There's no way that could factor into their decision, is there? Nah...not a chance.
On the other hand, you've got people who use higher biases in order to keep the amplifier functioning in class A for as long as possible. For an average--or even above average--listener, even when they decide to kick some backside, 50W of class A will pretty much cover all the bases and then some. Remember that musical peaks are greatly higher than the average listening level. By the time you've used up the 50W of class A, odds are pretty good that you're clipping peaks anyway.
And clipping is a whole lot worse than any namby-pamby crossover distortion. If you've used up your alloted 50W class A, and the 300W of class B behind that, you've got far worse problems than any of the usual bench-test distortions.
Grey
P.S.: The most inefficient commercial speaker I can remember was the Stax F-81--an electrostat. Incredible midrange. Highs good to very good. Lows were weak due to the relatively small size of the panels. I believe it was rated at something like 80 or 82dB. It took 200W to get it moving. Nonetheless, I lusted for a pair of those things. That midrange got hold of me and would not let go.
These days, assuming I was in the mood to buy electrostats, I'd probably go for something else, but back then I wanted a pair of F-81s in the worst kind of way.
(There was a stacked version--the F-83 [I think I got that right]--but I never had a chance to hear them.)
As far as the A vs. AB discussion...it looks distressingly like the cook throwing spaghetti at the wall to see if it sticks. Indeed, some did, but an awful lot fell to the floor and it's in danger of turning into a terrible mess.
The X amplifiers don't use any 'technology' to 'switch' the bias from class A to class AB. There's no dynamic biasing, no sliding bias, no funky sales-spawned bafflegab. The output stage is a conventional complementary follower. Nothing weird. The "X-ness" is all in the front end. Nothing in the output. It just happens to be biased somewhat hotter than some of the other class AB designs out there.
The idea of running class A up to some arbitrary power level before shifting to class B is a nifty one. Almost like having your cake and eating it too. The question is where to set the bias, which in turn determines where the A-to-B transition will occur. Some people set it as low as possible; their intention being to keep heat dissipation to a minimum. As a corollary--something you never see pointed out in the sales literature--they get to use cheaper transformers and heatsinks. Hmmm. There's no way that could factor into their decision, is there? Nah...not a chance.
On the other hand, you've got people who use higher biases in order to keep the amplifier functioning in class A for as long as possible. For an average--or even above average--listener, even when they decide to kick some backside, 50W of class A will pretty much cover all the bases and then some. Remember that musical peaks are greatly higher than the average listening level. By the time you've used up the 50W of class A, odds are pretty good that you're clipping peaks anyway.
And clipping is a whole lot worse than any namby-pamby crossover distortion. If you've used up your alloted 50W class A, and the 300W of class B behind that, you've got far worse problems than any of the usual bench-test distortions.
Grey
P.S.: The most inefficient commercial speaker I can remember was the Stax F-81--an electrostat. Incredible midrange. Highs good to very good. Lows were weak due to the relatively small size of the panels. I believe it was rated at something like 80 or 82dB. It took 200W to get it moving. Nonetheless, I lusted for a pair of those things. That midrange got hold of me and would not let go.
These days, assuming I was in the mood to buy electrostats, I'd probably go for something else, but back then I wanted a pair of F-81s in the worst kind of way.
(There was a stacked version--the F-83 [I think I got that right]--but I never had a chance to hear them.)
Let's say hypothetically that the output of an amplifier consists of one transistor pulling the output towards positive voltage, and the other pushing the output towards negative voltage. The pulling transistor is always on at positive output voltages, and the pushing transistor is always on at negative output voltages. The amount of bias current controls what point the transistors turn off.
A Class B amp would have no bias, so that both the pulling and pushing transistor turn off when they hit zero voltage. This give large crossover distortion, as one transistor is turning on (causing noise) at the exact same time the other transistor is turning off (causing more noise), and because both transistors are turning on/off at the same time, any feedback loop present can't do anything to correct or reduce the noise.
A Class A amp has a lot of bias current so that the pulling transistor stays on all the way to the maximum negative voltage, and the pushing transistor stays on all the way to the maximum positive voltage. Or in other words, both transistors are on all the time. Thus the transistors never go through the on/off stage, so there is no crossover distortion.
A Class AB amp has some bias, but not enough to be Class A. So the pulling transistor turns off at few volts negative, and the pushing transistor turns off at a few volts positive. So now there is a range of output voltages where both transistors are on, which is Class A operation. When the output goes to a positive voltage above that range and the pushing transistor turns off causing noise at the transistion, but since the pulling transistor is on at that point feedback can operate to reduce the transistion noise. The same applies to when the output goes to a negative voltage outside the Class A range. Now the pulling transistor turns off at the transition causing noise, but the pushing transistor stays on and allows feedback to reduce the transistion noise.
So there isn't any any real technique to achieving Class AB operation. It's a simple function of setting how much bias current is used in an amplifiers output stage.
Now, I wouldn't worry too much about only having the first 50W operating in Class A. My main amps are pure Class A and clip at 50W output. Even when using lowish efficiency speakers (down around 85dB/W/m), I have yet to clip my amps because I find it too loud before that happens. With 50W Class A before transistioning to Class AB, it will only be the very occasional dynamic peak that actually takes you out of Class A.
OK, let's see if I've confused things enough. Flame on!
A Class B amp would have no bias, so that both the pulling and pushing transistor turn off when they hit zero voltage. This give large crossover distortion, as one transistor is turning on (causing noise) at the exact same time the other transistor is turning off (causing more noise), and because both transistors are turning on/off at the same time, any feedback loop present can't do anything to correct or reduce the noise.
A Class A amp has a lot of bias current so that the pulling transistor stays on all the way to the maximum negative voltage, and the pushing transistor stays on all the way to the maximum positive voltage. Or in other words, both transistors are on all the time. Thus the transistors never go through the on/off stage, so there is no crossover distortion.
A Class AB amp has some bias, but not enough to be Class A. So the pulling transistor turns off at few volts negative, and the pushing transistor turns off at a few volts positive. So now there is a range of output voltages where both transistors are on, which is Class A operation. When the output goes to a positive voltage above that range and the pushing transistor turns off causing noise at the transistion, but since the pulling transistor is on at that point feedback can operate to reduce the transistion noise. The same applies to when the output goes to a negative voltage outside the Class A range. Now the pulling transistor turns off at the transition causing noise, but the pushing transistor stays on and allows feedback to reduce the transistion noise.
So there isn't any any real technique to achieving Class AB operation. It's a simple function of setting how much bias current is used in an amplifiers output stage.
Now, I wouldn't worry too much about only having the first 50W operating in Class A. My main amps are pure Class A and clip at 50W output. Even when using lowish efficiency speakers (down around 85dB/W/m), I have yet to clip my amps because I find it too loud before that happens. With 50W Class A before transistioning to Class AB, it will only be the very occasional dynamic peak that actually takes you out of Class A.
OK, let's see if I've confused things enough. Flame on!
GRollins said:
Mr Rollins,
even worse, 76 dB for both the ELS-F81 and the ELS-F83.
The ELS-F83 indeed consisted of 2 stacked driver units, the same driver as in the F81x.
The F83 had independant input terminals for each driver, to go bi-amping.
The F83x from around 1990 was even nicer than the 1983? F83, difference between the models was in the auxilliary hardware, not much change to the driver panels.
The panels only handled 200 and 300 watts, the minimum impedance of 4 ohms for the F83 was considerably higher than that of the F81.
Quite a few Stax were sold in France, the French still have a conection with Japan, a set of F83s with DMA-X2s can be described as pretty painfull.
Lusted is a rather good description, got the same feeling from the 1st ed. ML Statement, caught those on their world tour in France as well.
I now have a better understanding of how the complementary pair works. Thank you.
At first glance, Zenv1 Fig1 appears to be very similar to a common emitter amp in BJT transistor terminology, and I am familiar with BJT biasing techniques. Applying Load Line techniques, I understand how the combination of Id and the voltage divider network at the gate establish the Q point and affect swing.
In reading through the above and Nelson's Zen articles, I am left with this question. Why are current sources (either Active/Aleph or CCS) used in conjunction with a MOSFET (say in Common Source config), as opposed to using a resistor to connect the drain to Vcc? What does this really do? I know that replacing emitter resistor in the Common Emitter amp with an ideal inductor significantly increases efficiency of that Class A amp. Is that why the CCS is used? To stiffen the supply or to change the characteristics of the load line to the AC signal?
If anyone is interested, there's some awesome info on the MIT OpenCourseware sites listed below. After viewing a few of these courses (esp the Physics 8.02/Electricity and Magnetism video course), it occurs to me just how bad my state University EE program (ranked #9 in the country at the time) really was. I believe I remember hearing the phrase "You'll never need to know the information in Chapter three, so we'll skip it, too." I still have the Electronic Circuits book, and Chapter three is titled "The Field-Effect Transistor".
http://ocw.mit.edu/OcwWeb/Electrica...Science/6-012Fall-2005/LectureNotes/index.htm
http://ocw.mit.edu/OcwWeb/Physics/8-02Electricity-and-MagnetismSpring2002/VideoLectures/
Thank you all very much for the information you have provided and for your continued patience as I struggle to understand MOSFETs in Analog mode.
Bryan
At first glance, Zenv1 Fig1 appears to be very similar to a common emitter amp in BJT transistor terminology, and I am familiar with BJT biasing techniques. Applying Load Line techniques, I understand how the combination of Id and the voltage divider network at the gate establish the Q point and affect swing.
In reading through the above and Nelson's Zen articles, I am left with this question. Why are current sources (either Active/Aleph or CCS) used in conjunction with a MOSFET (say in Common Source config), as opposed to using a resistor to connect the drain to Vcc? What does this really do? I know that replacing emitter resistor in the Common Emitter amp with an ideal inductor significantly increases efficiency of that Class A amp. Is that why the CCS is used? To stiffen the supply or to change the characteristics of the load line to the AC signal?
If anyone is interested, there's some awesome info on the MIT OpenCourseware sites listed below. After viewing a few of these courses (esp the Physics 8.02/Electricity and Magnetism video course), it occurs to me just how bad my state University EE program (ranked #9 in the country at the time) really was. I believe I remember hearing the phrase "You'll never need to know the information in Chapter three, so we'll skip it, too." I still have the Electronic Circuits book, and Chapter three is titled "The Field-Effect Transistor".
http://ocw.mit.edu/OcwWeb/Electrica...Science/6-012Fall-2005/LectureNotes/index.htm
http://ocw.mit.edu/OcwWeb/Physics/8-02Electricity-and-MagnetismSpring2002/VideoLectures/
Thank you all very much for the information you have provided and for your continued patience as I struggle to understand MOSFETs in Analog mode.
Bryan
For those who feel that 50W is enough: My BMW goes 145MPH before the rev limiter kicks in (yes, I tested it), and that is not enough. To generalize, whatever it is that Americans have, we want more. This passage from Key Largo sums it up:
Johnny Rocco: There's only one Johnny Rocco.
James Temple: How do you account for it?
Frank McCloud: He knows what he wants. Don't you, Rocco?
Johnny Rocco: Sure.
James Temple: What's that?
Frank McCloud: Tell him, Rocco.
Johnny Rocco: Well, I want uh ...
Frank McCloud: He wants more, don't you, Rocco?
Johnny Rocco: Yeah. That's it. More. That's right! I want more!
James Temple: Will you ever get enough?
Frank McCloud: Will you, Rocco?
Johnny Rocco: Well, I never have. No, I guess I won't. You, do you know what you want?
Frank McCloud: Yes, I had hopes once, but I gave them up.
Johnny Rocco: Hopes for what?
Frank McCloud: A world in which there's no place for Johnny Rocco.
PS: My goal is not to create a 500W or even a 200W Class A amp. I built a 9V headphone amp for my iPod and that's how I found DIYAudio. I currently don't own a single pair of speakers. I have a pair of the Hifizen Aleph-X boards (thank you, Grey) and that'll probably be the end of my audio experimentation. But for some reason I don't really understand, I have an interest in filling in the gaps in my EE education. I'm an American. So I want more.
Bryan
Johnny Rocco: There's only one Johnny Rocco.
James Temple: How do you account for it?
Frank McCloud: He knows what he wants. Don't you, Rocco?
Johnny Rocco: Sure.
James Temple: What's that?
Frank McCloud: Tell him, Rocco.
Johnny Rocco: Well, I want uh ...
Frank McCloud: He wants more, don't you, Rocco?
Johnny Rocco: Yeah. That's it. More. That's right! I want more!
James Temple: Will you ever get enough?
Frank McCloud: Will you, Rocco?
Johnny Rocco: Well, I never have. No, I guess I won't. You, do you know what you want?
Frank McCloud: Yes, I had hopes once, but I gave them up.
Johnny Rocco: Hopes for what?
Frank McCloud: A world in which there's no place for Johnny Rocco.
PS: My goal is not to create a 500W or even a 200W Class A amp. I built a 9V headphone amp for my iPod and that's how I found DIYAudio. I currently don't own a single pair of speakers. I have a pair of the Hifizen Aleph-X boards (thank you, Grey) and that'll probably be the end of my audio experimentation. But for some reason I don't really understand, I have an interest in filling in the gaps in my EE education. I'm an American. So I want more.
Bryan
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.